breast cancer atlas of clinical oncology

American Cancer Society

Atlas of Clinical Oncology

Series Volumes

Blumgart, Jarnagin, Fong Hepatobiliary Cancer

Cameron Pancreatic Cancer

Carroll Prostate Cancer

Char Cancer of the Eye and Orbit

Clark, Duh, Jahan, Perrier Endocrine Tumors

Eifel, Levenback Cervical, Vulvar and Vaginal Cancer

Ginsberg Lung Cancer

Grossbard Malignant Lymphomas

Ozols Ovarian Cancer

Pollock Soft Tissue Sarcomas

Posner, Vokes, Weichselbaum Esophagus, Stomach and Small Bowel Cancer

Prados Brain Cancer

Raghavan Germ Cell Tumors

Rice, Taylor Endometrial and Uterine Cancer

Shah Head and Neck Cancer

Shipley Urothelial Cancer

Silverman Oral Cancer

Sober, Haluska Skin Cancer

Wiernik Adult Leukemias

Willett Colon and Rectal Cancer

Winchester, Winchester Breast Cancer

Yasko Bone Cancer



Breast Cancer



Editors

GLENN D. STEELE JR, MD

University of Chicago

THEODORE L. PHILLIPS, MD

University of California

BRUCE A. CHABNER, MD

Harvard Medical School

Managing Editor

TED S. GANSLER, MD, MBA

Director of Health Content, American Cancer Society



Breast CancerDavid J. Winchester, MD, FACSAssociate Attending, Department of Surgery

Evanston Northwestern HealthcareEvanston, Illinois

Assistant Professor of SurgeryNorthwestern University Medical School

Chicago, Illinois

David P. Winchester, MD, FACSChairman, Department of SurgeryEvanston Northwestern Healthcare

Evanston, IllinoisProfessor of Surgery

Northwestern University Medical SchoolChicago, Illinois

2000

B.C. Decker Inc.

Hamilton • London

JOEL R. BERNSTEIN, MDDepartment of RadiologyEvanston Northwestern HealthcareEvanston, IllinoisScreening and Diagnostic Imaging

WILLIAM D. BLOOMER, MDDepartment of RadiologyEvanston Northwestern HealthcareEvanston, IllinoisBreast Cancer and Radiation Therapy

DAVID R. BRENIN, MDColumbia-Presbyterian Medical CenterNew York, New YorkUnusual Breast Pathology

WENDY R. BREWSTER, MDDepartment of Obstetrics and GynecologyUniversity of CaliforniaIrvine, CaliforniaEstrogen Replacement Therapy for Breast Cancer Survivors

MASSIMO CRISTOFANILLI, MDDepartment of Breast Medical OncologyMD Anderson Cancer CenterHouston, TexasBreast Cancer Risk and Management:Chemoprevention, Surgery, and Surveillance

PHILIP J. DISAIA, MDDepartment of Obstetrics and GynecologyUniversity of CaliforniaIrvine, CaliforniaEstrogen Replacement Therapy for Breast CancerSurvivors

WILLIAM L. DONEGAN, MDDepartment of Surgery, Sinai Samaritan Medical CenterMilwaukee, WisconsinCarcinoma of the Breast in Men

GEOFFREY C. FENNER, MDEvanston Northwestern HealthcareEvanston, IllinoisBreast Reconstruction

RICHARD E. FINE, MDThe Breast CenterMarietta, GeorgiaImage-Directed Breast Biopsy

ROBERT A. GOLDSCHMIDT, MDDepartment of PathologyEvanston Northwestern HealthcareEvanston, IllinoisHistopathology of Malignant Breast Disease

HANINA HIBSHOOSH, MDColumbia University, College of Physicians andSurgeonsColumbia-Presbyterian Medical CenterNew York, New YorkUnusual Breast Pathology

GABRIEL N. HORTOBAGYI, MDDepartment of Breast Medical OncologyMD Anderson Cancer CenterHouston, TexasBreast Cancer Risk and Management:Chemoprevention, Surgery, and Surveillance

JAN M. JESKE, MDDepartment of RadiologyEvanston Northwestern HealthcareEvanston, IllinoisScreening and Diagnostic Imaging

JANARDAN D. KHANDEKAR, MDDepartment of MedicineEvanston Northwestern HealthcareEvanston, IllinoisSurveillance of the Breast Cancer Patient

Contributors

Contributors v

DAVID W. KINNE, MDColumbia-Presbyterian Medical CenterNew York, New YorkUnusual Breast Pathology

MICHAEL A. LACOMBE, MDDepartment of RadiologyEvanston Northwestern HealthcareEvanston, IllinoisBreast Cancer and Radiation Therapy

LAURIE H. LEE, PA-CDivision of General SurgeryEvanston Northwestern HealthcareEvanston, IllinoisSurgical Management of Ductal Carcinoma In Situ

GERSHON Y. LOCKER, MDDivision of HematologyEvanston Northwestern HealthcareEvanston, IllinoisAdjuvant Systemic Therapy of Early Breast Cancer

HENRY T. LYNCH, MDCreighton UniversityOmaha, NebraskaGenetics, Natural History, and DNA-BasedGenetic Counseling in Hereditary Breast Cancer

JANE F. LYNCH, BSNCreighton UniversityOmaha, NebraskaGenetics, Natural History, and DNA-BasedGenetic Counseling in Hereditary Breast Cancer

DOUGLAS E. MERKEL, MDDivision of Hematology-OncologyEvanston Northwestern HealthcareEvanston, IllinoisTreatment of Metastatic Breast Cancer

DAN H. MOORE II, PHDGeraldine Brush Cancer Research Institute Pacific Medical Center Department of Epidemiology and BiostatisticsUniversity of CaliforniaSan Francisco, CaliforniaPrognostic and Predictive Markers in Breast Cancer

THOMAS A. MUSTOE, MDDivision of Plastic Surgery, Northwestern UniversityMedical SchoolChicago, IllinoisBreast Reconstruction

LISA NEWMAN, MD, FACSDepartment of Surgical OncologyUniversity of Texas, MD Anderson Cancer CenterHouston, TexasBreast Cancer Risk and Management:Chemoprevention, Surgery, and Surveillance

PHILIP N. REDLICH, MD, PHDDepartment of SurgeryMedical College of WisconsinMilwaukee, WisconsinCarcinoma of the Breast in Men

STEPHEN F. SENER, MDDivision of General SurgeryEvanston Northwestern HealthcareEvanston, IllinoisSurgical Management of Ductal Carcinoma In Situ

S. EVA SINGLETARY, MD, FACSMD Anderson Cancer CenterHouston, TexasLocally Advanced Breast Cancer

MARGARET A. STULL, MDDepartment of RadiologyUniversity of IllinoisChicago, IllinoisScreening and Diagnostic Imaging

ANN D. THOR, MDDepartment of PathologyEvanston Northwestern HealthcareEvanston, IllinoisPrognostic and Predictive Markers in BreastCancer

DAVID J. WINCHESTER, MD, FACSDepartment of SurgeryEvanston Northwestern HealthcareEvanston, Illinois Northwestern University Medical SchoolChicago, IllinoisEvaluation and Surgical Management of Stage Iand II Breast Cancer

DAVID P. WINCHESTER, MD, FACSDepartment of SurgeryEvanston Northwestern HealthcareEvanston, IllinoisNorthwestern University Medical SchoolChicago, Illinois

B.C. Decker Inc.4 Hughson Street SouthP.O. Box 620, L.C.D. 1Hamilton, Ontario L8N 3K7Tel: 905-522-7017; 1-800-568-7281Fax: 905-522-7839E-mail: [email protected]: http://www.bcdecker.com

© 2000 American Cancer Society

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without priorwritten permission from the publisher.

00 01 / PC / 9 8 7 6 5 4 3 2 1

ISBN 1–55009–112–3Printed in Canada

Sales and Distribution

United States U.K., Europe, Scandinavia, Middle East South AmericaB.C. Decker Inc. Blackwell Science Ltd. Ernesto Reichmann, DistribuidoraP.O. Box 785 Osney Mead de Livros Ltda.Lewiston, NY U.S.A. 14092-0785 Oxford OX2 0EL Rua Coronel MarquesTel: 905-522-7017/1-800-568-7281 United Kingdom 335-Tatuape, 03440-000Fax: 905-522-7839 Tel: 44-1865-206206 Sao Paulo-SP-Brazile-mail: [email protected] Fax: 44-1865-721205 Tel/Fax: 011-218-2122website: http://www.bcdecker.com e-mail: [email protected]

Canada Australia Foreign RightsB.C. Decker Inc. Blackwell Science Asia Pty, Ltd. John Scott & Co.4 Hughson Street South 54 University Street International Publishers’ AgencyP.O. Box 620, L.C.D. 1 Carlton, Victoria 3053 P.O. Box 878Hamilton, Ontario L8N 3K7 Australia Kimberton, PA 19442Tel: 905-522-7017/1-800-568-7281 Tel: 03 9347 0300 Tel: 610-827-1640Fax: 905-522-7839 Fax: 03 9349 3016 Fax: 610-827-1671e-mail: [email protected] e-mail: [email protected]: http://www.bcdecker.com

Japan South KoreaIgaku-Shoin Ltd. Seoul Medical Scientific Books Co.Foreign Publications Department C.P.O. Box 97943-24-17 Hongo, Bunkyo-ku, Seoul 100-697Tokyo 113-8719, Japan Seoul, KoreaTel: 3 3817 5676 Tel: 82-2925-5800Fax: 3 3815 6776 Fax: 82-2927-7283e-mail: [email protected]

Notice: The authors and publisher have made every effort to ensure that the patient care recommended herein, including choice of drugsand drug dosages, is in accord with the accepted standard and practice at the time of publication. However, since research and regulationconstantly change clinical standards, the reader is urged to check the product information sheet included in the package of each drug, whichincludes recommended doses, warnings, and contraindications. This is particularly important with new or infrequently used drugs.

1 Genetics, Natural History, and DNA-Based Genetic Counseling in Hereditary Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Henry T. Lynch, MD, Jane F. Lynch, BSN

2 Breast Cancer Risk and Management: Chemoprevention, Surgery, and Surveillance . . . 19Massimo Cristofanilli, MD, Lisa Newman, MD, FACS, Gabriel N. Hortobagyi, MD

3 Screening and Diagnostic Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Jan M. Jeske, MD, Joel R. Bernstein, MD, Margaret A. Stull, MD

4 Image-Directed Breast Biopsy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Richard E. Fine, MD

5 Histopathology of Malignant Breast Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89Robert A. Goldschmidt, MD

6 Unusual Breast Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99David R. Brenin, MD, Hanina Hibshoosh, MD, David W. Kinne, MD

7 Prognostic and Predictive Markers in Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Ann D. Thor, MD, Dan H. Moore II, PhD

8 Surgical Management of Ductal Carcinoma In Situ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131Stephen F. Sener, MD, Laurie H. Lee, PA-C

9 Evaluation and Surgical Management of Stage I and II Breast Cancer. . . . . . . . . . . . . . 139David J. Winchester, MD, FACS

10 Locally Advanced Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153S. Eva Singletary, MD, FACS

11 Breast Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171Geoffrey C. Fenner, MD, Thomas A. Mustoe, MD

Contents

viii Contents

12 Adjuvant Systemic Therapy of Early Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201Gershon Y. Locker, MD

13 Breast Cancer and Radiation Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219Michael A. LaCombe, MD, William D. Bloomer, MD

14 Carcinoma of the Breast in Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239Philip N. Redlich, MD, PhD, William L. Donegan, MD

15 Estrogen Replacement Therapy for Breast Cancer Survivors . . . . . . . . . . . . . . . . . . . . . 253Wendy R. Brewster, MD, Philip J. DiSaia, MD

16 Surveillance of the Breast Cancer Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263Janardan D. Khandekar, MD

17 Treatment of Metastatic Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271Douglas E. Merkel, MD

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

The organization of this book reflects a logical,stepwise evaluation and treatment of the patientwith breast cancer. It emphasizes the importanceof early detection, but highlights a move towardrisk identification and reduction. The understand-ing of the breast cancer patient has evolved fromthe radical mastectomy for all patients to a tai-lored approach employing aggressive applica-tions of treatment modalities according to theirrespective risk reductions.

Despite shifting efforts to identify high-riskpatients and address their risk with pre-emptivestrategies, there remains a worldwide educationalchallenge to adopt early detection screening guide-lines. Although there is continuing progress inimplementing mortality reducing surveillanceguidelines as reflected by the increased prevalenceof preinvasive breast cancer, the full spectrum ofdisease remains a challenge to the medical commu-nity. The high prevalence of breast cancer continuesto drive improvements in all areas of detection,diagnostic evaluation, disease characterization,multimodality therapy, quality of life issues, and,finally, in the treatment of patients whose diseasehas extended beyond our capabilities to detect orcontain local or regional cancer.

One of the most important innovations in theunderstanding of breast cancer has been the identi-fication of genetic mutations that have allowed theopportunity to intervene with proven surgical orchemopreventive strategies for high-risk patients.Diagnostic imaging technology continues to pro-vide increased resolution and precision, resultingin an enhanced ability to preserve tissue. The sur-gical treatment of breast cancer is in the process oftaking another significant step forward, with thedevelopment of sentinel lymph node biopsy. Thedefinition of prognostic factors has helped to guideimportant adjuvant therapy decisions. Moving

beyond the regimented doctrines of an overwhelm-ing preoccupation of cancer treatment, immediatereconstruction with microvascular surgery andother techniques have provided an answer to someof the physical and psychological challenges ofbreast cancer. Recognition of other competingcauses of mortality in the breast cancer survivor hasled to a more comprehensive consideration of hor-mone replacement therapy to address quality of lifeissues and to reduce the risk of cardiovasculardisease and osteoporosis. In the patient withmetastatic disease, the introduction of novel formsof treatment, such as with Herceptin, has led to sig-nificant improvements in survival. In total, theseinnovations represent significant progress and pro-vide important directions for future interventions.

One of the biggest challenges to the clinicianhas been the recognition of improved methods ofdiagnosis and treatment and utilization of theseimprovements despite ingrained practices. Thispattern is well documented by the great variationobserved nationally in the implementation ofbreast preserving surgery and the utilization ofadjuvant treatment.

It is the goal of this book to identify signifi-cant improvements in each area of breast cancerdiagnosis and treatment, and to help acceleratethe dispersion of this knowledge to an ever-broad-ening spectrum of physicians and scientists whoare dedicated to preventing and treating one of themost common afflictions of women.

We wish to thank our distinguished authors fortheir timely and expert contributions to this effort.We also wish to thank the American Cancer Soci-ety, particularly Ted Gansler, for a helpful reviewof this book.

DJWDPW

Preface

To our parents, for their example and knowledge,and to Marilyn, Doris, Eric, Laura, Elena, andColin, who have allowed us to spend their timeworking on this project.

David J. Winchester

David P. Winchester

1

The extraordinary advances in molecular genet-ics during the past decade have establishedbeyond doubt that there is a Mendelian inheritedbasis for a subset of virtually all forms of can-cer.1 Specifically, more than 30 hereditary cancersyndromes have been shown to harbor germ-linemutations. These culprit molecular genetic fac-tors include oncogenes such as the RET proto-oncogene for the multiple endocrine neoplasiatype 2 syndromes, the mismatch repair genes(hMSH2, hMLH1) in hereditary nonpolyposiscolorectal cancer (HNPCC) of the Lynch I andLynch II syndrome variants, and tumor suppres-sor genes. Examples of the latter include APC,which predisposes to familial adenomatouspolyposis (FAP), and BRCA1 and BRCA2 muta-tions in hereditary breast cancer, the subject ofthis chapter.

In the United States in 1999, it was pro-jected that 176,300 new cases of carcinoma ofthe breast would be diagnosed, and 43,700would die from this disease.2 The currentauthors estimate that approximately 10 percent(17,600) of these newly diagnosed patients willmanifest a hereditary breast cancer (HBC) dis-order, the most common of which will be thehereditary breast-ovarian cancer (HBOC) syn-drome.3 It is the purpose of this chapter toupdate what is known about HBC, including its

1

Genetics, Natural History, and DNA-Based Genetic Counseling in Hereditary Breast CancerHENRY T. LYNCH, MDJANE F. LYNCH, BSN

genetic and phenotypic heterogeneity, naturalhistory, genetic counseling issues, and cancercontrol implications.

GENETICS

The combination of carcinoma of the breastand ovary in families, now known as theHBOC syndrome, was first reported in theearly 1970s.4–6 The molecular genetic discov-eries that confirmed beyond any doubt thehereditary basis for HBC and HBOC haveprogressed at an explosive rate, particularlyover the past decade. This avalanche of knowl-edge was heralded by the gene linkage studyof Hall and colleagues,7 which identified alocus on chromosome 17q for families withsite-specific breast cancer. Subsequently,Narod and colleagues8 reported that this samelocus was responsible for the HBOC syndrome.The culprit gene, now known as BRCA1, wasthen cloned.9 More recently, a second breastcancer susceptibility locus on chromosome13q, known as BRCA2, was identified by link-age analysis10 and subsequently cloned.11

Approximately 45 percent of all hereditarybreast cancer-prone families, including thosecharacterized as HBOC, owe this condition tomutations of the BRCA1 gene;12 a slightly lower

2 BREAST CANCER

percentage is due to BRCA2 mutations. Initialstudies estimated that carriers of the BRCA1germ-line mutation harbor a lifetime risk forbreast cancer of about 85 percent12,13 and a riskfor ovarian cancer that ranges between 40 and66 percent.12,13 Carriers of the BRCA2 muta-tion, on the other hand, have a lifetime risk ofbreast cancer of about 85 percent, but their riskfor ovarian cancer is somewhat lower (10 to20 %).12,13 Male BRCA2 mutation carriers havean approximate 7 percent lifetime risk forbreast cancer. Other cancers occurring in BRCA2mutation carriers include carcinoma of the pan-creas, head and neck, and intraocular malignantmelanoma. Males who are harbingers of germ-line mutations in BRCA1/BRCA2 will have atwo- to three-fold increased lifetime risk forprostate cancer.

These initial breast and ovarian cancer genepenetrance risks were based upon publicationsof highly extended pedigrees selected becauseof their profound familial cancer aggregations;they are thereby biased in the direction of can-cer excess. In contrast, recent observations ofbreast and ovarian cancer occurrence in theAshkenazi Jewish founder mutations (185delAGand 5382insC mutations on the BRCA1 gene,and 6174delT on the BRCA2 gene) indicate thatthe lifetime risk for breast cancer is only 56percent, and that of ovarian cancer 16 percent.14

Claus and colleagues15 examined the familyhistory of carcinoma of the breast and ovary in adata set involving 4,730 patients with breast can-cer and 4,688 controls who were enrolled in theCancer and Steroid Hormone Study. Attentionwas given to the association between family his-tory of carcinoma of the breast and/or ovary andbreast cancer risk when controlling for the car-rier status of BRCA1 and BRCA2 mutations. Thequestion examined pertained to whether the fam-ily history of carcinoma of the breast remained apredictive risk factor once the carrier status forBRCA1 and/or BRCA2 was given consideration.Findings disclosed that among those women“…with a moderate family history of breast can-cer, that is, predicted noncarriers of BRCA1

and/or BRCA2 mutations, family historyremains a factor in predicting breast cancer risk.In families with breast and ovarian cancers, theaggregation of these two cancers appears to beexplained by BRCA1/BRCA2 mutation-carrierprobability.” This study clearly enunciates theneed for obtaining a well-orchestrated cancerfamily history for the assessment of breast/ovar-ian cancer risk.

Previously, the best prediction of a patient’slifetime breast cancer risk was 50 percent. Thisestimate was based upon the patient’s positionin the pedigree, namely, having one or morefirst-degree relatives with a syndrome cancerin the direct genetic lineage of an HBC orHBOC family. The identification of the muta-tions for BRCA19 and BRCA211 now enablesphysicians and genetic counselors to predict apatient’s lifetime risk for carcinoma of thebreast and ovary, in context with the penetranceof these genes.

Cancer Family History and Mutation Search

The search for a germ-line mutation should beperformed only on families with substantialevidence of a hereditary cancer syndrome.Therefore, to establish a hereditary breast can-cer syndrome diagnosis, a detailed collection ofa patient’s cancer family history, with as muchpathologic corroboration as possible, is manda-tory. The family history may potentially consti-tute the most cost-beneficial component of apatient’s medical workup; its collection andevaluation in the typical clinical setting, how-ever, remains notoriously neglected.16,17 Thisproblem was well documented by David andSteiner-Grossman17 through a survey of 76acute care, non-psychiatric hospitals in NewYork City. Only four of the 64 reporting hospi-tals indicated that family history informationwas reported in their medical records. Suchserious omissions must be resolved in order toenhance cancer control. Otherwise, the oppor-tunity to search for germ-line cancer prone

Genetics, Natural History, and DNA-Based Genetic Counseling in Hereditary Breast Cancer 3

mutations may be lost. This is a pity since,when a cancer-causing mutation is identified,this information can, in concert with geneticcounseling, be used effectively to benefit thepatient and family members.

ASSESSMENT OF BREAST CANCER-PRONE FAMILIES

Building the case for hereditary cancer is fre-quently based upon the cardinal clinical fea-tures of hereditary cancer (namely, early age ofcancer onset, pattern of multiple primary can-cers [such as breast and ovarian cancer], verti-cal transmission of cancer, and increased num-ber of cancer occurrences) (Table 1–1). It isvirtually axiomatic that the larger the breastcancer-prone family, the greater the number ofexpected carcinomas of the breast or ovary. Onecan be more confident of a likely hereditary eti-ology for breast cancer if there is evidence forearlier age of onset of cancer, especially whenthere is familial clustering of these cancers,particularly ovarian cancer, among primary andsecondary relatives (see Table 1–1). In such asetting, there is an increased probability that agerm-line mutation (BRCA1, BRCA2) will befound. On the other hand, when dealing withfamilies that are small, there may be a limitednumber of patients with cancer, a deficit of

females, or the few cancers that are occurringmay be in the paternal lineage. The overalleffect is that it becomes exceedingly difficult topredict whether such a small family should be acandidate for searching for a mutation inBRCA1 or BRCA2.

HETEROGENEITY AND HEREDITARYBREAST CANCER

Virtually all forms of hereditary cancer showsignificant genetic and phenotypic heterogene-ity. For example, breast cancer occurs in signif-icant excess in disorders associated with extra-breast cancer sites, such as Li-Fraumenisyndrome (Figure 1–1), Bloom’s syndrome,Cowden’s disease, ataxia-telangiectasia, thebreast-gastrointestinal tract cancer syndrome,extraordinarily early-onset breast cancer (Fig-ure 1–2), and the HBOC syndrome (Figure1–3). Undoubtedly, other tumor combinationsand/or hereditary syndromes which will qualifyas hereditary breast cancer are yet to be identi-fied. Space does not allow a discussion of eachof these breast cancer-associated disorders (formore detail see Lynch and colleagues18).

Clearly, it is no longer appropriate to char-acterize hereditary breast cancer as a genericterm. Rather, one must be more precise anddenote the particular breast cancer-associated

Table 1–1. ESTIMATED PROBABILITY OF BRCA1 MUTATION BASED ON FAMILY HISTORY

Family History Probability of BRCA1 Mutation (%)

Single affected personBreast cancer at < 30 years of age 12Breast cancer at < 40 years of age 6Breast cancer at 40–49 years of age 3Ovarian cancer at < 50 years of age 7

Sister pairsBoth with breast cancer at < 40 years of age 37Both with breast cancer at 40–49 years of age 20Breast cancer at < 50 years of age, ovarian cancer at < 50 years of age 46Both with ovarian cancer at < 50 years of age 61

FamiliesBreast cancer only, three or more cases at < 50 years of age 40Two or more breast cancers and one or more ovarian cancers 82Two or more breast cancers and two or more ovarian cancers 91

Reprinted with permission of author (Barbara Weber, MD) and publisher from Weber B. Breast cancer susceptibility genes: current challengesand future promises. Ann Intern Med 1996;124:1088–90.

4 BREAST CANCER

Fig

ure

1–1.

Upd

ated

ped

igre

e of

a fa

mily

with

sar

com

a, b

reas

t can

cer

and

brai

n tu

mor

s, lu

ng a

nd la

ryng

eal c

ance

r, an

d ad

reno

cort

ical

car

cino

ma

(SB

LA s

yndr

ome,

als

o kn

own

as L

i-Fr

aum

eni s

yndr

ome)

.Rep

rinte

d w

ith p

erm

issi

on f

rom

Lyn

ch e

t al

.Gen

etic

and

pat

holo

gic

findi

ngs

in a

kin

dred

with

her

edita

ry s

arco

ma,

bre

ast

canc

er,

brai

n tu

mor

s, le

ukem

ia,

lung

,la

ryge

al,

and

adre

nal c

ortic

al c

arci

nom

a.C

ance

r 19

78;4

1:20

55–6

4.

Indi

vidu

al n

umbe

rU

naffe

cted

Cur

rent

age

Can

cer

by p

atho

logy

,ag

e at

dia

gnos

isC

urre

nt a

geC

ance

r by

fam

ily h

isto

ry,

age

of d

eath

Mul

tiple

prim

ary

canc

ers

by p

atho

logy

Can

cer

by d

eath

cer

tific

ate

orm

edic

al r

ecor

dsN

umbe

r of

una

ffect

ed p

roge

ny(B

oth

sexe

s)

Pro

band

Twin

s

Can

cer

Site

s

Ad

Adr

enal

cor

tical

Br

Bre

ast

Bt

Bra

in t

umor

Co

Col

onC

SU

Can

cer

site

unk

now

nLe

uLe

ukem

iaLu

Lung

Ly

mLy

mph

oma

Lyx

Lary

nxN

bN

euro

blas

tom

aP

anP

ancr

eas

Pro

Pro

stat

eS

arS

arco

ma


syndrome relating to a particular patient/family.Such syndrome identification is important notonly for molecular genetic assessment but,moreover, for targeted surveillance and man-agement purposes.

GENOTYPE-PHENOTYPE DIFFERENCES

More than 200 different BRCA1 germ-linemutations have been identified in HBOC fami-lies. Certain types of these mutations may giverise to differing patterns of cancer occurrence.For example, Gayther and colleagues19 suggestthat the position of the BRCA1 mutation has asignificant influence on the ratio of breast toovarian cancer in HBOC kindreds. Specifically,they reported that mutations in the 3' third of thegene are associated with a lower proportion ofovarian cancer. However, these findings must bereviewed cautiously. Serova and colleagues20

were unable to confirm these findings. How-ever, the findings of the latter did suggest thatthe risk of ovarian cancer is greater in familieswith mutations associated with reduced RNAlevels. In the case of BRCA2, Gayther and col-leagues21 found that “... truncating mutations infamilies with the highest risk of ovarian cancerrelative to breast cancer are clustered in a regionof approximately 3.3 kb in exon 11 (p = .0004).”Further research in this area may establish linksbetween specific mutations and specific cancerrisk that will be extremely useful for geneticcounseling. Until confirmation of these geno-type-phenotype findings is more firmly estab-lished, however, the current authors prefer towithhold this information in the genetic coun-seling setting.

Most of the hereditary cases will harbor aBRCA1 or BRCA2 germ-line mutation. However,one should expect (albeit rarely) to encounterfamilies where both BRCA1 and BRCA2 muta-tions are segregating, considering the high preva-lence of carcinoma of the breast and ovary in thegeneral population, coupled with the fact thatapproximately 5 to 10 percent of the total breast

cancer burden will be hereditary. Interestingly,Ramus and colleagues22 reported a patient from aHungarian family who manifested both breastand ovarian cancer and was found to have trun-cating mutations in both the BRCA1 and BRCA2genes. This patient “... carried the 185delAGmutation in BRCA1 as well as the 6174delTmutation in BRCA2. Both of these mutations arecommon in Ashkenazi Jewish breast cancerpatients.”23–25 Recently, Liede and colleagues26

identified an Ashkenazi Jewish kindred withthree mutations, namely BRCA1 185delAG,BRCA1 5382insC, and BRCA2 6174delT. Eachfounder mutation has been shown to have a fre-quency of approximately 1 percent in the Ashke-nazi population.27–29

Figure 1–2. Pedigree of a family showing extremely early ageof onset of hereditary breast cancer. Reprinted with permis-sion from Lynch et al. Extremely early onset hereditary breastcancer (HBC): surveillance/management implications. NebrMed J 1988; 73:97–100.

Individual numberUnaffectedCurrent age

Cancer by pathology,age at diagnosisCurrent ageMultiple primary cancers by pathologyage at deathCancer by death certificate ormedical recordsNumber of unaffected progeny(Both sexes)Proband

Cancer Sites

Br BreastSt Stomach

6 BREAST CANCER

Fig

ure

1–3.

Upd

ated

ped

igre

e of

a l

arge

her

edita

ry b

reas

t-ov

aria

n ca

ncer

fam

ily.

Ada

pted

with

per

mis

sion

fro

m L

ynch

et

al.

Man

agem

ent

of f

amili

al b

reas

t ca

ncer

, pa

rts

I an

d II.

Arc

h S

urg

1978

;113

:106

0–71

.

Indi

vidu

al n

umbe

rU

naffe

cted

Cur

rent

age

Can

cer

by p

atho

logy

,ag

e at

dia

gnos

isC

urre

nt a

geC

ance

r by

fam

ily h

isto

ry,

age

of d

eath

Mul

tiple

prim

ary

canc

ers

by

med

ical

rec

ords

or

deat

h ce

rtifi

cate

Mul

tiple

prim

ary

canc

ers

by p

atho

logy

Can

cer

by d

eath

cer

tific

ate

orm

edic

al r

ecor

ds

Can

cer

Site

sB

rB

reas

tC

oC

olon

CS

UC

ance

r si

te u

nkno

wn

LuLu

ng

Mm

elM

alig

nant

mel

anom

aO

men

Om

entu

mO

vO

varia

nP

roP

rost

ate

Psu

Prim

ary

site

unk

now

nS

tS

tom

ach

Twin

sN

umbe

r of

una

ffect

ed p

roge

ny(B

oth

sexe

s)P

roba

nd


PATHOLOGY OF HEREDITARYBREAST/OVARIAN CANCER IN

CONCERT WITH BRCA1 OR BRCA2 MUTATIONS

Pathology studies have shown differencesbetween BRCA1- and BRCA2-related breastcancers when compared with sporadic controls.Specifically, Marcus and colleagues30–32 haveshown that BRCA1 HBC has a highly distinctivepathology phenotype, consisting of an increasednumber of aneuploid cancers, more medullarycarcinomas, and high proliferation rates as mea-sured by DNA flow cytometry and mitoticgrade, and lesser ductal carcinoma in situ(DCIS) than in nonfamilial cases. In alluding tohigh S-phase fraction in HBC and attributing itto the BRCA1-linked subset, it was suggestedthat the mutation resulted in enhanced cellularproliferation.33 This prediction was borne out bythe demonstration of the antiproliferative effectof BRCA1 mRNA protein in vitro and invivo34–36 after the gene was cloned. The currentauthors have proposed a model for the BRCA1pathophenotype that considers the tumors to bein an advanced state of genetic evolution.30

In contrast, “other” HBCs (cases from HBCfamilies with no BRCA1 mutations, no 17qlinkage, and a paucity of ovarian cancer affect-eds, or with BRCA2 mutations or 13q linkage)appear to lack the systematic high grade, aneu-ploidy, and high proliferation of BRCA1 HBCs,and they are not deficient in in situ carci-noma.30,31 This “other” group also has moreinvasive lobular, tubular, tubulolobular, andcribiform special type carcinomas, which wehave designated as “tubular-lobular group”(TLG). Indeed, the excess of TLG and “no spe-cial type” (NST) invasive carcinomas with TLG“features” (10 to 50% tumor composition) par-allels a trend for more lobular neoplasia (lobu-lar carcinoma in situ and atypical lobularhyperplasia) in “other” HBC. These featuresare present in the subset of mutation-confirmedBRCA2 HBC cases in the “other” HBC group,32

which suggests that TLG carcinomas and lobu-

lar neoplasia are signatures of the BRCA2 HBCphenotype.32 In contrast, BRCA1 HBC casesmanifest a deficit of TLG carcinomas and lob-ular neoplasia.32 Armes and colleagues37 con-firm an excess of TLG carcinoma (in theircases, pleomorphic lobular carcinomas) inBRCA2 HBC in a population-based study ofBRCA2 cases that were not specificallyrecruited from large HBOC families.

The pathophenotype of BRCA2 HBC maybe more heterogeneous than BRCA1 HBCwhen the amount of high-grade carcinoma inthe syndrome is considered. There have beenreports of BRCA2 families with predominantlyhigh-grade carcinomas.38,39 However, we havenot seen high-grade predominance in the fourBRCA2 families we have studied nor as theaverage phenotype of the “other” HBC group inwhich most Creighton BRCA2 families wouldreside.32 Similarly, the Breast Cancer LinkageConsortium has not observed unusually highgrades in its BRCA2 family series.40 The highergrades reported from Iceland39 and, to a lesserextent, from the Linkage Consortium,40,41 maywell be associated with a site on the BRCA2gene, in these cases the 999del5 mutation.37

Lakhani and colleagues41 confirmed manyof the original observations of Marcus and col-leagues.32,33,42 Specifically, they showed that“Cancers associated with BRCA1 mutationsexhibited higher mitotic counts (p = .001), agreater proportion of the tumor with a continu-ous pushing margin (p < .0001), and more lym-phocytic infiltration (p = .002) than sporadic(ie, control) cancers. Cancers associated withBRCA2 mutations exhibited a higher score fortubule formation (fewer tubules) (p = .0002), ahigher proportion of the tumor perimeter with acontinuous pushing margin (p < .0001), and alower mitotic count (p = .003) than control can-cers.” These authors concluded that thishistopathology information may improve theclassification of breast cancers in those show-ing a positive family history for this disease.Specifically, employing multifactorial analysisresults from their previous estimates, they

8 BREAST CANCER

found that 7.5 percent of individuals withbreast cancer in Britain who had been diag-nosed at between the ages of 20 and 29 yearsharbor a BRCA1 mutation.43 Further, “Assum-ing that the odds ratios from our analysis areindependent of age, only about 2 percent ofcase subjects in this age group in whom themitotic count is below 5 per 10 hpf, withoutcontinuous pushing margins, and in whomthere is no lymphocytic infiltrate would beexpected to carry a BRCA1 mutation. By con-trast, about 45 percent of case subjects in the20- to 29-year-old group with 20 to 39 mitosesper 10 hpf, continuous pushing margins occu-pying more than 75 percent of the tumorperimeter, and a prominent lymphocytic infil-trate would be expected to be BRCA1 carriers.The corresponding proportions based onmitotic count would be 4 and 16 percent.”

RADIATION EFFECTS AND BRCA1AND BRCA2 MUTATIONS

Questions have been raised relevant to potentialcarcinogenic risk of radiation exposure forwomen who harbor the BRCA1 or BRCA2mutations. Scully and colleagues44 raised thepossibility that there may be an interactionbetween BRCA1 and BRCA2 gene productswith respect to proteins involved in the repair ofradiation-induced DNA errors. However, thisissue remains controversial due to lack of con-firmation of this risk in the past by other inves-tigators.45 Nevertheless, recent evidence hasindicated that both BRCA1 and BRCA2 areassociated with defective repair of radiation-induced DNA damage.44,46

In a study by Chabner and colleagues45 of201 patients, 29 of whom had positive familyhistories of breast cancer (a mother or sisterpreviously diagnosed before the age of 50years, or ovarian cancer at any age) and whohad undergone breast-conserving surgery andradiation therapy for early-stage breast cancer,there was no evidence associated with a higherrate of “...local recurrence, distant failure, or

second (nonbreast) cancers in young womenwith a family history (FH) suggestive of inher-ited breast cancer susceptibility compared withyoung women without an FH.” As expected, thepatients with a positive FH showed anincreased risk of contralateral breast cancer.This matter of contralateral breast cancer mustbe given careful consideration when counselingwomen with positive family histories who areconsidering the option of breast conservingsurgery and radiation therapy versus modifiedradical mastectomy. Given these findings,Chabner and colleagues45 conclude that “...young women can be offered conservativesurgery and radiation therapy as a reasonableoption at breast cancer diagnosis.” However,the investigators appropriately call attention tothe limitations in their study: a relatively shortfollow-up time, small size of their cohort, andthe absence of specific genetic findings ontheir patients, including an absence of BRCA1or BRCA2 mutation findings.

OVARIAN CARCINOMA

Any discussion of the genetics of carcinoma ofthe breast must include ovarian cancer. This dis-ease will affect approximately 1 percent ofwomen in the United States during their lifetime,where it accounts for about 14,500 deaths annu-ally2 with a five-year survival rate of < 30 per-cent. The biological mechanism of transformingbenign cells to carcinoma remains elusive,although it likely involves a multistep processrequiring an accumulation of genetic lesionsinvolving different gene classes. As mentioned,the etiologic association of ovarian cancer withbreast cancer was first reported in the early1970s in a series of breast cancer-prone pedi-grees;4–6,8 both BRCA1 and BRCA2 mutationswere subsequently found to predispose to ovar-ian cancer. In BRCA1, the lifetime risk for ovar-ian cancer is in the range of approximately 50percent, while in BRCA2 the lifetime ovariancancer risk is about 20 percent. Ovarian carci-noma is also an integral lesion in Lynch syn-


drome II.47 Lynch and colleagues48 have pro-vided an extensive review of the genetics ofovarian cancer.

GENETIC COUNSELING

We believe that genetic counseling should bemandatory for patients who are at high risk forbreast cancer and are contemplating DNA test-ing in the search for specific germ-line muta-tions. Counseling should take place prior to col-lection of DNA and at the time of disclosure ofresults. The ideal individual for initial gene test-ing in a family where a hereditary form of breastcancer is considered likely would be one whohas had a syndrome cancer, particularly if diag-nosed at an early age, and who is in the directline of descent of syndrome cancer expression.The clinician’s task during the genetic counsel-ing process is to help the patient answer crucialquestions which may arise during the genetictesting process. Importantly, patients need todecide whether to be tested for the presence of agerm-line mutation in BRCA1 or BRCA2, oncethe facts are understood. They should be awareof the potential for fear, anxiety, apprehension,intrafamily strife, as well as insurance/ employ-ment discrimination. Finally, they need to knowthe best type of medical management for them,based upon the test result.

Because there are a limited number of certi-fied genetic counselors who have sufficientknowledge of oncology to effectively counselcancer-prone families, the American Society ofClinical Oncology (ASCO) has recommendedthat, whenever possible, physicians should per-form genetic counseling.49 In its published posi-tion49 on genetic testing for cancer susceptibility,ASCO recognizes that the clinical oncologist’srole should be to document the family history ofcancer, provide counseling with respect to apatient’s inordinate lifetime cancer risk, and pro-vide options for prevention and early detectionto those families for whom genetic testing mayaid in the genetic counseling process. Informedconsent by the patient is considered to be an inte-

gral part of the process of genetic predispositiontesting, on either a clinical or research basis. Pre-disposition testing should be performed onpatients for whom there is a strong family his-tory of cancer that is consonant with a likelyhereditary etiology, where the results can be ade-quately interpreted, and where there is a poten-tial to aid in the medical management of thepatient and/or family members.

The American Society of Clinical Oncologyalso recognizes the need to strengthen regulatoryauthority over laboratories that provide cancerpredisposition tests that will ultimately be usedin making informed clinical decisions. In theinterest of protecting patients and their families,ASCO endorses the adoption of legislation toprohibit discrimination by insurance companiesor employers based on an individual’s suscepti-bility to cancer. Finally, ASCO and the AmericanCancer Society prudently endorses the need forall individuals at hereditary risk for cancer tohave, in concert with medical care, appropriategenetic counseling which should be covered bypublic and private third-party payers.

Figure 1–4 is an algorithm depicting theprocess used by the Creighton cancer geneticresearch team to ascertain, test, and counselHBC/HBOC-prone families. Detailed informa-tion about the natural history of HBC/HBOCwas provided and the pros and cons of DNAtesting were discussed with more than 2,000members of 29 large families with BRCA1mutations and 8 families with BRCA2 mutations(Table 1–2).50 The current authors found thatthe perceived risk for cancer was associatedwith the individual’s position in the pedigree.There was a significant tendency to overesti-mate risk rather than underestimate it (p < .001)by a chi-square test from Fleiss and colleagues51

(Table 1–3). Fifty-seven family members who had pro-

vided blood samples several years ago declinedthe opportunity to receive the results of theirDNA testing. Thirty of the 57 responded to ananonymous questionnaire by giving one ormore reasons for declining. Their responses

10 BREAST CANCER

varied, but fear of insurance discrimination wascited by 37 percent of this group, and fear of apositive result was cited by 20 percent.

Of those choosing to learn their mutationstatus, the majority identified their children asa primary reason for being tested (Table 1–4).

Figure 1–4. Algorithm for education and assessment of BRCA1 and BRCA2 families. Reprinted withpermission from Lynch et al. Cancer. In press.


Although most of these individuals freely choseto be tested, occasionally they reported pressurewithin the family either for or against theirbeing tested.

Overall, 167 of 403 queried family members(41%) cited health management (especiallysurveillance for early detection and prophylac-tic surgery) as a reason for seeking genetic test-ing, making it the second most common reasonreported. In the subgroup where mutation sta-tus information would have the greatest impacton recommendations for surveillance (womenunder the age of 40), 72 of 118 queried, or 61percent, cited this reason (other subgroup dataare not shown).

Patients’ self reports indicated that currentlythe majority were having breast cancer screeningtests at the recommended frequency (Table 1–5).Nevertheless, the majority said that they wouldconsider increasing surveillance if the mutationtest showed they were mutation carriers. Only aslim majority (13 of 25 overall) would considerdecreasing surveillance if the test showed theywere not mutation carriers. Sixty-two percent(59 of 95) of all women over the age of 25 yearssaid they would consider prophylactic mastec-tomy if the test showed they were mutation car-riers. Patient self reports indicated that a minor-ity of these women are having CA-125 and

transvaginal ovarian ultrasound as often as rec-ommended (Table 1–6). Most said they wouldconsider increasing surveillance and/or havingprophylactic oophorectomy if the test showedthem to be mutation carriers. Note that all dataon post-test management reflect what patientssay they will consider in the decision-makingprocess, not necessarily what they plan to do.

The emotional responses to disclosure ofgerm-line mutation results cannot always beanticipated. Our data show that the majority ofpatients who are negative express relief. How-ever, some may experience disbelief or survivorguilt. Those who are told they do have thegerm-line mutation express a variety of reac-

Table 1–2. DEMOGRAPHIC CHARACTERISTICS OF STUDY SUBJECTS IN 29 BRCA1 FAMILIES AND 8 BRCA2 FAMILIES

BRCA1 BRCA2

Total number counseled and given gene status 339 85

Mutation positive 137 (40) 42 (49)Mutation negative 198 (58) 43 (51)Ambiguous 4 ( 1) 0

Male 89 (26) 20 (24)Female 250 (74) 65 (76)

Age Group< 25 11 ( 4) 4 ( 6)25-50 89 (36) 26 (40)> 50 150 (60) 35 (54)

Cancer statusBreast and/or ovarian cancer affected 59 (24) 16 (25)

Reproduced with permission of author and publisher from Lynch et al. An update on DNA-based BRCA1/BRCA2 genetic counseling in hereditarybreast cancer. Cancer Genet Cytogenet 1999;109:91–8.

Table 1–3. ESTIMATE OF PERCEIVED RISK OF HARBORING A BRCA1/BRCA2 MUTATION

Perceived Risk of Mutation Carriage (n (%))

Pedigree Risk* < 50% 50% > 50%

< 25% 27 (38.0%) 27 (38.0%) 17 (23.9%)50% 44 (17.7%) 74 (29.8%) 130 (52.4%)Affected/obligate 3 (3.8%) 5 (6.3%) 71 (89.9%)

gene carrier

*These individuals tended to overestimate their risk more often thanto underestimate it (p < .001) (chi square test51).Reproduced with permission of author and publisher from Lynch et al.An update on DNA-based BRCA1/BRCA2 genetic counseling inhereditary breast cancer. Cancer Genet Cytogenet 1999;109:91–8.

12 BREAST CANCER

tions, including acceptance because the resultsare not a surprise to them, relief of anxiety withthe removal of uncertainty about their geneticrisk status, a positive attitude in terms of pre-vention, feelings of sadness, or even anger. Thegenetic counselor must be responsive to all ofthese emotions.

At a baseline interview in a study by Ler-man and colleagues,52 breast-ovarian cancer-related stress symptoms were predictive of theonset of depressive symptoms in family mem-bers who were invited but declined testing.“Among persons who reported high baseline

levels of stress, depression rates in declinersincreased from 26 percent at baseline to 47 per-cent at one-month follow-up; depression ratesin noncarriers decreased and in carriers showedno change (odds ratio [OR] for decliners versusnoncarriers = 8.0; 95% confidence interval [CI]1.9 to 3.5; p = .0004). These significant differ-ences in depression rates were still evident atthe six-month follow-up evaluation (p = .04).”It was concluded that in BRCA1/BRCA2-linkedfamilies, individuals showing high levels ofcancer-related stress who ultimately declinedgenetic testing appeared to be at increased riskfor depression. It was reasoned that they couldderive benefit through education and counsel-ing even though they might ultimately declineto be tested; these are the individuals whorequire monitoring for the potential occurrenceof adverse psychological effects.

An important genetic counseling question is,Do patients who received information about theirgenetic risk status, including the presence of aBRCA1/BRCA2 germ-line mutation, heed sur-veillance and management recommendations?These recommendations include the need forincreased frequency of mammography, breastself-examination, and physician breast examina-tion. Recommendations for ovarian screening

Table 1–4. REASONS FOR SEEKING RISK ASSESSMENT IN COUNSELED MEMBERS OF 29

BRCA1 FAMILIES AND 8 BRCA2 FAMILIES

Counseled BRCA1 Counseled BRCA2Individuals Individuals(n = 319) (n = 84)

Top three reasonsfor seeking risk assessment

Children 170 (53) 47 (56)Surveillance/ 125 (39) 42 (50)

prophylaxisCuriosity 94 (29) 24 (29)

Reproduced with permission of author and publisher from Lynch et al.An update on DNA-based BRCA1/BRCA2 genetic counseling inhereditary breast cancer. Cancer Genet Cytogenet 1999;109:91-8.

Table 1–5. SURVEILLANCE PRACTICES AND ATTITUDES TOWARD PROPHYLACTIC MASTECTOMIES PRIOR TO GENETIC TEST RESULT DISCLOSURE*

BRCA1 BRCA2

Bilateral mastectomies prior to the counseling sessionBreast cancer-affected 39/250 (16) 8/65 (12)Breast cancer-free (Prophylactic) 23/250 ( 9) 3/65 ( 5)

Current surveillance practices‡

Mammography Ever 145/153 (95) 42/46 (91)As recommended 92/118 (78) 25/36 (69)

Physician BE Ever 105/107 (98) 39/40 (97)As recommended 84/90 (93) 37/38 (97)

Self BE Ever 102/116 (88) 38/42 (90)As recommended 54/88 (61) 27/37 (73)

If carrier, will consider prophylactic mastectomy† 51/77 (66) 8/18 (44)If carrier, will consider increasing surveillance‡ 62/65 (95) 24/25 (96)If noncarrier, will consider decreasing surveillance‡ 9/20 (45) 4/5 (80)

*In counseled female members of 29 BRCA1 families and counseled female members in 8 BRCA2 families (number/number queried† (percent)).†The number queried varies from item to item since not all questions were asked and/or responded to within the genetic counseling setting.‡Excluding women aged < 25 or with bilateral mastectomies.BE= breast examination.Reproduced with permission of author and publisher from Lynch et al. Cancer Genet Cytogenet 1999;109:91–8.


included transvaginal ovarian ultrasound,Doppler color bloodflow imagery, and CA-125.However, the patients were thoroughly educatedabout the limitations of ovarian cancer screening.The option of prophylactic mastectomy and/orprophylactic oophorectomy was also discussedduring genetic counseling sessions.50

Preliminary data show that psychologicalassessment 6 months following BRCA1/BRCA2testing among unaffected individuals (bothmale and female) from HBOC families did notreflect adverse psychological effects.53 How-ever, with respect to screening, we did find thatrates of adherence to mammography recom-mendations among mutation carriers was notincreased. It was also noted that carriers ofdeleterious genes who said they would considerprophylactic surgery nevertheless showed lowrates of actually adopting such options. How-ever, these observations are based upon short-term experience; longer-term data will berequired to determine how often women mayopt for prophylactic surgery.53

PROPHYLACTIC MASTECTOMY

In a landmark study, Hartmann and colleagues54

pursued the efficacy of prophylactic mastec-

tomy on a retrospective cohort study of allwomen with a breast cancer-positive family his-tory who underwent bilateral prophylactic mas-tectomy at the Mayo Clinic between 1960 and1993. These women were divided into high-riskversus moderate-risk groups based on their fam-ily history. Those at high risk showed pedigreesconsistent with a single-gene, autosomal domi-nant predisposition to carcinoma of the breast,whereas those at moderate risk showed positivefamily histories that did not meet these high-riskcriteria. To predict the number of breast cancersexpected in these two groups had prophylacticmastectomies not been performed, theresearchers used a nested-sister control studyfor the high-risk group and the Gail Model forthe moderate-risk group. Their findings werebased upon 693 women with a family history ofbreast cancer (214 high-risk and 425 moderate-risk) who had bilateral prophylactic mastec-tomies. Their median follow-up was 14.4 years,while the median age at prophylactic mastec-tomy was 43 (mean 42.4) years. The Gail Modelprediction for breast cancer occurrenceexpected in the moderate-risk group was 37.4.However, only four breast cancers occurred fol-lowing prophylactic mastectomy in this group(89.5% reduction, p < .001).

Table 1–6. SURVEILLANCE PRACTICES AND ATTITUDES TOWARD PROPHYLACTIC OOPHORECTOMIES PRIOR TO GENETIC TEST RESULT DISCLOSURE*

BRCA1 BRCA2

Bilateral oophorectomies prior to the counseling sessionOvarian cancer-affected 7/250 ( 3) 0/65 ( 0)Ovarian cancer-free 83/250 (33) 5/65 ( 8)

Current surveillance practices‡

CA-125 Ever 26/65 (40) 4/19 (21)As recommended 10/49 (20) 2/17 (12)

Ultrasound Ever 32/71 (45) 11/25 (44)As recommended 11/51 (22) 2/16 (12)

Pelvic Ever 86/96 (90) 50/51 (98)As recommended 64/79 (81) 45/46 (98)

If carrier, will consider prophylactic oophorectomy‡ 72/81 (89) 10/20 (50)If carrier, will consider increasing surveillance‡ 50/53 (94) 25/26 (96)If noncarrier, will consider decreasing surveillance‡ 8/15 (53) 5/6 (83)

*In counseled female members of 29 BRCA1 families and counseled female members in 8 BRCA2 families (number/number queried† (percent)).†The number queried varies from item to item since not all questions were asked and/or responded to within the genetic counseling setting.‡Excluding women aged < 25 or with bilateral oophorectomies.Reproduced with permission of author and publisher from Lynch et al. An update on DNA-based BRCA1/BRCA2 genetic counseling in hereditarybreast cancer. Cancer Genet Cytogenet 1999;109:91–8.

14 BREAST CANCER

Breast cancer occurrences in the 214 high-risk probands were compared to their sisterswho had not undergone prophylactic mastec-tomy. These 214 probands had a total of 403sisters. Of keen interest was the finding thatthere have been 156 breast cancers in the sis-ters; 115 occurred before the sister proband’smastectomy, 38 after the sister proband’s pro-phylactic mastectomy, and 3 with dateunknown. In comparison, 3 of the 214 probandshad developed breast cancer. This represents a> 90 percent reduction in the incidence ofbreast cancer with current follow-up. Breastcancer mortality was also reduced significantlyin both the high and moderate-risk groups.

The investigators concluded that prophy-lactic mastectomy resulted in a significantreduction in the incidence of and mortalityfrom breast cancer among these women withpositive family histories of breast cancer.This information will be useful for geneticcounseling.54

Schrag and colleagues55 discuss the deci-sion analysis involved in prophylactic mastec-tomy and oophorectomy and life expectancyoutcome among patients with BRCA1 andBRCA2 germ-line mutations. They found that,on average, a 30-year-old woman harboringsuch a mutation would gain from 2.9 to 5.3years of life expectancy from prophylacticmastectomy and from 0.3 to 1.7 years fromprophylactic oophorectomy. These findingswere dependent upon their cumulative risk ofcancer. Gains in life expectancy also woulddecline with age at the time of prophylacticsurgery. They would be minimal for a 60-year-old woman. Importantly, in women aged 30, anoophorectomy may be delayed for 10 yearswith minimal loss of life expectancy. Thiswould allow women to complete their families.These investigators concluded that “On thebasis of a range of estimates of the incidenceof cancer, prognosis, and efficacy of prophy-lactic surgery, our model suggests that prophy-lactic mastectomy provides substantial gains inlife expectancy and prophylactic oophorec-

tomy more limited gains for young womenwith BRCA1 or BRCA2 mutations.”

CONSERVATIVE VERSUS CONVENTIONAL SURGERY IN

HEREDITARY BREAST CANCER

Should a patient with HBOC, particularly onewho is harboring a BRCA1 or BRCA2 germ-line mutation, be managed differently from apatient with the more common sporadic formof this disease? We have taken the positionthat, because of the early age of breast canceronset and excess lifetime risk for bilaterality,coupled with the potential deficiency of repairof radiation-induced DNA damage,44 thepatient should be given the option of total mas-tectomy as opposed to conservative (“lumpec-tomy”) management, and seriously considercontralateral prophylactic mastectomy, assum-ing that the ipsilateral breast cancer is likely tohave adequate control.

POTENTIAL FOR TARGETEDBRCA1/BRCA2 MUTATION THERAPY

In addition to identifying cancer risk statusthrough mutations such as BRCA1 andBRCA2, this knowledge has the potential toprovide individualized highly-targeted molec-ular genetic therapies based upon mutationdiscoveries.56 Specifically, once the functionsof cancer susceptibility genes have been iden-tified, knowledge as to how such gene-deter-mined biochemical functions can be employedfor targeted radiation and chemotherapyshould emerge.

Abbott and colleagues57 examined the pro-tein product of the BRCA2 gene in terms of itshaving an important role in mediating repair ofdouble-strand breaks in DNA. They identified ahuman pancreatic carcinoma cell line whichlacked one copy of the BRCA2 gene and con-tained a mutation (617delT) in the remainingcopy.58 They performed in vitro and in vivoexperiments in this cell line as well as with other


carefully matched cell lines. They then exam-ined double-strand break repair with attentiongiven to sensitivity to drugs and radiation effectthat induce double-strand breaks. Their findingsdisclose that “... BRCA2-defective cells areunable to repair the double-strand DNA breaksinduced by ionizing radiation. These cells werealso markedly sensitive to mitoxantrone,amsacrine, and etoposide... (two-sided p = .002)and to ionizing radiation (two-sided p = .001).Introduction of antisense BRCA2 deoxyribonu-cleotides into cells possessing normal BRCA2function led to increased sensitivity to mitox-antrone (two-sided p = .008). Tumors formed byinjection of BRCA2-defective cells into nudemice were highly sensitive (> 90% tumor sizereduction, two-sided p = .002) to both ionizingradiation and mitoxantrone when comparedwith tumors exhibiting normal BRCA2 func-tion.” Abbott and colleagues concluded thatthese BRCA2-defective cancer cells were highlysensitive to agents that contribute to double-strand breaks in DNA.

SURVEILLANCE AND MANAGEMENT FOR HEREDITARY BREAST CANCER

When the diagnosis of a hereditary breast can-cer-prone syndrome has been established, thesurveillance and management strategies are thenmelded to the natural history of the particularHBC syndrome. We recommend that patientsreceive intensive education regarding the naturalhistory, genetic risk, and availability of DNAtesting such as BRCA1, BRCA2, or p53, depend-ing upon the hereditary breast cancer syndromeof concern. We initiate such education betweenthe ages of 15 and 18 years but do not performany DNA testing until the patient is > 18 yearsof age and has given informed consent.

When patients are 18 years old, the authorsprovide instruction in breast self examination(BSE) with physician assessment of their per-formance. Although the effectiveness of BSEhas been controversial, the authors are con-vinced that it can be effective if the woman is

taught how to perform this procedure anddemonstrates proficiency in doing so on returnmedical visits. When patients reach the age of20 years, we begin semiannual breast examina-tion by the physician and at age 25 initiateannual mammography.

With respect to ovarian cancer, we discusstransvaginal ovarian ultrasound, Doppler colorbloodflow imaging, and CA-125, with their lim-itations, and initiate this at age 30 and performit annually. The option of prophylactic bilateraloophorectomy is also discussed. If the patient isinterested in this option and has completed herchild bearing, the oophorectomy can be per-formed between the ages of 35 and 45 years.

CONCLUSION

It is necessary to keep an open mind about thepros and cons of DNA testing and genetic coun-seling and its translation into medical practice bythe basic scientist, medical and moleculargeneticist, physician, genetic counselor, andethicist. For example, how does one interpretsome of the ethical positions of today suggestingthat genetic testing be limited to a research set-ting or even curtailed until specific benefit ofsuch DNA testing can be more fully established?

Prodigious advances in science and technol-ogy (ie, better surveillance, more effective surgi-cal management including data supporting pro-phylactic surgery, and improved moleculargenetic technology with lower cost for germ-linemutation discovery) may resolve some of theconcerns about molecular genetic testing thatcause certain physicians, basic scientists, andethicists to believe it should be limited. Howdoes one educate these colleagues about newlyemerging benefits of molecular genetic testingwhich could prove to be lifesaving?

Molecular genetic advances are occurring atsuch a rapid pace that it is exceedingly difficultto keep physicians fully informed of thisprogress. For example, it is predicted that theentire human genome will be identified by theyear 2003.

16 BREAST CANCER

The lay press strives to keep the public fullyinformed about the impact new gene discover-ies may have on patients and their close rela-tives. Unfortunately, certain members of themedia have overinterpreted the benefit ofgerm-line testing and have not fully dealt withsome of its drawbacks. In turn, some moleculargenetic laboratories have made testing appearto be the panacea for cancer control. Some mayoffer molecular testing without sufficient evi-dence that the family of concern merits testing.Genetic counseling may not be provided tothose being tested. In spite of these misgivings,we believe that hereditary breast cancerpatients, when properly counseled and DNAtested, will benefit immensely during this excit-ing era of molecular genetics.

Patients must be encouraged to meticulouslyexamine their family histories of cancer. Shouldthey be found to harbor a germ-line cancer pre-disposing mutation, this knowledge may be usedto encourage screening and detect cancer at anearly stage so that a cure might be possible and/orcancer prevented through the option of prophy-lactic surgery.

REFERENCES

1. Vogelstein B, Kinzler KW. The genetic basis ofhuman cancer. New York: McGraw-Hill; 1998.

2. Landis SH, Murray T, Bolden S, Wingo PA. Cancerstatistics, 1999. CA Cancer J Clin 1999;49:8–31.

3. Lynch HT, Lemon SJ, Durham C, et al. A descrip-tive study of BRCA1 testing and reactions to dis-closure of test results. Cancer 1997;79:2219–28.

4. Lynch HT, Krush AJ. Carcinoma of the breast andovary in three families. Surg Gynecol Obstet1971;133:644–8.

5. Lynch HT, Krush AJ, Lemon HM, et al. Tumorvariation in families with breast cancer. JAMA1972;222:1631–5.

6. Lynch HT, Guirgis HA, Albert S, et al. Familialassociation of carcinoma of the breast andovary. Surg Gynecol Obstet 1974;138:717–24.

7. Hall JM, Lee MK, Newman B, et al. Linkage ofearly-onset breast cancer to chromosome17q21. Science 1990;250:1684–9.

8. Narod SA, Feunteun J, Lynch HT, et al. Familialbreast-ovarian cancer locus on chromosome17q12–q23. Lancet 1991;388:82–3.

9. Miki Y, Swensen J, Shattuck-Eidens D, et al. Astrong candidate for the breast and ovarian can-cer susceptibility gene BRCA1. Science 1994;266:66–71.

10. Wooster R, Neuhausen SL, Mangion J, et al. Local-ization of a breast cancer susceptibility gene,BRCA2, to chromosome 13q12–13. Science1994;265:2088–90.

11. Wooster R, Bignell G, Lancaster J, et al. Identifi-cation of the breast cancer susceptibility geneBRCA2. Nature 1995;378:789–92.

12. Easton DF, Bishop DT, Ford D, Crockford GP, TheBreast Cancer Linkage Consortium. Geneticlinkage analysis in familial breast and ovariancancer: results from 214 families. Am J HumGenet 1993;52:678–701.

13. Easton DF, Ford D, Bishop DT, The Breast CancerLinkage Consortium. Breast and ovarian can-cer incidence in BRCA1 mutation carriers. AmJ Hum Genet 1995;56:265–71.

14. Struewing JP, Hartge P, Wacholder S, et al. Therisk of cancer associated with specific muta-tions of BRCA1 and BRCA2 among AshkenaziJews. N Engl J Med 1997;336:1401–8.

15. Claus EB, Schildkraut J, Iversen ES Jr, Berry D,Parmigiani G. Effect of BRCA1 and BRCA2 onthe association between breast cancer risk andfamily history. J Natl Cancer Inst 1998;90:1824–9.

16. Lynch HT, Follett KL, Lynch PM, et al. Familyhistory in an oncology clinic. Implications forcancer genetics. JAMA. 1979;242:1268–72.

17. David KL, Steiner-Grossman P. The potential useof tumor registry data in the recognition andprevention of hereditary and familial cancer.NY State J Med 1991;91:150–2.

18. Lynch HT, Lemon SJ, Marcus JN, et al. Breastcancer genetics: heterogeneity, moleculargenetics, syndrome diagnosis, and geneticcounseling. In: Bland KI, Copeland EMI, edi-tors. The breast: comprehensive managementof benign and malignant diseases. 2 ed.Philadelphia (PA): W. B. Saunders Company;1998. p. 370–94.

19. Gayther SA, Warren W, Mazoyer S, et al.Germline mutations of the BRCA1 gene inbreast and ovarian cancer families provide evi-dence for a genotype-phenotype correlation.Nat Genet 1995;11: 428–33.

20. Serova O, Montagna M, Torchard D, et al. A highincidence of BRCA1 mutations in 20 breast-ovarian cancer families. Am J Hum Genet1996;58:42–51.


21. Gayther SA, Mangion J, Russell P, et al. Variationof risks of breast and ovarian cancer associatedwith different germline mutations of theBRCA2 gene. Nat Genet 1997;15:103–5.

22. Ramus SJ, Friedman LS, Gayther SA, PonderBAJ. A breast/ovarian cancer patient withgermline mutations in both BRCA1 andBRCA2. Nat Genet 1997;15:14–15.

23. Tonin P, Moslehi R, Green R, et al. Linkage analy-sis of 26 Canadian breast and breast-ovariancancer families. Hum Genet 1995;95:545–50.

24. Neuhausen S, Gilewski T, Norton L, et al. Recur-rent BRCA2 617delT mutations in AshkenaziJewish women affected by breast cancer. NatGenet 1996;13:126–8.

25. Couch FJ, Farid LM, DeShano ML, et al. BRCA2germline mutations in male breast cancer casesand breast cancer families. Nat Genet 1996;13:123–5.

26. Liede A, Metcalfe K, Offit K, et al. A family withthree germline mutations in BRCA1 andBRCA2. Clin Genet 1998;54:215–8.

27. Struewing JP, Abeliovich D, Peretz T, et al. The car-rier frequency of the BRCA1 185delAG mutationis approximately 1 percent in Ashkenazi Jewishindividuals. Nat Genet 1995;11:198–200.

28. Roa BB, Boyd AA, Volcik K, Richards CS.Ashkenazi Jewish population frequencies forcommon mutations in BRCA1 and BRCA2. NatGenet 1996;14:185–7.

29. Oddoux C, Struewing JP, Clayton CM, et al. Thecarrier frequency of the BRCA2 6174delT muta-tion among Ashkenazi Jewish individuals isapproximately 1%. Nat Genet 1996;14:188–90.

30. Marcus JN, Watson P, Page DL, et al. Hereditarybreast cancer: pathobiology, prognosis, andBRCA1 and BRCA2 gene linkage. Cancer 1996;77:697–709.

31. Marcus JN, Page DL, Watson P, et al. BRCA1 andBRCA2 hereditary breast carcinoma pheno-types. Cancer 1997;80 Suppl:543–56.

32. Marcus JN, Watson P, Page DL, et al. BRCA2hereditary breast cancer phenotype. BreastCancer Res Treat 1997;44:275–7.

33. Marcus JN, Watson P, Page DL, Lynch HT. Thepathology and heredity of breast cancer inyounger women. J Natl Cancer Inst Monogr1994;16:23–34.

34. Thompson ME, Jensen RA, Obermiler PS, et al.Decreased expression of BRCA1 acceleratesgrowth and is often present during sporadicbreast cancer progression. Nat Genet 1995;9:444–50.

35. Rao VN, Shao N, Ahmad M, Reddy ESP. Anti-sense RNA to the putative tumor suppressorgene BRCA1 transforms mouse fibroblasts.Oncogene 1996;12:523–8.

36. Holt JT, Thompson ME, Szabo CI, et al. Growthretardation and tumour inhibition by BRCA1.Nat Genet 1996;12:298–302.

37. Armes JE, Egan AJM, Southey MC, et al. The his-tologic phenotypes of breast carcinoma occur-ring before age 40 years in women with andwithout BRCA1 or BRCA2 germline mutations:a population-based study. Cancer 1998;83:2335–45.

38. Collins N, McManus R, Wooster R, et al. Consistentloss of the wild type allele in breast cancers froma family linked to the BRCA2 gene on chromo-some 12q12–13. Oncogene 1995;10:1673–5.

39. Sigurdsson H, Agnarsson BA, Jonasson JG, et al.Worse survival among breast cancer patients infamilies carrying the BRCA2 susceptibilitygene. [abstract] Breast Cancer Res Treat 1996;37:33.

40. Breast Cancer Linkage Consortium. Pathology offamilial breast cancer: differences betweenbreast cancers in carriers of BRCA1 or BRCA2mutations and sporadic cases. Lancet 1997;349:1505–10.

41. Lakhani SR, Jacquemier J, Sloane JP, et al. Multi-factorial analysis of differences between spo-radic breast cancers and cancers involvingBRCA1 and BRCA2 mutations. J Natl CancerInst 1998;90:1138–45.

42. Eng C, Li FP, Abramson DH. Mortality from sec-ond tumors among long-term survivors ofretinoblastoma. J Natl Cancer Inst 1993;85:1121–6.

43. Ford D, Easton DF, Peto J. Estimates of the genefrequency of BRCA1 and its contribution tobreast and ovarian cancer incidence. Am JHum Genet 1995;57:1457–62.

44. Scully R, Chen J, Plug A, et al. Association ofBRCA1 with Rad51 in mitotic and meioticcells. Cell 1997;88:265–75.

45. Chabner E, Nixon A, Gelman R, et al. Family his-tory and treatment outcome in young womenafter breast-conserving surgery and radiationtherapy for early-stage breast cancer. J ClinOncol 1998;16:2045–51.

46. Lim DS, Hasty P. A mutation in mouse Rad51results in early embryonic lethal that is sup-pressed by a mutation in p53. Mol Cell Biol1996;16:7133–43.

18 BREAST CANCER

47. Watson P, Lynch HT. Extracolonic cancer inhereditary nonpolyposis colorectal cancer.Cancer 1993;71:677–85.

48. Lynch HT, Casey MJ, Lynch J, et al. Genetics andovarian carcinoma. Semin Oncol 1998;25:265–81.

49. American Society of Clinical Oncology. State-ment of the American Society of ClinicalOncology: genetic testing for cancer suscepti-bility. J Clin Oncol 1996;14:1730–6.

50. Lynch HT, Watson P, Tinley S, et al. An update onDNA-based BRCA1/BRCA2 genetic counsel-ing in hereditary breast cancer. Cancer GenetCytogenet 1999;109:91–8.

51. Fleiss JL. Statistical methods for rates and pro-portions. 2nd ed. New York: John Wiley &Sons; 1981.

52. Lerman C, Hughes C, Lemon SJ, et al. What youdon’t know can hurt you: adverse psychologiceffects in members of BRCA1–linked andBRCA2-linked families who decline genetictesting. J Clin Oncol 1998;16:1650–4.

53. Lerman C, Narod S, Schulman K, et al. BRCA1

testing in families with hereditary breast-ovarian cancer: a prospective study of patientdecision-making and outcomes. JAMA 1996;275:1885–92.

54. Hartmann LC, Schaid DJ, Woods JE, et al. Effi-cacy of bilateral prophylactic mastectomy inwomen with a family history of breast cancer.N Engl J Med 1999;340:77–84.

55. Schrag D, Kuntz KM, Garber JE, Weeks JC. Deci-sion analysis—effects of prophylactic mastec-tomy and oophorectomy on life expectancyamong women with BRCA1 or BRCA2 muta-tions. N Engl J Med 1997;336:1465–71.

56. Livingston DM. Genetics is coming to oncology.JAMA 1997;277:1476–7.

57. Abbott DW, Freeman ML, Holt JT. Double-strandbreak repair deficiency and radiation sensitiv-ity in BRCA2 mutant cancer cells. J Natl Can-cer Inst 1998;90:978–85.

58. Goggins M, Schutte M, Lu J, et al. GermlineBRCA2 gene mutations in patients with appar-ently sporadic pancreatic carcinomas. CancerRes 1996; 56:5360–4.

19

In the past decade, the systematic use of mam-mography as part of diagnostic screening pro-grams and the extensive use of stereotacticfine-needle biopsy techniques have greatlyimproved our ability to detect pre-invasive aswell as microinvasive breast carcinomas. Theconsequent earlier detection of breast lesions isconsidered the most important factor explain-ing the recent decline in overall mortality frombreast cancer observed in the United States.1

This progress has translated into many effortsto gain insight into the biologic mechanismsresponsible for cancer development and pro-gression and to identify potential areas of inter-vention for prevention studies.2

About 20 percent of women diagnosed withproliferative breast disease display atypicalhyperplasia, a condition associated with increas-ed risk of breast cancer and probably a precursorto the disease.3 The recognition of the importanceof these benign breast lesions as risk factors andthe identification of genetic alterations (BRCA1,BRCA2, p53) associated with genetic predisposi-tion to breast cancer have spurred investigationin the areas of prophylactic surgery and chemo-prevention in the management of women at highrisk of breast cancer and directed attention tospecific interventions and to the design of com-prehensive surveillance programs.4–6

The development of effective chemopreven-tion strategies requires a systematic approachfocusing on multiple investigational aspects ofthe problem including (1) a clear description ofthe specific risk factors that may be used inselecting cohorts of women at increased risk andas end points for chemoprevention studies; (2)the identification and preclinical evaluation ofcancer chemoprevention agents and subsequentdevelopment of definitive, large-scale clinical tri-als; and (3) the identification and characteriza-tion of specific molecular biomarkers that maybe quantitatively assessed and used as surrogateend-point biomarkers (SEBs) in future trials.

BREAST CANCER RISK ASSESSMENT

Chemoprevention is traditionally defined as theinhibition or reversal of carcinogenesis, beforeovert malignancy, by intervention with chemi-cal agents. Most breast cancer chemopreventivestudies are conducted within cohorts of womenconsidered at high risk of the disease.7 In thecontext of chemoprevention investigations, themost important cancer risk factors are consid-ered to be those that can be measured quantita-tively in the subject at risk. These factors arecalled risk biomarkers, and they can be used toidentify cohorts for chemoprevention trials.8

2

Breast Cancer Risk and Management:Chemoprevention, Surgery, and SurveillanceMASSIMO CRISTOFANILLI, MDLISA NEWMAN, MD, FACSGABRIEL N. HORTOBAGYI, MD

Generally, the risk biomarkers are groupedin the following categories: (1) genetic predis-position, (2) carcinogen exposure, (3) carcino-gen effect/exposure, (4) previous cancers, and(5) intermediate biomarkers. In some cases,risk biomarkers that are measurable and mayundergo selective modulation by chemopreven-tive agents can be used as SEBs in clinicalchemoprevention studies.9 Today, there is not asingle ideal risk biomarker or SEB for breastcancer, and the selection of high-risk subjectsfor breast cancer chemoprevention studies isgenerally done on the basis of the presence ofproliferative atypical breast disease or epidemi-ologic and genetic factors known to increase awoman’s risk of developing breast cancer.10,11

The quantification of the risk of developingcancer is usually estimated from epidemiologic

models, which consider a variety of risk factorsand project a cumulative risk for the develop-ment of disease over a finite period of time.Among the various models proposed, the modeldeveloped by Gail and colleagues is probably themost widely accepted.12 Proposed in the originalform in 1989 and subsequently modified, thismodel was designed to calculate the absolutebreast cancer risk in women on the basis of thedata from the Breast Cancer Detection Demon-stration Project (BCDDP) that recruited 280,000women in 28 centers in the United States in themid-1970s. It represented an attempt to definethe contribution of accumulated breast cancerrisk from a number of risk factors combined in amultivariate logistic regression model. The vari-ables identified included age of the patient, num-ber of first-degree relatives with breast cancer,

20 BREAST CANCER

Table 2–1. RISK FACTORS CONSIDERED FOR THE CALCULATION OF ABSOLUTE BREAST CANCER RISK

Risk Factor Associated Relative Risk Expansion Variables

Age at menarche (y)³ 14 1.00012 to 13 1.099 A< 12 1.207

No. of breast biopsiesAge < 50 y

0 (0) 1.0001 (1) 1.698³ 2 (2) 2.882

Age ³ 50 y B0 (0) 1.0001 (1) 1.273³ 2 (2) 1.620

Age at first-term live birth (y) No. of first-degree relativeswith breast cancer

< 20 0 (0) 1.0001 (1) 2.607

³ 2 (2) 6.79820 to 24 0 (0) 1.244

1 (1) 2.681³ 2 (2) 5.775 C

25 to 29 or nulliparous 0 (0) 1.5481 (1) 2.756

³ 2 (2) 4.907³ 30 0 (0) 1.927

1 (1) 2.834³ 2 (2) 4.169

Absolute risk = A ´ B ´ C ´ 1.82 *A ´ B ´ C ´ 0.93†

*Presence of atypical hyperplasia in biopsies.†No atypical hyperplasia.Adapted from Gail MH, Brinton LA, Byar DP, et al. Projecting individualized possibilities of developing breast cancer for white females who arebeing examined annually. J Natl Cancer Inst 1989;81:1879–86.

nulliparity or age at first live birth, number ofbreast biopsies, presence or absence of atypicalhyperplasia, and age at menarche (Table 2–1).The individualized absolute risk can be esti-mated by combining the relative risk based onthe woman’s age and individual risk factors andthen combining this information with an esti-mate of the baseline hazard rate for a womanwith no risk factors.12,13 In particular, womenwith a strong family history of breast cancer (2first-degree relatives affected, or 1 first-degreeand 2 second-degree relatives affected) or ofbreast and ovarian cancers (3 affected relativeswith breast cancer and 1 relative with ovariancancer) have a 25 percent or greater lifetime riskof developing breast cancer. A computer pro-gram called RISK has been subsequently devel-oped and allows physicians to calculate awoman’s absolute risk to develop breast cancerwith 95 percent confidence interval bounds.14

Familial and Genetic Factors

Identification of cohorts at genetic risk for can-cer is an appealing concept because it may offerthe opportunity to explore the steps in breastcarcinogenesis, from the inheritance of a pre-disposing mutation through the development ofpreinvasive lesions or overt malignancy. How-ever, germline genetic changes are rare andreported in only a small proportion of womenwho will eventually develop breast cancer.15

Breast cancer attributed to a family historyof the disease has been reported to account for 6to 19 percent of all cases of breast cancer.Hereditary breast cancers, which constitute aproportion of these cases, are characterized byearly onset, a high incidence of bilateral disease,association with other malignancies, and auto-somal dominant inheritance. Genetic-linkagestudies of families with multiple members withbreast cancer have allowed major improvementsin our understanding of the genetic alterationsassociated with hereditary predisposition. Suchstudies led to the discovery of germline muta-tions in the BRCA1 and BRCA2 genes.16

The BRCA1 gene, a breast cancer susceptibil-ity gene localized to chromosome 17q, wasdescribed for the first time in 1990. It is a tumor-suppressor gene postulated to be important inregulating the growth of breast epithelial cellsand in the process of deoxyribonucleic acid(DNA) damage repair.17 The breast cancer sus-ceptibility gene BRCA1 accounts for 45 percentof hereditary cases of breast cancer and 80 to 90percent of hereditary cases of combined breastand ovarian cancers. The BRCA1 mutations arefound in approximately 7 to 12 percent of womenwith breast cancer of early onset. Mutations havebeen described as spreading evenly across theentire gene. The most commonly described muta-tion is a specific alteration causing a deletion ofadenine and guanine (185delAG). This mutationis present in 1 percent of the Ashkenazi Jewishpopulation and contributes to a risk of breast can-cer as high as 21 percent among Jewish women.Women who carry a BRCA1 mutation were ini-tially estimated to have an 87 percent lifetime riskof breast cancer and a 44 percent lifetime risk ofovarian cancer.18 Subsequent studies suggestrisks of 56 percent and 16 percent, respectively.

More recent data from population-basedstudies suggest that BRCA1 mutations accountfor only 10 to 20 percent of inherited breastcancer, and BRCA2 mutations are responsiblefor half this fraction. The discrepancy betweenestimates from early studies is probably relatedto the fact that initial data were derived fromlinkage analysis that probably tends to overesti-mate the true incidence of BRCA1 and BRCA2mutations in hereditary breast cancer.19–21

The breast cancer susceptibility geneBRCA2, localized to chromosome 13q, wasdescribed in 1994. Linkage studies suggest that35 percent of high-risk families may haveBRCA2 mutations.22 Male breast cancer, a rarecondition that represents less than 1 percent ofall cancer and 0.1 percent of cancer-relateddeaths in men, has been found to be associatedwith mutations in the BRCA2 gene.23

Familial clustering of breast cancer has alsobeen described in families diagnosed with Cow-

Breast Cancer Risk and Management: Chemoprevention, Surgery, and Surveillance 21

den syndrome and Li-Fraumeni syndrome, andmore recently, ataxia-telangiectasia.24,25 Cow-den syndrome involves multiple hamartomatouslesions, especially of the skin, and mucousmembranes, and carcinoma of the breast andthyroid. Li-Fraumeni syndrome, associated witha high incidence of p53 mutations, consists of afamilial aggregation of breast carcinomas, softtissue sarcomas, brain tumors, osteosarcomas,leukemias, and adrenocortical carcinomas.

Hereditary breast cancer syndromes areclinically relevant because they raise the possi-bility of effective identification, throughgenetic testing, of patients at high risk andoptimal quantification of the risk. In view ofthe high risk of developing an invasive breastcancer during their lifetime and in considera-tion of the imperfect nature of early detection,patients with hereditary breast cancer syn-dromes constitute an ideal target for chemo-prevention studies.

Environmental Radiation Exposure

Epidemiologic observations suggest that expo-sure of breast tissue to ionizing radiation is asso-ciated with an increased risk of breast cancer. Inparticular, an increased risk of breast carcinomahas been clearly documented in young womenthat have survived atomic bombs, in patientswho have undergone repeated fluoroscopies (eg,in patients with tuberculosis), and in patientstreated with radiation for postpartum mastitis,thymic enlargement, and Hodgkin’s disease.26–37

Treatment-associated second neoplasms haveemerged as a major complication in patientscured of Hodgkin’s disease. Sporadic cases ofbreast carcinoma after mantle irradiation forHodgkin’s disease were reported beginning in1978, by which time clinical data on a largecohort of women treated and cured for their dis-ease in the 1960s had become available.25

Subsequent reports directed attention to thecarcinogenic risk associated with radiationexposure and the latency time between expo-sure and clinical manifestations of breast can-

cer. A latency period of at least 10 years is usu-ally required in the majority of the patients; ashorter latency period (5 to 10 years) has beenreported occasionally.31

These observational studies suggest that thecarcinogenic process is related to two factors:the dose of radiation delivered and the person’sage at the time of exposure. In general, youngwomen (less than 30 years of age) represent acohort with a higher relative risk comparedwith the risk for other age groups. The cluster-ing of breast carcinoma in women irradiated forHodgkin’s disease at an early age is probablyrelated to the increased sensitivity of theincompletely differentiated breast epithelium tothe carcinogenic action of radiation.33

The high risk of breast carcinoma in youngwomen with Hodgkin’s disease treated withirradiation mandates regular follow-up withbreast examination and yearly mammographystarting within 8 years of completion of theradiation treatment.37

The hypothetical risk of breast cancerderived from prolonged screening needs to bementioned. Using a risk estimate provided bythe Biological Effects of Ionizing Radiation(BEIR) V Report of the National Academy ofSciences and a mean breast glandular dose of4 mGy from bilateral mammography, with twoviews per breast, one can estimate that annualmammography of 100,000 women for 10 con-secutive years beginning at age 40 years willresult in, at most, 8 breast cancer deaths duringtheir lifetime. On the other hand, researchershave shown a 24 percent mortality reductionfrom biennial screening of women in this agegroup; this will result in one life lost. Anassumed mortality reduction of 36 percent fromannual screening would result in 36.5 livessaved per life lost and 91.3 years of life savedper year of life lost. Thus, the theoretic radia-tion risk from screening mammography isextremely small compared with the establishedbenefit from this life-saving procedure andshould not unduly distract women under age 40years who are considering screening.38

22 BREAST CANCER

Intermediate Biomarkers

A key element in the control and prevention ofinvasive breast cancer is the recognition of early,preclinical changes with or without associatedcharacteristic molecular abnormalities that mayidentify a woman as being at high risk for devel-opment of the disease.39 These alterations, calledintermediate biomarkers, may be used as SEBsand provide a more cost-effective and rapidmeans of testing chemopreventive interventions.To date, the most specific intermediate biomark-ers for invasive breast cancer development areprecancerous and preinvasive lesions, such ashistologic changes of atypical hyperplasia, lobu-lar carcinoma in situ (LCIS), and ductal carci-noma in situ (DCIS). Observations about thevalue of molecular alterations, for example,c-erbB-2 overexpression and p53 accumulation,as markers of disease progression and increasedrisk in patients with benign breast conditions arecontradictory. 40

Intermediate biomarkers are essentially pre-cancerous lesions identified as being directlyon the control pathway to cancer. Their pres-ence puts carriers at high risk for invasive dis-ease. A hypothetical model of intraepithelialbreast neoplasms postulates progression fromfocal aberrant ductal or lobular proliferation(hyperplasia) to cellular pleomorphism, disor-ganized growth, and abnormal growth (dyspla-sia), and, finally, focally invasive cancer.41

The histologic abnormalities, described byPage and colleagues3 and collectively known asproliferative breast disease (PBD), are associatedwith increased risk of breast cancer; the lesionsand associated risks of cancer development varyfrom moderate hyperplasia (two-fold risk), toatypical ductal hyperplasia (four-fold risk), tocarcinoma in situ (CIS) (11-fold risk).3,4,42

The presence of these lesions in the con-tralateral breast of patients with a history ofbreast cancer has been associated with up to a0.8 percent chance per year of developing anew primary tumor. In this context, theselesions appear to be the ideal target for chemo-

prevention studies, the main advantage beingthe possibility of a more objective assessmentof the effectiveness of interventions throughrepeated core biopsies.

In recent years, the availability of sensitiveassays of DNA damage and genetic instabilityhave prompted novel investigations and may helpdefine subsets of women who are at high risk ofdeveloping breast cancer. Among the various mol-ecular markers investigated, p53 overexpression,dysplasia, and aneuploidy have been found to beassociated with increased risk of invasive breastcarcinoma.43,44 Rohan and colleagues conducted acase-control study within a cohort of 4,888 Cana-dian women in the National Breast ScreeningStudy (NBSS) who were diagnosed with benignbreast disease. Case subjects were the women inwhom breast cancer subsequently developed. Thec-erbB-2 protein overexpression and p53 accu-mulation were determined by immunohisto-chemical techniques. Accumulation of p53 wasassociated with an increased risk of breast cancer(adjusted odds ratio 2.5), while c-erbB-2 proteinoverexpression was not.44 Even with multiplemethodologic problems, this investigation pro-vides an interesting model for the clinical use ofbiomarkers in benign breast disease.

The design of prospective trials that includesthe evaluation of p53 status along with othermarkers studied in randomly obtained fine-needle aspirates may provide a unique opportu-nity to identify specific, measurable, intermedi-ate biomarkers that may be used in short-termtrials to verify the efficacy of chemopreventiveagents.

ESTROGENS AND BREAST CANCER RISK

Endogenous Hormones

The controversy surrounding exogenous hor-mone use and breast cancer risk is predicatedon the concept that breast carcinogenesis is ahormone-dependent process. The ability tocontrol breast cancer with hormonal manipula-


tion has been recognized since 1986, whenBeatson reported on oophorectomy as a suc-cessful treatment for this disease.45 Since thattime several other epidemiologic, experimentaland clinical lines of evidence have developedthat also support this concept.46

It is well established that menstrual factorsresulting in exposure of the breast to increasednumbers of ovulatory estrogen cycles over alifetime, such as early menarche (< 13 years),late menopause (> 50 years), and nulliparitycan increase the risk of breast cancer.46,47 Con-versely, bilateral oophorectomy at a young ageand interruptions of the menstrual cycle in theform of multiple pregnancies may confer a pro-tective effect.48

The impact of pregnancy on the risk ofbreast cancer is strongest in the case of thefirst pregnancy occurring at a young age(before 20 years). The rate of proliferation ofthe ductal epithelium is normally high afterpuberty. The hormonal influences associatedwith pregnancy induce a process of terminalductal and lobular stem cell differentiation,theoretically rendering the breast more resis-tant to carcinogenesis.47,48 Henderson and col-leagues hypothesized that completion of afull-term pregnancy is crucial for this protec-tive effect because the rapid increase in freeestradiol during the first trimester of preg-nancy is “equivalent to several ovulatorycycles over a relatively short period of time.”They hypothesized that failure to over-ride thisestrogenic surge with the subsequent hormonalchanges of advanced pregnancy (as occurswith first-trimester abortions) can result inincreased risk of breast cancer.49

It is also well established that estrogen andprogesterone exert proliferative effects onhuman breast tissue49,50 and that estrogen canpromote mammary tumorigenesis in animalmodels as well as in in vitro tissue cultures.47–51

Postmenopausal obesity has been associatedwith increased breast cancer risk, and this rela-tionship appears to be mediated by age-relatedvariations in estrogen metabolism. In the post-

menopausal woman, androstenedione, synthe-sized in the adrenal gland, is the principal estro-gen precursor following the decline of ovarianfunction. Increased conversion of androstened-ione to estrone by fat cells that results in elevatedlevels of this predominant postmenopausalestrogen is reputed to be the underlying explana-tion for the increased risk of breast cancer seenin obese postmenopausal women.47–52 In con-trast, in premenopausal obese women, derange-ment of the estrogen-progesterone balance andsubsequent menstrual disturbances result in adecreased risk of breast cancer.50,53

Male breast cancer is also likely to be relatedto factors resulting in abnormalities of estrogenmetabolism, such as liver disease or geneticdefects such as Klinefelter’s syndrome.53–55

Exogenous Hormones

Oral Contraceptives

Oral contraceptives (OCs) have been marketedextensively over the past 30 to 40 years; world-wide users are estimated to be over 200 millionwomen, and in the United States, it is projectedthat approximately 80 percent of all womenwill have used OCs by the age of 40 years.48

Studies evaluating a possible associationbetween OCs and breast cancer risk have beenhampered by changes in the composition ofOCs over time and by individual patient varia-tion in the duration of use.56–64

Early evidence that OCs could significantlyincrease the risk of breast cancer was reportedby Pike and colleagues in 1983.56 In their case-control study of 314 breast cancer patients <37 years of age and 314 matched controls, theuse of OCs with a relatively high progesteronecontent for more than 6 years and starting useof OCs before age 25 years were associatedwith a relative risk of 4.9 for breast cancerdevelopment. Since that time, many studieshave been conducted in the United Statesattempting to quantify the level of risk ofbreast cancer conferred by the use of OCs. The

24 BREAST CANCER

results of the largest studies are summarized inTable 2–2. Most studies56–61,63 have been case-control studies, in which the rates of use ofOCs among groups of breast cancer patientswere compared with the rates of use of OCsamong matched groups of noncancer patients.The results are relatively inconsistent, withsome studies demonstrating an increased riskof breast cancer associated with use of OCs,whereas others demonstrate a protective effectassociated with the use of OCs. It should benoted that in most studies, the relative risk esti-mate is close to unity, indicating that any effectof OCs is modest in magnitude. However, thehigh incidence of breast cancer in the UnitedStates suggests that even small increases in rel-ative risk could translate into more cases ofbreast cancer.

One of the most widely quoted case-controlstudies on OCs and breast cancer is the Cancerand Steroid Hormone (CASH) Study by theCenters for Disease Control.57 This project, firstreported in 1983, evaluated 689 patients diag-nosed with breast cancer between the ages of 20and 54 years who were identified through theSurveillance, Epidemiology, and End Results(SEER) program. These patients were matchedwith 1,077 controls, and on initial analysis, therisk of breast cancer was lower for women thathad been users of OCs than for women whonever used OCs (relative risk, 0.9). Results of theCASH study were re-evaluated and reported in1991 by Wingo and colleagues.64 In this report,a particular focus was placed on the issue ofOCs having variable, age-related associationswith breast cancer risk. This analysis revealed atrend toward increased risk for users of OCsyounger than 35 years (relative risk, 1.4), and aslight decrease in risk for users of OCs from 45to 54 years of age (relative risk, 0.9).

The Nurses’ Health Study61 provides dataregarding the relative risk of breast canceramong users of OCs followed up in a prospectivefashion. In a cohort of more than 100,000 nurseswith more than 1 million person-years of follow-up, no significant increase in breast cancer risk

was associated with the use of OCs. This relativerisk was not affected appreciably by the durationof use, history of fibrocystic breast disease, orfamily history of breast cancer. These resultswere validated in an updated review of theNurses’ Health Study reported in 1997, withover 1.6 million person-years of follow-up.62

In 1996, a meta-analysis of 54 epidemiologicstudies of OCs and breast cancer was publishedin Lancet.65 This review compiled data on morethan 53,000 breast cancer patients and morethan 100,000 controls from 25 countries. Cur-rent users of OCs had a small but statisticallysignificant increase in the risk of breast cancer(relative risk, 1.24; p < .00001), and this risk didnot persist after 10 years following discontinua-tion of the OCs. The tumors detected in users ofOCs were also found to be of earlier stage thanthose that were detected in women that did notuse OCs. Two theoretical mechanisms couldexplain these findings. One explanation is relatedto the concept of estrogen acting as a promoterrather than as a cause of the neoplastic process.Under this circumstance, it would be expectedthat more tumors would be detected during andfollowing use of OCs because the estrogen con-tent would merely be expediting the clinicalappearance of a pre-existing but previouslyoccult tumor. The second explanation is that


Table 2–2. SUMMARY OF RANDOMIZED STUDIES EVALUATING THE RISK ASSOCIATED WITH THE USE

OF ORAL CONTRACEPTIVES (OCs)

Number of Age Years Relative Study Patients (y) Use Risk

Pike56 314P < 37 ³ 6 4.9314C

CASH57 689P < 55 > 4 1.11,077C

Stadel58 2,088P < 45 > 4 1.12,065C

Miller59 521P < 45 3–4 0.8521C > 7 1.4

Jick60 127P < 43 > 5 0.7174C ³ 10 1.4

Romieu61 >118K < 65 < 1 1.2Cohort 3 0.9

P = patients; C = controls; CASH = Cancer and Steriod HormoneStudy, Centers for Disease Control.

women who are users of OCs are necessarilyreceiving follow-up care, which presumablyincludes surveillance for breast cancer.

Hormone Replacement Therapy

The controversy surrounding hormone replace-ment therapy (HRT) and breast cancer is com-plicated not only by the prevalence of breastcancer but also by the fact that we live in anaging society. The health risks associated withthe postmenopausal estrogen-deficient state,namely, cardiovascular disease and osteoporo-sis, are being faced by increasing numbers ofwomen still in the prime years of their life.

Approximately one quarter of the Americanpopulation is currently over the age of 55 years,and cardiovascular disease is the leading causeof death among postmenopausal women. Car-diovascular disease accounts for nearly threetimes as many deaths as cancer among womenover the age of 65 years.66 Osteoporosis afflictsapproximately 25 million Americans, and it isestimated that half of all women will experi-ence an osteoporotic fracture by the age of 75years. In particular, hip fractures are a majorproblem; they are associated with a 34 percentmortality rate within 6 months, and the corre-sponding health-care costs are several billiondollars annually.67

Menopause also causes several othersymptoms that have a significant adverseimpact on the quality of life. Vasomotor symp-toms are experienced by 80 percent ofmenopausal women; urinary incontinence,vaginal dryness, sleep disturbances, depres-sion, anxiety, and memory losses are beingreported increasingly and have all been relatedto changes in estrogen.66,68–70

Estrogen replacement therapy (ERT) in post-menopausal women has been proved to reverseseveral risk factors for cardiovascular disease,such as low high-density lipoprotein (HDL)cholesterol levels,71 and ERT has been associ-ated with a 40 to 60 percent reduction in com-plications of cardiovascular disease, such as

myocardial infarction and sudden death.66,72–76

Estrogen replacement therapy also reduces ratesof bone resorption by as much as 60 percent,thereby lowering rates of osteoporosis andosteoporotic fractures.66,75,76 Unfortunately,ERT exerts a dose-dependent and duration-of-use–dependent proliferative effect on the uter-ine lining, and several studies have demon-strated an increased rate of endometrial cancerassociated with ERT.66,74,77–79 Results from thePostmenopausal Estrogen/Progestin Interven-tions (PEPI) trial, however, demonstrated thatthe addition of cyclic micronized progesteroneto ERT negated the risk of endometrial hyper-plasia (as measured by baseline and annualendometrial aspiration biopsy), withoutadversely affecting the favorable impact of ERTon the cholesterol profile.71

The effect of ERT on the risk of breast can-cer remains an unresolved question. To date,no prospective, randomized study has beencompleted, and many of the patterns andinconsistencies demonstrated in the availabledata are similar to those seen in the publishedseries of OC use and breast cancer. The resultsof several studies of HRT and breast cancerconducted in the United States and abroad aregiven in Table 2–3. The relative risk esti-mates80–85 vary considerably. A protectiveeffect was seen in the study by Gambrell andcolleagues,81 in which a relative risk of 0.3 was

26 BREAST CANCER

Table 2–3. HORMONE REPLACEMENT THERAPY:BREAST CANCER RISK IN THE GENERAL POPULATION

Number Mean Numberof of Years Relative

Study Year Women Follow-Up Risk

Hoover80 1976 1,891 12 1.3

Gambrell81 1983 5,563 7 0.4

Hunt82 1987 4,544 6 1.59

Bergkvist83 1989 23,244 5.7 1.1

Mills84 1989 20,341 6 1.7

Colditz85 1995 23,965 16 1.321

associated with HRT (combined estrogen andprogestin) among more than 5,500 women fol-lowed up at the Wilford Hall United States AirForce Medical Center. In this study, the relativerisk estimate was calculated by comparing theobserved breast cancer incidence with the inci-dence expected on the basis of the SEER pro-gram. Mills and colleagues84 followed up acohort of more than 20,000 Seventh DayAdventist women and found a relative risk of1.7 associated with any history of HRT use;this relative risk increased to 2.5 for currentHRT users, and to 2.8 for HRT users with ahistory of benign breast disease. In 1995,Colditz and colleagues85 reported on findingsfrom the Nurses’ Health Study. A significantincrease in the relative risk of breast cancerassociated with current HRT use (relative risk,1.32) was demonstrated.

Updated results of the Nurses’ Health Studywere published in 1997.86 At the time of theupdated report, more than 120,000 registerednurses had been followed up with biannualquestionnaires since 1976, and more than 3,600deaths had been documented. Several impor-tant findings were reported. Use of HRT wasassociated with significant decreases in mortal-ity compared with nonuse (relative risk ofdeath, 0.63). This benefit of HRT was strongestfor women who had pre-existing risk factors foratherosclerotic heart disease (eg, currenttobacco use, parental history of prematuremyocardial infarction, diabetes, or hyperten-sion). For women considered to be at low risk

for cardiac disease, there was a lesser benefitfrom HRT (relative risk of death, 0.89). After10 years of HRT, however, mortality ratesbegan to rise, predominantly from an increas-ing rate of breast cancer–related deaths.

In conclusion, the use of ERT and OCsappears to be associated with an increased riskof breast cancer, particularly in youngerwomen.87 In the presence of atheroscleroticdisease and osteoporosis, the use of ERTseems to reduce the incidence of severe com-plications (eg, myocardial infarction and bonefractures) and appears to be associated withimproved survival. Recommendation for rou-tine use should be restricted to a select groupof women, and the decision on treatmentshould be based on a detailed evaluation of therisk/benefit ratio.

BREAST CANCER CHEMOPREVENTIVE AGENTS

The importance of estrogens as breast cancerpromoters has been sustained by direct andindirect observations since the 1960s. Animportant source of indirect evidence is pro-vided by the relevant clinical observationsderived from clinical trials of the adjuvant useof tamoxifen. Several large randomized studieshave demonstrated that adjuvant therapy withthe nonsteroidal antiestrogen tamoxifen citrateis associated with reduction in the risk of devel-oping a second contralateral primary breastcancer by 30 to 50 percent.88–90


Table 2–4. SUMMARY OF ONGOING RANDOMIZED CHEMOPREVENTIVE TRIALS

Number Study Agent of Patients Patient Population

BCPT-P1 (USA)94 TAM vs. placebo 13,388 High-risk, pre-/postmenopausalRoyal Marsden (UK)93 TAM vs. placebo 2,012 High-risk pre-/postmenopausalNational Tumor Institute (ITALY)96 TAM vs. placebo 5,408 Post-hysterectomy

No risk of breast cancer, pre-/postmenopausalMD Anderson Cancer Center (USA)115 TAM +/-4-HPR Ongoing Ductal carcinoma in situ

(pre-operative)National Tumor Institute (ITALY) 94112 4-HPR 2,972 History of breast cancerSTAR (USA)94 TAM vs. Raloxifene Ongoing High-risk, pre-/postmenopausal

BCPT = Breast Cancer Prevention Trial; NSABP = National Surgical Adjuvant Breast and Bowel Project; TAM = tamoxifen; 4-HPR = fenretinide(N-[4-hydroxyphenyl] retinamide, 4-HPR); STAR = Study of Tamoxifen and Raloxifene.

These observations confirm the centralrole for estrogens in the promotion of breastcancer and suggest an opportunity for devel-oping a preventive strategy based on usingselective antiestrogens to modulate estrogenicactivity.

Preclinical data and clinical observationsderived from clinical chemoprevention trialsconducted in other malignancies indicated apotential activity of retinoids as chemopreven-tive agents. The major ongoing chemopreven-tion trials are listed on Table 2–4.

Tamoxifen

Tamoxifen is a nonsteroidal triphenylethylenederivative, which is generally classified as anantiestrogen with partial estrogen-agonist activ-ity in some tissues. In fact, results from chemo-preventive and adjuvant trials suggest that treat-ment with tamoxifen is associated with anincrease in bone mineral density and decreasedserum cholesterol, particularly in postmeno-pausal women.91,92

More importantly, the use of adjuvanttamoxifen following primary surgery for estro-gen-sensitive early breast cancer has been asso-ciated with prolonged disease-free survival andreduction in the risk of death by 20 to 30 per-cent. One of the major arguments in favor ofthe use of tamoxifen as a chemopreventiveagent is derived from the observation of a sig-nificant reduction in the incidence of primarytumors in the contralateral breast in womentreated with adjuvant tamoxifen.88,89,93,94

The National Surgical Adjuvant Breast andBowel Project (NSABP) and other Europeanstudies showed a reduction between 40 and 50percent in the incidence of primary tumors in thecontralateral breast among women taking tamox-ifen as adjuvant therapy.88,91,92 Among the Euro-pean trials, the Stockholm trial, designed to eval-uate the efficacy and toxicity of adjuvanttamoxifen (40 mg daily) for 2 years in post-menopausal patients (23 percent older than 50years) with unilateral breast cancer, deserves par-

ticular mention.90 This trial demonstrated a sig-nificant decrease in the incidence of contralateralbreast cancer but also a six-fold increase in theincidence of endometrial cancer and an unex-pected excess of gastrointestinal malignancies.

Of particular interest is the overview analy-sis of the major randomized trials of adjuvanttamoxifen among nearly 30,000 women withearly breast cancer.95 A recent update of thisanalysis demonstrated a reduction in the inci-dence of contralateral breast cancer incidence of47 percent with 5 years of treatment. The pro-portional reduction in contralateral breast cancerappeared to be unrelated to the estrogen receptor(ER) status of the original tumor. The treatmentappeared to be associated with a significantincrease in the incidence of endometrial cancerand a slight, not significant increase in colorec-tal cancer (larger with only 1 year of tamoxifen).

In 1986, a pilot study was started in theUnited Kingdom to test the feasibility and tox-icity associated with long-term tamoxifen treat-ment in women at high risk of breast cancer.93

Between October 1986 and June 1993, a totalof 2,012 women were accrued and randomlyassigned to tamoxifen (20 mg/day) or placebofor up to 8 years. A total of 265 women were onHRT at entry and 131 were randomized totamoxifen treatment. With a median follow-upof 36 months and with a compliance of 77 per-cent of the women assigned to the treatmentarm, no obvious effect on bone mineral densitywas observed and only marginal effects on clot-ting factors. Tamoxifen was associated with asignificant reduction in the serum cholesterollevel. More importantly, there was an increasedincidence of uterine fibromata and benign ovar-ian cysts; however, no increase in endometrialcancer incidence was reported.

On the basis of these encouraging data, in1992, a large, multicenter, randomized, double-blind trial funded by the National Cancer Insti-tute (NCI) was begun to test whether tamoxifencould prevent breast cancer in high-riskwomen.94 The population at risk was defined,taking into account the following risk factors,

28 BREAST CANCER

as indicated by the Gail model: age, age atmenarche and at first live birth, number offirst-degree relatives affected, and, finally, thenumber of previous breast biopsies and thepresence or absence of atypical hyperplasia.

The Breast Cancer Prevention Trial (BCPT-P1) enrolled 13,388 women older than 35 yearsbetween April 1992 and September 1997. Theresearch, coordinated by the NSABP, involvedmore than 300 centers across the United Statesand Canada. The study was closed and prelimi-nary results released 14 months earlier thanplanned.94 In the median average follow-up timeof 54.6 months, 89 cases of invasive breast can-cer occurred in the group of women assigned totamoxifen treatment (6,681 women) comparedwith 175 cases in the group assigned to placebo(6,707 women), corresponding to a 49 percentrisk reduction. There was also a 50 percentreduction in the incidence of noninvasive breastcancer. Tamoxifen reduced the occurrence ofER-positive tumors by 69 percent, but no effi-cacy was seen in the prevention of ER-negativetumors. The risk reduction was not age depen-dent; the risk reduction in women aged 49 yearsor younger was 44 percent, and it was 55 percentin women older than 60 years. The NCI and theEndpoint Review, Safety Monitoring, and Advi-sory Committee agreed that the participants andtheir physicians should be told which treatmenthas been assigned because of the clear evidencethat tamoxifen reduced breast cancer risk.

The Breast Cancer Prevention Trial was alsodesigned to evaluate the possible benefit oftamoxifen in reducing cardiac events and osteo-porosis-related complications. There was nodifference between the tamoxifen group and theplacebo group in the number of heart attacks,whereas there was a 19 percent reduction in theincidence of fractures of the hip, wrist, andspine (111 cases in the tamoxifen group versus137 cases in the placebo group). Treatmentwith tamoxifen was associated with anincreased incidence of endometrial cancer (33cases versus 14 cases in the placebo group), inparticular in women aged 50 years or older. A

slight increase in the incidence of deep veinthrombosis and pulmonary embolism in thetamoxifen group was also reported; all theseevents were more frequently observed inwomen older than 50 years of age. Because ofthese reported complications, the decisionabout whether to use tamoxifen as a chemopre-ventive agent must be carefully weighed forevery patient on the basis of an accurate evalu-ation of the patient’s age group, personal breastcancer risks, and comorbid conditions.

Similar trials of tamoxifen for chemopreven-tion have been conducted in Italy and the UnitedKingdom.96 The Italian study has completedaccrual of 5,408 women who have had hysterec-tomy and have no factors associated withincreased risk of breast cancer. Preliminaryanalysis at a median follow-up of 46 months didnot show any difference in breast cancer inci-dence in the tamoxifen group compared withthe placebo-control group.96 Differences in thestudy populations, age distributions, history ofHRT, and family history may account for theinability of studies to confirm the effectivenessof tamoxifen for chemoprevention.

Selective Estrogen Receptor Modulators

In the past decade, the reports of significantside effects associated with the prolonged useof tamoxifen have stimulated research directedtoward the development of other selectiveestrogen receptor modulators (SERMs). Amongthe various products investigated, raloxifenehas demonstrated antitumor activity and afavorable toxicity profile and is being furtherinvestigated.97–99

Preclinical data have shown that raloxifene,an antiestrogen with no estrogen-agonist effecton the uterus, inhibits mammary carcinogenesisin a rat model of breast cancer in a manner sim-ilar to tamoxifen when raloxifene is used incombination with 9-cis retinoic acid.97 Clinicaltrials have been started in an attempt to establishthe role of raloxifene in preventing osteoporosisin postmenopausal women, and preliminary


results from two randomized clinical trials haverecently become available. The Multiple Out-comes of Raloxifene Evaluation (MORE) trialwas specifically designed to evaluate the possi-bility of reducing the risk of fractures in post-menopausal women receiving raloxifene; amarkedly reduced risk of newly diagnosed breastcancer was demonstrated with raloxifene com-pared with placebo (0.21% versus 0.82%).100

Jordan and colleagues recently reported theresults of a multicenter, double-blind random-ized trial conducted in about 12,000 women.Treatment with raloxifene was associated with a58 percent reduction in the risk of developingprimary breast cancer.101 These results havestimulated the design of a second major breastcancer prevention trial, the Study of Tamoxifenand Raloxifene (STAR) that will compare thetoxicity, risks, and benefits of raloxifene withthose of tamoxifen. Women enrolled in thestudy will be randomly assigned to receiveeither 20 mg of tamoxifen or 60 mg of ralox-ifene for 5 years, with follow-up planned for anadditional 2 years.

A number of other SERMs have emerged(eg, toremifene, trioxifene, droloxifene, TAT-59)for clinical use.102–105 The majority of thesedrugs are presently in phase I to II clinical tri-als and have already demonstrated their clinicalactivity in the management of breast cancer.They represent possible candidates for futurechemoprevention studies.

Retinoids

Retinoids are a family of natural and syntheticcompounds structurally related to vitamin A.They are a group of molecules capable of influ-encing many biologic functions, such as prolif-eration, differentiation, and induction of apop-tosis.106–109 Retinoids function via two types ofnuclear receptors, the retinoid alpha-receptors(RARs) and the retinoid X receptors (RXRs),each of which is encoded by three genes.93

Preclinical data have demonstrated that carcinogen-induced mammary carcinomas are

sensitive to the antiproliferative effects ofretinoids. Enhanced sensitivity has been shownby estrogen receptor–positive cell lines, whereasestrogen receptor–negative cell lines have shownminimal sensitivity to these compounds.110,111

The mechanisms of the antiproliferative effectare still being investigated. At least, in somecell lines, apoptosis, instead of differentiation,seems to be the prevalent mechanism of growthinhibition.111

A synthetic retinoid, fenretinide (N-[4-hydroxyphenyl] retinamide, 4-HPR), has demon-strated the capacity to inhibit the growth ofbreast cancer cell lines and chemically inducedmammary tumors in rats, without the toxicityassociated with other retinoids. This compoundwas the first retinoid to be tested in clinical tri-als and has been proved to be well tolerated ata daily dose of 200 mg with a 3-day monthlydrug holiday.112,113

An Italian prospective randomized trialdesigned to evaluate the role of fenretinide inreducing the incidence of contralateral breastcancer began in March 1987. In July 1993,accrual of 2,972 women, age 30 to 70 years,with history of T1-2, N0 breast cancer wascompleted.112 Preliminary data suggested thatfenretinide can reduce the incidence of ovariancarcinomas.113 Though safer than otherretinoids in experimental models, fenretinideproduced visual (dark adaptation) and ophthal-mologic complaints (ocular dryness, lacrima-tion, conjunctivitis, photophobia) in 20 percentand 8 percent of women, respectively, at 5years.114 Such effects are thought to be causedby the reduction of plasma retinal levels, whichoccurs after administration of the retinoid.

In animal models, 9-cis-retinoic acid incombination with antiestrogens (tamoxifen orraloxifene) resulted in effective chemopreven-tion of a rat model of breast cancer induced bythe carcinogen nitrosomethylurea.97,111

The University of Texas M. D. AndersonCancer Center is presently investigating the roleof tamoxifen and fenretinide in reducing the riskof invasive breast cancer in patients diagnosed

30 BREAST CANCER

with DCIS.115 This phase II trial is accruingwomen presenting with small breast lesions andmammographic calcifications suspicious formalignancy. After histologic confirmationthrough core biopsies, participants are randomlyassigned to receive tamoxifen, fenretinide, or acombination of both. The treatment is plannedfor 3 weeks before definitive surgery. An impor-tant objective of this trial is to perform a detailedquantitative assessment of biomarkers to be usedas surrogate end points. The proposed SEBsinclude estrogen and progesterone receptors,nuclear retinoid receptors, transforming growthfactor (TGF)-b, HER-2/neu, proliferation (Ki-67immunostaining), angiogenesis (factor VIII)markers, and chromosomal aberrations. Thistrial is expected to provide insight into the bio-logic mechanisms of antiestrogens and retinoids,or the combination of both, in reducing the pro-gression of DCIS to invasive cancer.

Dietary Interventions

Epidemiologic observations of large interna-tional differences in the incidence of breastcancer have provided a basis for formulatinghypotheses on a possible relation between dietand the development of cancer. The age-adjusted incidence of breast cancer varies from22 per 100,000 in Japan to 68 per 100,000 inthe Netherlands.116 The ratio of breast cancermortality between the United States and Japanis 3:1 for premenopausal women and 8:1 forpostmenopausal women.117 These importantdifferences may possibly be related to fat intakeand total calories in the diet. Clinical data col-lected from case-control studies have demon-strated a positive correlation between diets highin fat and meat and breast cancer.118–122 Experi-mental studies have shown that omega-6polyunsaturated fatty acids (PUFAs) containedin high-fat diets promote both mammarytumorigenesis and cell proliferation in chemi-cally induced mammary tumors, whereasomega-3 PUFAs, contained in fish oil, caninhibit these effects.119,120

Heterocyclic amines, a group of mutageniccompounds identified in cooked foods, seem tobe related to the increased risk of breast cancerassociated with high intake of well-done meat.Recently, a case-control study among 41,836women demonstrated that women who con-sumed well-done meats, including hamburger,beef steak, and bacon, had higher adjusted oddsratios for breast cancer (up to 4.62), if they con-sumed all three different meats well done.121

These data have provided the rationale fordiet interventions, consisting of a low-fat dietand fish-oil supplements, that have been foundto be able to produce increases in total omega-3 PUFAs in adipose tissue and in the ratio ofomega-3/omega-6 PUFAs in patients withbreast cancer.120 The Canadian Diet and BreastCancer Prevention Study Group has conducteda multicenter randomized trial involvingwomen with breast densities detected on mam-mography and showed that after 2 years of alow-fat diet, with less than 15 percent of calo-ries from fat, there was a significant reductionin the number of radiographic abnormalities.122

Adjuvant dietary recommendations of 15 per-cent of calories from fat for women with post-menopausal breast cancer are currently beingevaluated in the Women’s Intervention Nutri-tion Study and in the Women’s Healthy Eatingand Living Study.123

The role of alcohol consumption and smok-ing are also being extensively investigated aspossible risk factors for breast cancer. Whilethe majority of the studies has documented thathigh alcohol intake is associated with a signifi-cant increased incidence of breast cancer, nodefinitive pathogenetic role for active or pas-sive smoking has been demonstrated.124,125

The use of natural products contained inessential oils and soy-based products, forexample, the monoterpenes limonene andperillyl alcohol and the isoflavonoid genis-tein, all showed preclinical evidence of tumorregression.126–129 The effects of limonene andlimonene-related monoterpenes, perillyl alco-hol and perillic acid, on cell growth, cell cycle


progression, and expression of cyclin D1 hasbeen investigated in T-4D, MCF-7, MDA-MB-231 breast cancer cell lines. The resultsrevealed that limonene-related monoterpenescaused a dose-dependent inhibition of cell pro-liferation. Of the three monoterpenes tested,perillyl alcohol was the most potent andlimonene was the least potent inhibitor of cellgrowth. Growth inhibition induced by perillylalcohol and perillic acid was associated with afall in the proportion of cells in the S phase,accumulation of cells in the G1 phase, and adecrease in cyclin D1 mRNA levels.128 Thepotential preventive role of genistein, a compo-nent of soy, has been evaluated in rats. Pharma-cologic doses of genistein given to immaturerats enhance mammary gland differentiation,resulting in a significantly less proliferategland that is not as suspectible to mammarycancer.129 These components are presentlybeing tested in several clinical chemopreven-tive studies.

SURGICAL PROPHYLAXIS AND MODULATION OF RISK

Prophylactic Mastectomy

It seems intuitive that mastectomy would be aneffective means of preventing breast cancer,especially in this era of immediate reconstruc-tion techniques that produce cosmeticallyacceptable results. However, animal models aswell as clinical data from trials in humans haveconfirmed that this is not always the case. Stud-ies have demonstrated that even total mastec-tomy (defined as removal of the entire breast,including the nipple-areolar complex, but spar-ing the axillary contents) is frequently incom-plete; microscopic amounts of breast tissuemay be left in the skin flaps, attached to thepectoralis fascia and extending into the axilla.Temple and colleagues130 evaluated 10 prophy-lactic mastectomy specimens from 5 patientsconsidered to be at high risk for developingbreast cancer. In this study, random frozen sec-

tion analyses (approximately 700 per breast) ofthe margins identified 3 cases of breast tissueextending into the pectoralis fascia, 1 case ofpectoralis muscle involvement, 2 cases of infe-rior skin flap involvement, and 1 case of axil-lary tissue involvement. It would be expectedthat prophylactic subcutaneous mastectomy(defined as removal of all gross breast tissue,usually via an inframammary incision, butsparing the nipple-areolar complex) wouldresult in additional breast tissue left on the skinflaps of the nipple-areolar complex.

The clinical significance of retained breasttissue in the setting of prophylactic mastec-tomy for humans is not yet defined. In therodent model of mammary tumors and prophy-lactic mastectomy, it is clear that the extent ofbreast tissue removed does not clearly corre-late with the extent of protection against breastcancer. Wong and colleagues131 performedvarying degrees of partial versus total mastec-tomy versus sham surgery in a series of rats,either before or after administration of the car-cinogen DMBA; at 8 months of age, there wereno significant differences in the number of car-cinogen-induced tumors between any of thegroups. Using a mouse model with a high inci-dence of spontaneous mammary tumor devel-opment (and therefore theoretically more sim-ilar to the human experience of spontaneousbreast cancer incidence) Nelson and col-leagues132 similarly found no difference in thenumber of tumors that developed in mice thatunderwent either sham surgery, 50 percentmastectomy, or total mastectomy.

Data on prophylactic bilateral mastectomy inhumans are limited. The often-cited studies byPennisi and Capozzi133 and Woods andMeland134 in the plastic surgery literature eachreported on at least 1,500 women who under-went subcutaneous mastectomies, and in bothstudies, the subsequent incidence of breast car-cinoma was less than 1 percent. However, boththese studies have been criticized for their lim-ited applicability to truly high-risk women,since many of the prophylactic procedures were

32 BREAST CANCER

performed in women who would be consideredat the present time to be at only low or interme-diate risk for breast cancer. Detailed follow-upinformation is also lacking in these large series.

It remains to be determined if a prophylac-tic mastectomy is clearly indicated in women athigh risk of developing breast cancer. Womenwith BRCA1 mutations, who may have a cumu-lative breast cancer risk of 40 to 85 percent,would be the obvious candidates.

Schrag and colleagues135 developed a statis-tical decision model to calculate the benefitthat BRCA1 or BRCA2 mutation carriers mightderive from prophylactic mastectomy. Using arisk-reduction estimate of 85 percent associatedwith prophylactic mastectomy, this study deter-mined that a 30-year-old BRCA1 mutation car-rier would gain 2.9 to 5.3 years of lifeexpectancy following preventive surgery.Lynch and colleagues136 reported on the resultsof a series of women who had undergone exten-sive genetic counseling and subsequently testedpositive for BRCA1 gene mutations. Only 35percent of these patients said they would con-sider undergoing prophylactic mastectomy.This finding underscores the complexity ofidentifying high-risk women who will benefitpsychologically as well as clinically from pre-ventive surgery.

In addition, there are many case reports doc-umenting the failure of prophylactic total mas-tectomy to protect against breast cancer.137–141

However, very intriguing data were recentlyreported by Hartmann and colleagues from theMayo Clinic.142 A retrospective analysis wasperformed on 639 women with moderate orhigh risk for breast cancer (by family history)that had undergone bilateral prophylactic sub-cutaneous mastectomy between 1960 and 1993.The breast cancer incidence in these womenwas compared with the number of expectedcases based on the Gail model and with thenumber of cases that occurred among femalesiblings who had not undergone prophylacticsurgery; these estimations revealed an approxi-mately 90 percent reduction in breast cancer

risk associated with prophylactic mastectomy. Another patient population that might be

considered to be suitable for prophylactic mas-tectomy is women with a history of unilateralbreast cancer, in whom the risk of subsequentcontralateral breast cancer is approximately 0.5to 0.7 percent per year.143 However, it has beenargued that the risk of death from the primarycancer is still greater than the risk of develop-ing a second primary cancer for the majority ofbreast cancer patients, making the survival ben-efit of prophylactic surgery questionable.144 Onthe other hand, the rationale has also beenoffered that optimal immediate reconstructioncosmesis can be attained with bilateral trans-verse rectus abdominis myocutaneous (TRAM)flaps. To address this issue Kroll and col-leagues145 studied 88 patients with unilateralbreast cancer who had undergone bilateral mas-tectomies with immediate breast reconstruc-tion. Previously unsuspected invasive breastcancer was found in 3.4 percent of the con-tralateral mastectomy specimens.

The Society of Surgical Oncology has delin-eated categories of patients for whom prophy-lactic mastectomy may reasonably be consid-ered on the basis of clinical features (and notincluding genetic testing results).144 For womenwith no history of breast cancer, the indicationsinclude atypical hyperplasia, family history ofpremenopausal bilateral breast cancer, anddense, nodular breasts associated with atypicalhyperplasia. For women with a known unilateralbreast cancer, the indications for consideringcontralateral prophylactic mastectomy includediffuse microcalcifications, LCIS, a large, diffi-cult-to-evaluate breast, history of LCIS, andfamily history of early-onset breast cancer.

Prophylactic Oophorectomy and/or Hysterectomy

Several studies have documented lower breastcancer incidence among women who underwentoophorectomy at a young age. The effect of hys-terectomy on breast cancer risk is less clear, but


it has been postulated that hysterectomy mayhave some secondary effects by affecting ovar-ian blood flow and ovulation. Schairer and col-leagues evaluated 15,844 women undergoingsurgery in the Uppsala health care region ofSweden and found a 50 percent reduction inbreast cancer risk in those women who under-went bilateral oophorectomy prior to age 50years, compared with the risk of the backgroundpopulation.146 Hysterectomy alone had no con-sistent association with change in breast cancerrisk. In a case-control series from Italy, womenwho underwent premenopausal oophorectomywith hysterectomy or hysterectomy alone hadreduced relative risk of developing breast cancer(0.8 and 0.7, respectively).147 However, giventhe importance of the ovarian function in main-taining cardiovascular and bone health, there arepresently no indications for recommendingthese procedures as prophylaxis against breastcancer in any subset of patients.

CONCLUSIONS

The recently reported encouraging results withthe use of tamoxifen and the ongoing clinical tri-als introducing SERMs (raloxifene), retinoids,and other approaches suggest an increasingawareness in physicians of the field of chemo-prevention and its potentially enormous socio-economic implications. Breast cancer chemo-prevention is a field in constant evolution andhas the potential to significantly affect the livesof thousands of women by reducing their risk ofbreast cancer.

In the last decade, the increasing informationavailable on the differential contribution ofgenetic, dietary, and environmental factors hasgreatly improved our ability to determine theabsolute breast cancer risk for every woman andproperly select high-risk groups for interven-tional studies. Interestingly, the concomitantevaluation of histopathologic factors and mul-tiple biomarkers has offered the opportunity toincrease our understanding of the important bio-logic modifications associated with tumor pro-

gression. In the future, prospective clinical trialsof chemoprevention strategies should be limitedto high-risk populations identified on the basisof a combination of epidemiologic, histopatho-logic, and genetic data and should make use ofSEBs to evaluate the efficacy of drug interven-tions. This approach will contribute greatly toreducing the patient population under study,eventually reducing the costs related to theseinvestigations and helping to clarify the biologyof each drug’s mechanism of action.

The initial results of the BCPT-P1 havedemonstrated, for the first time, the possibilityof reducing the risk of breast cancer in a high-risk group of women, with a marginal toxicity.The ongoing STAR preventive trial is designedto determine if raloxifene has a chemopreven-tive efficacy comparable with tamoxifen withless associated toxicity. In the meantime, alonger follow-up of the BCPT-P1 trial will clar-ify if the treatment with tamoxifen represents atrue preventive intervention or a treatment forpreclinical conditions with consequent delay-ing of the onset of invasive breast cancer.

The potential role of dietary intervention inmodifying the risk of breast cancer is probablypresently underestimated; the ongoing clinicaltrials may contribute essential information totheir potential clinical applicability.

In conclusion, more attention should bedirected to the biologic relationships amonghormone modulation, diet, and the risk ofbreast cancer to develop an “ideal lifestylemodel” to propose for the high-risk groups. Inthis context, the role of prophylactic surgery,with the psychologic consequences related tothe change in body image, even if associatedwith improved outcome in high-risk women,will come to be considered as a secondary,rather than a primary, option for breast cancerrisk management.

REFERENCES

1. Chu KC, Tarone RE, Kessler LG, et al. Recenttrends in U.S. breast cancer incidence, survival,

34 BREAST CANCER

and mortality rates. J Natl Cancer Inst 1996;88:1571–9.

2. Vogel VG. Subjects and recruitment strategies fora short-term phase II chemoprevention trial ofbreast cancer using surrogate endpoint bio-markers. J Cell Biochem 1993;17G:257–8.

3. Page DL, Dupont WD, Rogers LW. Atypicalhyperplastic lesions of the female breast. Along-term follow-up study. Cancer 1985;55:2698–708.

4. Dupont WD, Page DL. Relative risk of breast can-cer varies with time since diagnosis of atypicalhyperplasia. Hum Pathol 1989;20:723–5.

5. Hill AD, Doyle JM, McDermott EW. Hereditarybreast cancer. Br J Surg 1997;84:1334–9.

6. Tseng SL, Yu IC, Yue CT, et al. Allelic loss atBRCA1, BRCA2, and adjacent loci in relationto TP53 abnormality in breast cancer. GenesChromosomes Cancer 1997;20:377–82.

7. Sporn MB, Newton DL. Chemoprevention of can-cer and retinoids. Fed Proc 1979;38:2528–34.

8. Kelloff GJ, Boone CW, Steele VE. Progress incancer chemoprevention: perspectives on agentselection and short-term clinical interventiontrials. Cancer Res 1994;54:2015S–24S.

9. Kelloff GJ, Boone CW, Crowell JA, et al. Surro-gate endpoint biomarkers for phase II cancerchemoprevention trials. J Cell Biochem Suppl1994;19:1–9.

10. Grizzle WE, Myers RB, Arnold MM, Srivastava S.Evaluation of biomarkers in breast and prostatecancer. J Cell Biochem Suppl 1994;19:259–66.

11. Fabian CJ, Kamel S, Zalles C, Kimler BF. Identi-fication of a chemoprevention cohort from apopulation of women at high risk for breastcancer. J Cell Biochem 1996;25S:112–22.

12. Gail MH, Brinton LA, Byar DP, et al. Projectingindividualized probabilities of developingbreast cancer for white females who are beingexamined annually. J Natl Cancer Inst 1989;81:1879–86.

13. Benichou J, Gail MH, Mulvihill JJ. Graphs to esti-mate an individualized risk of breast cancer. J Clin Oncol 1996;14(1):103–10.

14. Smith J. New computer program assesses awoman’s risk for developing breast cancer. J Natl Cancer Inst 1998;90(18):1332.

15. Lynch HT, Krush AJ. The cancer family syndromeand cancer control. Surg Gynecol Obstet 1971;132:247–50.

16. Marcus JN, Watson P, Page DL, et al. Hereditarybreast cancer: pathobiology, prognosis, andBRCA1 and BRCA2 gene linkage. Cancer1996;77:697–709.

17. Ford D, Easton DF, Petro J. Estimates of the genefrequency of BRCA1 and its contribution tobreast and ovarian cancer incidence. Am JHum Genet 1995;57:1457–62.

18. Struewing JP, Hartge P, Wacholder S, et al. Therisk of cancer associated with specific muta-tions of BRCA1 and BRCA2 among Ashke-nazi Jews. N Engl J Med 1997;336:1401–8.

19. Shattuck-Eidens D, Oliphant A, McClure M, et al.BRCA1 sequence analysis in women at highrisk for susceptibility mutations. JAMA 1997;278:1242–50.

20. Couch FJ, DeShano M, Blackwood MA, et al.BRCA1 mutations in women attending clinicsthat evaluate the risk of breast cancer. N Engl JMed 1997;336:1409–15.

21. Iglehart JD, Miron A, Rimer BK, et al. Overesti-mation of hereditary breast cancer risk. AnnSurg 1998;228(3):375–84.

22. Krainer M, Silva-Arrieta S, FitzGerald MG, et al.Differential contributions of BRCA1 andBRCA2 to early onset breast cancer. N Engl JMed 1997;336:1416–21.

23. Prechtel D, Werenskiold AK, Prechtel K, et al.Frequent loss of heterozygosity at chromo-some 13q12-13 with BRCA2 markers in spo-radic male breast cancer. Diagn Mol Pathol1998;7(1):57–62.

24. Malkin D, Li FP, Strong LC, et al. Germ line p53mutations in a familial syndrome of breast can-cer, sarcomas, and other neoplasms. Science1990;250:1233–8.

25. Swift M, Morrell D, Massey RB, Chase DL. Inci-dence of cancer in 161 families affected byataxia-telangiectasia. N Engl J Med 1991;325:1831–6.

26. Tokunaga M, Land CE, Yammoto T, et al. Inci-dence of female breast cancer among atomicbomb survivors, Hiroshima and Nagasaki,1950–1980. Radiat Res 1987;112:243–72.

27. Mackenzie I. Breast cancer following multiplefluoroscopies. Br J Cancer 1965;19:1–8.

28. Miller AB, Howe GR, Sherman GJ, et al. Mortal-ity from breast cancer after irradiation duringfluoroscopic examinations in patients beingtreated for tuberculosis. N Engl J Med 1989;321:1285–9.

29. Shore RE, Hildreth N, Woodard E, et al. Breastcancer among women given x-ray therapy foracute postpartum mastitis. J Natl Cancer Inst1986;77:689–96.

30. Hildreth NG, Shore RE, Dvoretsky PM. The riskof breast cancer after irradiation of the thymusin infancy. N Engl J Med 1989;321:1281–4.


31. Goss PE, Sierra S. Current perspectives on radia-tion-inducted breast cancer. J Clin Oncol 1998;16(1):338–47.

32. van Leeuwen FE, Klockman WJ, Hagenbeek A, etal. Second cancer risk following Hodgkin’s dis-ease: a 20-year follow-up study. J Clin Oncol1994;12:312–25.

33. Kaldor JM, Day NE, Brand P, et al. Second malig-nancies following testicular cancer, ovariancancer and Hodgkin’s disease: an internationalcollaborative study among cancer registries. IntJ Cancer 1987;39:571–85.

34. Tucker MA, Coleman CN, Cox RS, et al. Risk ofsecond cancers after treatment for Hodgkin’sdisease. N Engl J Med 1988;318:76–81.

35. Hancock SL, Tucker MA, Hoppe RT. Breast can-cer after treatment of Hodgkin’s disease. J NatlCancer Inst 1993;85:25–31.

36. Sankila R, Garwicz S, Olsen JH, et al. Risk ofsubsequent malignant neoplasm among 1641Hodgkin’s disease patients diagnosed in child-hood and adolescence: a population-basedcohort study in the five Nordic countries. JClin Oncol 1996;14:1442–6.

37. Aisenberg AC, Finkelstein DM, Doppke KP, et al.High risk of breast carcinoma after irradiationof young women with Hodgkin’s disease. Can-cer 1997;79(6):1203–10.

38. Feig SA, Hendrick RE. Radiation risk fromscreening mammography of women aged 40-49 years [Monogr]. J Natl Cancer Inst 1997;(22):119–24.

39. Marshall CJ, Schumann GB, Ward JH, et al. Cyto-logic identification of clinically occult prolifer-ative breast disease in women with a familyhistory of breast cancer. Ann Pathol 1991;95:157–65.

40. McKittrick R, Fabian C, Kamel S, et al. Dysplasiaand other biomarker abnormalities as potentialsurrogate endpoint biomarkers in breastchemoprevention trials [abstract]. Proc AmSoc Clin Oncol 1995;14:A348.

41. Alpers CE, Wellings SR. The prevalence of carci-noma in situ in normal and cancer-associatedbreast. Hum Pathol 1985;16:796–807.

42. Frykberg ER, Bland KI. In situ breast carcinoma.Adv Surg 1993;26:29–72.

43. Eriksson ET, Schimmelpenning H, Aspenblad U,et al. Immunohistochemical expression of themutant p53 protein and nuclear DNA contentduring the transition from benign to malig-nant breast disease. Hum Pathol 1994;25:1228–33.

44. Rohan TE, Hartwick W, Miller AB, Kandel RA.Immunohistochemical detection of c-erbB-2and p53 in benign breast disease and breastcancer risk. J Natl Cancer Inst 1998;90:126–9.

45. Beatson GT. On the treatment of inoperable casesof carcinoma of the mamma: suggestions for anew method of treatment with illustrativecases. Lancet 1986;2:104–7.

46. Kelsey JL, Gammon MD. The epidemiology ofbreast cancer. CA Cancer J Clin 1991;41:147–65.

47. Hulka BS, Liu ET, Lininger RA. Steroid hor-mones and risk of breast cancer. Cancer 1994;74:1111–24.

48. Russo J, Tay LK, Russo IH. Differentiation of themammary gland and susceptibility to carcino-genesis. Breast Cancer Res Treat 1982;2:5–73.

49. Henderson BE, Ross R, Bernstein L. Estrogens asa cause of human cancer: the Richard andHilda Rosenthal Foundation Award Lecture.Cancer Res 1988;48:246–53.

50. Pike MC, Spicer DV, Dahmoush L, Press MF.Estrogens, progestogens, normal breast cellproliferation, and breast cancer risk. EpidemiolRev 1993;15:17–35.

51. Spicer DV, Pike MC. Breast cancer preventionthrough modulation of endogenous hormones.Breast Cancer Res Treat 1993;28:179–93.

52. Dickson RB, Thompson EW, Lippman ME. Reg-ulation of proliferation, invasion, and growthfactor synthesis in breast cancer by steroids. JSteroid Biochem Mol Biol 1990;37:305–16.

53. Cauley JA, Gutal JP, Kuller LH, et al. The epidemi-ology of serum sex hormones in postmenopausalwomen. Am J Epidemiol 1989;129:1120–31.

54. Moore M. Male breast cancer. In: Harris JR, Lipp-man ME, Morrow M, Hellman S, editors. Dis-eases of the breast. Philadelphia: Lippincott-Raven; 1996. p. 859.

55. Committee on the Relationship Between OralContraceptives and Breast Cancer. Institute ofMedicine, Division of Health Promotion andDisease Prevention. Oral contraceptives andbreast cancer. Washington, D.C.: NationalAcademy Press; 1991.

56. Pike MC, Henderson BE, Krailo MD, et al. Breastcancer in young women and use of oral contra-ceptives: possible modifying effect of formula-tion and age at use. Lancet 1983;2:926–9.

57. The Centers for Disease Control Cancer andSteroid Hormone Study. Long-term oral con-traceptive use and the risk of breast cancer.JAMA 1983;249:1591–5.

58. Stadel BV, Rubin GL, Webster LA, et al. Oral con-

36 BREAST CANCER

traceptives and breast cancer in young women.Lancet 1985;2:970–3.

59. Miller DR, Rosenberg L, Kaufman DW, et al.Breast cancer risk in relation to early contra-ceptive use. Obstet Gynecol 1986;68:863–8.

60. Jick SS, Walker AM, Stergachis A, Jick H. Oralcontraceptives and breast cancer. Br J Cancer1989;59:618–21.

61. Romieu I, Willett WC, Colditz GA, et al. Prospec-tive study of oral contraceptive use and risk ofbreast cancer in women. J Natl Cancer Inst1989;81:1313–21.

62. Hankinson SE, Colditz GA, Manson JE, et al. Aprospective study of oral contraceptive use andrisk of breast cancer (Nurses’ Health Study,United States). Cancer Causes Control 1997;8:65–72.

63. Weinstein AL, Mahoney MC, Nasca PC, et al.Breast cancer risk and oral contraceptive use:results from a large case-control study. Epi-demiology 1991;2:353–8.

64. Wingo PA, Lee NC, Ory HW, et al. Age-specificdifferences in the relationship between oralcontraceptive use and breast cancer. ObstetGynecol 1991;78:161–70.

65. Collaborative Group on Hormonal Factors inBreast Cancer. Breast cancer and hormonalcontraceptives. Collaborative reanalysis ofindividual data on 53,297 women with breastcancer and 100,239 women without breast can-cer from 54 epidemiological studies. Lancet1996;347:1713–27.

66. Smith HO, Kammerer-Doak DN, Barbo DM,Sarto GE. Hormone replacement therapy in themenopause: a pro opinion. CA Cancer J Clin1996;46:343–63.

67. Consensus Conference. Osteoporosis. JAMA1984;252:799–802.

68. Bhatia NN, Bergman A, Karram MM. Effects ofestrogen on urethral function in women withurinary incontinence. Am J Obstet Gynecol1989;160:176–81.

69. Campbell S, Whitehead M. Oestrogen therapy andthe menopausal syndrome. Clin Obstet Gynecol1977;4:31–47.

70. Sherwin BB. Estrogen and cognitive function inwomen. Proc Soc Exper Biol Med 1998;217:17.

71. The Writing Group for the PEPI Trial. Effects ofestrogen or estrogen/progestin regimens onheart disease risk factors in postmenopausalwomen: the Postmenopausal Estrogen/Prog-estin Interventions (PEPI) trial. JAMA 1995;273:199–208.

72. Stampfer MJ, Colditz GA. Estrogen replacementtherapy and coronary heart disease: a quantita-tive assessment of the epidemiologic evidence.Prev Med 1991;20:47–63.

73. Wenger NK, Speroff L, Packard B. Cardiovascu-lar health and disease in women. N Engl J Med1993;329:247–56.

74. Grady D, Rubin SM, Petitti DB, et al. Hormonetherapy to prevent disease and prolong life inpostmenopausal women. Ann Intern Med1992;117:1016–37.

75. Lindsay R. Estrogen therapy in the prevention andmanagement of osteoporosis. Am J ObstetGynecol 1987;156:1347–51.

76. Weiss NS, Ufe CL, Ballard JH, et al. Decreasedrisk of fractures of the hip and lower forearmwith postmenopausal use of estrogen. N Engl JMed 1980;303:1195–8.

77. Mack TM, Pike MC, Henderson BE, Pfeffer RI.Estrogens and endometrial cancer in a retirementcommunity. N Engl J Med 1976;294:1261–2.

78. Boring JE, Bain CJ, Ehrmann RL. Conjugatedestrogen use and risk of endometrial cancer.Am J Epidemiol 1986;124:434–41.

79. Antunes CM, Stolley PD, Rosenshein NB, et al.Endometrial cancer and estrogen use: report ofa large case-control study. N Engl J Med 1979;300:9–13.

80. Hoover R, Gray LA, Cole P, MacMahon B.Menopausal estrogens and breast cancer. NEngl J Med 1976;295:401–5.

81. Gambrell RD, Maier RC, Sanders BI. Decreasedincidence of breast cancer in postmenopausalestrogen-progestogen users. Obstet Gynecol1983;62:435–43.

82. Hunt K, Vessey M, McPherson K, Coleman M.Long-term surveillance of mortality and can-cer incidence in women receiving hormonereplacement therapy. Br J Obstet Gynecol1987;94:620–35.

83. Bergkvist L, Adami H-O, Persson I, et al. The riskof breast cancer after estrogen and estrogen-progestin replacement. N Engl J Med 1989;321:293–7.

84. Mills PK, Beeson L, Phillips RL, Fraser GE.Prospective study of exogenous hormone useand breast cancer in Seventh-Day Adventists.Cancer 1989;64:591–7.

85. Colditz GA, Hankinson SE, Hunter DJ, et al. Theuse of estrogens and progestins and the risk ofbreast cancer in postmenopausal women. NEngl J Med 1995;332:1589–93.

86. Grodstein F, Stampfer MJ, Colditz GA, et al. Post-


menopausal hormone therapy and mortality. NEngl J Med 1997;336:1769–75.

87. Colditz GA. Relationship between estrogen lev-els, use of hormone replacement therapy, andbreast cancer. J Natl Cancer Inst 1998;90(11):814–23.

88. Fornander T, Rutqvist LE, Cedermark B, et al.Adjuvant tamoxifen in early breast cancer:occurrence of new primary cancers. Lancet1989;1:117–20.

89. Andersson M, Storm HH, Mouridsen HT. Incidenceof new primary cancers after adjuvant tamox-ifen therapy and radiotherapy for early breastcancer. J Natl Cancer Inst 1991;83:1013–7.

90. Rutqvist LE, Johansson H, Signomklao T, et al.Adjuvant tamoxifen therapy for early stagebreast cancer and second primary malignan-cies. J Natl Cancer Inst 1995;87:645–51.

91. Chang J, Powles TJ, Ashley SE, et al. The effectof tamoxifen and hormone replacement ther-apy on serum cholesterol, bone mineral densityand coagulation factors in healthy post-menopausal women participating in a random-ized, controlled tamoxifen prevention study.Ann Oncol 1996;7(7):671–5.

92. Powles TJ, Hickish T, Kanis JA, et al. Effect oftamoxifen on bone mineral density measuredby dual-energy x-ray absorptiometry in healthypremenopausal and postmenopausal women. JClin Oncol 1996;14(1):78–84.

93. Powles TJ, Jones AL, Ashley SE, et al. The RoyalMarsden Hospital pilot tamoxifen chemopre-vention trial. Breast Cancer Res Treat 1994;31:73–82.

94. Fisher B, Joseph P, Costantino D, et al., and otherNSABP investigators. Tamoxifen for preven-tion of breast cancer: report of the NationalSurgical Adjuvant Breast and Bowel ProjectP-1 study. J Natl Cancer Inst 1998;90:1371–88.

95. Early Breast Cancer Trialists’ CollaborativeGroup. Tamoxifen for early breast cancer: anoverview of the randomized trials. Lancet1998;351:1451–67.

96. Veronesi U, Maisonneuve P, Costa A, et al. Pre-vention of breast cancer with tamoxifen: pre-liminary findings from the Italian randomizedtrial among hysterectomised women. ItalianTamoxifen Prevention Study. Lancet 1998;352:93–7.

97. Anzano MA, Peer CW, Smith JM, et al. Chemo-prevention of mammary carcinogenesis in therat: combined use of raloxifene and 9-cis-retinoic acid. J Natl Cancer Inst 1996;88:123–5.

98. Gradishar WJ, Jordan VC. Clinical potential of newantiestrogens. J Clin Oncol 1997;15(2):840–52.

99. Boss SM, Huster WJ, Neild JA, et al. Effects ofraloxifene hydrochloride on the endometriumof postmenopausal women. Am J ObstetGynecol 1997;177:1458–64.

100. Cummings SR, Norton L, Eckert S, et al. Ralox-ifene reduces the risk of breast cancer and maydecrease the risk of endometrial cancer in post-menopausal women. Two-year findings fromthe multiple outcomes of raloxifene evaluation(MORE) trial [abstract 3]. Proc Am Soc ClinOncol 1998;17:2a.

101. Jordan VC, Glusman JE, Eckert S, et al. Incidentprimary breast cancers are reduced by ralox-ifene: integrated data from multicenter, double-blind, randomized trials in ~ 12,000 postmeno-pausal women [abstract 466]. Proc Am SocClin Oncol 1998;17:122a.

102. Valavaara R. Phase II trials with toremifene inadvanced breast cancer: a review. Breast Can-cer Res Treat 1990;16:S31–5.

103. Brunning PF. Droloxifene, a new anti-oestrogenin postmenopausal advanced breast cancer: pre-liminary results of a double blind dose findingphase II trial. Eur J Cancer 1992;28A:1404–7.

104. Sato M, Turner CH, Wang T, et al. LY353381.HC1:a novel raloxifene analog with improved SERMpotency and efficacy in vivo. J Pharmacol ExpTher 1998;287(1):1–7.

105. Toko T, Saito H, Fujioka A, et al. Antitumoractivity of miproxifene phosphate (TAT-59)against human mammary carcinoma. Gan toKagaku Ryoho 1998;25(6):829–38.

106. Sporn MB, Roberts AB. Role of retinoids in dif-ferentiation and carcinogenesis. Cancer Res1983;43:3034–40.

107. Lotan R. Retinoids and apoptosis: implicationsfor cancer chemoprevention and therapy. J NatlCancer Inst 1995;87:1655–7.

108. Delia D, Aillo A, Formelli F, et al. Regulation ofapoptosis induced by the retinoid N-(4-hydroxy-phenyl) retinamide and effect of deregulatedbcl-2. Blood 1995;85:359–67.

109. Mangelsdorf DJ, Umesono K, Evans RM. Theretinoid receptors. In: Sporn MB, Roberts AB,Goodman DS, editors. The retinoids: biology,chemistry, and medicine. 2nd ed. New York:Raven Press; 1994. p. 319–49.

110. Toma S, Isnardi L, Raffo P, et al. Effects of ALL-trans-retinoic acid and 13-cis-retinoic acid onbreast-cancer cell lines: growth inhibition andapoptosis induction. Int J Cancer 1997;70:619–27.

38 BREAST CANCER

111. Bishoff ED, Gottardis MM, Moon TE, et al.Beyond tamoxifen: the retinoid X receptor-selective ligand LGD1069 (TARGRETIN)causes complete regression of mammary carci-noma. Cancer Res 1998;53:479–84.

112. De Palo G, Camerini T, Marubini E, et al. Chemo-prevention trial of contralateral breast cancerwith fenretinimide. Rationale, design, method-ology, organization, data management, statisticsand accrual. Tumori 1997;83:884–94.

113. De Palo G, Veronesi U, Camerini T, et al. Canfenretinide protect women against ovarian can-cer ? J Natl Cancer Inst 1995;87:146–7.

114. Cancer incidence in five continents. Age-stan-dardized incidence rates, four-digit rubrics,and age-standardized and cumulative incidencerates, three-digit rubrics. IARC Sci Publ 1992;120:871–1011.

115. Dhingra K. A phase II chemoprevention trialdesigned to identify surrogate endpoint bio-markers in breast cancer. J Cell Biochem 1995;23Suppl:19–24.

116. Mariani L, Formelli F, De Palo G, et al. Chemo-prevention of breast cancer with fenretinide (4-HPR): study of long-term visual and ophthal-mologic tolerability. Tumori 1996;82:444–9.

117. National Cancer Institute. Annual Cancer Statis-tics Review including Cancer Trends: 1950-1985. Bethesda, MD: U.S. Department ofHealth and Human Services; 1988.

118. Wynder EL, Cohen LA, Muscat JE, et al. Breastcancer: weighing the evidence for a promotingrole of dietary fat. J Natl Cancer Inst 1997;89:766–75.

119. Rose DP, Connolly JM, Liu XH. Dietary fattyacids and human breast cancer cell growth,invasion, and metastasis. Adv Exper Med Biol1994;364:83–91.

120. Bagga D, Capone S, Wang HJ, et al. Dietary mod-ulation of omega-3/omega-6 polyunsaturatedfatty acid ratios in patients with breast cancer.J Natl Cancer Inst 1997;89:1123–31.

121. Zheng W, Gustafson DR, Sinha R, et al. Well-done meat intake and the risk of breast cancer.J Natl Cancer Inst 1998;90:1724–9.

122. Boyd NF, Greenberg C, Lockwood G, et al.Effects at two years of a low-fat, high-carbohy-drate diet on radiologic features of the breast:results from a randomized trial. J Natl CancerInst 1997;89:488–96.

123. Chlebowski RT, Blackburn GL, Buzzard JM, etal. Adherence to a dietary fat intake reductionprogram in postmenopausal women receiving

therapy for early breast cancer. The Women’sIntervention Nutrition Study. J Clin Oncol1993;11:2072–80.

124. Willett WC, Stampfer MJ. Sobering data on alco-hol and breast cancer. Epidemiology 1997;8(3):225–7.

125. Mezzetti M, La Vecchia C, Decarli A, et al. Pop-ulation attributable risk for breast cancer: diet,nutrition, and physical exercise. J Natl CancerInst 1998;90(5):389–94.

126. Jirtle RL, Haag JD, Ariazi EA, Gould M.Increased mannose 6-phosphate/insulin-likegrowth factor II receptor and transforminggrowth factor b-1 levels during monoterpene-induced regression of mammary tumors. Can-cer Res 1993;53:3849–52.

127. Messina MJ, Pensky V, Setchell KD, Barnes S. Soyintake and cancer risk: a review of the in vitroand in vivo data. Nutr Cancer 1994;21:113–31.

128. Bardon S, Picard K, Martel P. Monoterpenesinhibit cell growth, cell cycle progression. NutrCancer 1998;32(1):1–7.

129. Lamartiniere CA, Zhang JX, Cotroneo MS.Genistein studies in rats: potential for breastcancer prevention and reproductive and devel-opmental toxicity. Am J Clin Nutr 1998;68(6Suppl):1400S–5S.

130. Temple WJ, Lindsay RL, Magi E, Urbanski SJ.Technical considerations for prophylactic mas-tectomy in patients at high risk for breast can-cer. Am J Surg 1991;161:413–5.

131. Wong JH, Jackson CF, Swanson JS, et al. Analy-sis of the risk reduction of prophylactic partialmastectomy in Sprague-Dawley rats with 7,12-dimethylbenzanthracene-induced breast can-cer. Surgery 1986;90:67–71.

132. Nelson H, Miller SH, Buck D, et al. Effectivenessof prophylactic mastectomy in the preventionof breast tumors in C3H mice. Plast ReconstrSurg 1989;83:662–9.

133. Pennisi VR, Capozzi A. Subcutaneous mastec-tomy data: a final statistical analysis of 1500patients. Plast Surg 1989;13:15–21.

134. Woods JE, Meland NB. Conservative manage-ment in full-thickness nipple-areolar necrosisafter subcutaneous mastectomy. Plast ReconstrSurg 1989;84:258–64.

135. Schrag D, Kuntz KM, Garber JE, Weeks JC.Decision analysis: effects of prophylactic mas-tectomy and oophorectomy on life expectancyamong women with BRCA1 or BRCA2 muta-tions. N Engl J Med 1997;336:1465–71.

136. Lynch HT, Lemon SJ, Durham C, et al. A descrip-


tive study of BRCA1 testing and reactions to dis-closure of test results. Cancer 1997;79:2219–28.

137. Willemsen HW, Kaas R, Peterse JH, Rutgers EJ.Breast carcinoma in residual breast tissue afterbilateral subcutaneous mastectomy. Eur J SurgOncol 1998;24:331–8.

138. Ziegler LD, Kroll SS. Primary breast cancer afterprophylactic mastectomy. Am J Clin Oncol1991;14:451–4.

139. Jameson MB, Roberts E, Nixon J, et al. Metastaticbreast cancer 42 years after bilateral subcuta-neous mastectomies. Clin Oncol 1997;9:119–21.

140. Goodnight JE, Quagliana JM, Morton DL. Failureof subcutaneous mastectomy to prevent devel-opment of breast cancer. J Surg Oncol 1984;26:198–201.

141. Eldar S, Meguid MM, Beatty JD. Cancer of thebreast after prophylactic subcutaneous mastec-tomy. Am J Surg 1984;148:692–3.

142. Hartmann LC, Schaid DJ, Woods JE, et al. Effi-cacy of bilateral prophylactic mastectomy in

women with a family history of breast cancer.N Engl J Med 1999;340:77–84.

143. Singletary SE, Taylor SH, Guinee VF. Occur-rence and prognosis of contralateral carcinomaof the breast. J Am Coll Surg 1994;178:390–6.

144. Bilimoria MM, Morrow M. The woman atincreased risk for breast cancer: evaluation andmanagement strategies. CA Cancer J Clin1995;45:263–78.

145. Kroll SS, Miller MJ, Schusterman MA, et al.Rationale for elective contralateral mastectomywith immediate bilateral reconstruction. AnnSurg Oncol 1994;1:457–61.

146. Schairer C, Persson I, Falkeborn M, et al. Breastcancer risk associated with gynecologicsurgery and indications for such surgery.Obstet Gynecol 1997;70:150–4.

147. Parazzini F, Braga C, LaVecchia C, et al. Hys-terectomy, oophorectomy in premenopause,and risk of breast cancer. Obstet Gynecol1997;90:453–6.

40 BREAST CANCER

41

Mammography is currently the best availablescreening modality for early detection anddiagnosis of breast cancer. Periodic examina-tion of asymptomatic females with mammo-graphy has been shown to reduce breast cancermortality.1 In accordance with the AmericanCancer Society recommendations, the availablescientific data suggest a benefit from annualmammographic screening of all women begin-ning at the age of 40 years, combined withannual physical examination and monthlybreast self examination.2 For women between20 and 39 years of age, the ACS recommends abreast physical exam every three years andmonthly breast self exam. Patients with a first-degree premenopausal relative diagnosed withbreast cancer may consider beginning annualscreening examinations 10 years prior to theage at which the relative was diagnosed, in anattempt to benefit from early detection.3

Screening mammography evaluates asymp-tomatic women with the goal of discoveringunsuspected breast cancer at an early andpotentially curable stage. The routine screeningmammogram is comprised of a craniocaudal(CC) and mediolateral oblique (MLO) image ofeach breast. Abnormalities detected on thescreening exam are further evaluated with adiagnostic mammogram.

Diagnostic mammography is performed onpatients presenting with signs or symptoms ofpotential breast pathology. The request for a diag-nostic mammogram should be considered a con-sultation to provide evaluation of the patient’ssymptoms or mammographic findings and makefurther management recommendations. Tai-lored mammographic images, physical exami-nation, and breast ultrasound are frequentlyused to further investigate and explain a partic-ular clinical or radiographic concern. Alterna-tive projections and magnification views oftensupplement the standard mammogram. Spotcompression views of a particular site of con-cern can improve visualization of an underlyinglesion and allow for more accurate assessmentof lesion margins. Magnification mammogra-phy is best suited for enhanced characterizationand visualization of microcalcifications.

It is well known that mammography isunable to detect every breast cancer. The falsenegative rate of mammography ranges fromfour to thirty-four percent.4 The diagnosis ofbreast cancer therefore, is not excluded by anegative mammogram. Patient managementshould take into account the clinical assess-ment, despite a negative mammogram.

When appropriate, the diagnostic examincludes ultrasonography of a mammographic

3

Screening and Diagnostic ImagingJAN M. JESKE, MD JOEL R. BERNSTEIN, MDMARGARET A. STULL, MD

finding, palpable abnormality, or site of pain.Breast sonography aids in the characterizationof mammographically detected masses and canconfirm equivocal radiographic findings.Breast ultrasound is the initial imaging modal-ity of choice for evaluating palpable masses inwomen < 30 years of age and in lactating andpregnant women, as per the American Collegeof Radiology standards.5

BREAST IMAGING REPORTING AND DATA SYSTEMS

To improve the quality of mammographyreporting and early breast cancer detection, aconsortium of medical experts has developedthe Breast Imaging Reporting Data System(BIRADS).6 The American College of Radiol-ogy (ACR), in collaboration with the NationalCancer Institute, the Centers for Disease Con-trol and Prevention, the Food and Drug Admin-istration, the American Medical Association, theAmerican College of Surgeons, and the Collegeof American Pathologists, created BIRADS tostandardize communication of mammographicresults, reduce ambiguous breast imagingreports, and facilitate the collection and analysisof medical audit data at individual mammogra-phy practices as well as at the national level.

There are four main sections in BIRADS:breast imaging lexicon, reporting system, fol-low-up and outcome monitoring, and ACRNational Mammography Database (NMD). Thelexicon provides standardized language forlesion characterization. It also provides descrip-tive terms for masses, calcifications, architec-tural distortion, and associated findings of skinand nipple retraction as well as trabecularthickening and axillary adenopathy.

The reporting system uses a standardizedformat for the mammographic report, notingavailable comparison films, breast tissue com-position, a concise description of any signifi-cant findings, and a final assessment withappropriate recommendations. The mammo-graphic study is classified by BIRADS accord-

ing to one of the following decision categories:

1. Category 0: incomplete. Needs additionalimaging evaluation. This category typicallyarises after a screening mammography,when the patient must be recalled for addi-tional evaluation before a final assessmentcan be made.

2. Category 1: negative. The study is normal.3. Category 2: benign finding. There is a

benign finding described and no evidence ofmalignancy.

4. Category 3: probably benign finding—shortinterval follow-up suggested. Lesions in thiscategory have imaging characteristics thatare most likely benign. Follow-up is per-formed at a 6-month interval to establish sta-bility of a lesion with low probability of can-cer and detect those few cancers that initiallypresent with benign morphology. Ninety-eight percent of these lesions subsequentlyprove to be benign.7 This approach limitsunnecessary tissue sampling.

5. Category 4: suspicious abnormality—biopsy should be considered. Lesions in thiscategory have a 30 percent positive predic-tive value for being malignant; therefore,biopsy is recommended.7

6. Category 5: highly suggestive of malig-nancy—appropriate action should be taken.These lesions have morphologic featurescharacteristic of cancer. Intervention isrequired. The positive predictive value forcategory 5 lesions is 97 percent.7

These assessment categories are consistentwith those mandated by the final regulationsunder the Mammography Quality Standard Act(MQSA).8 These categories are not intended toreplace clinical evaluation of the breast. Clini-cal assessment directs the ultimate course ofaction when the mammogram is “negative” andthere is concern about a clinically suspiciousabnormality.

The BIRADS section on follow-up and out-come monitoring enables an individual radiolo-

42 BREAST CANCER

gist to assess his or her overall mammographyinterpretation skills by performing a mammog-raphy audit. There is a detailed description ofthe necessary core data to be collected and cal-culated for a comprehensive medical audit. TheACR BIRADS committee is encouraging mam-mography practices to participate in theNational Mammography Database. The pro-gram will enable evaluation of mammographicscreening in the diagnosis of clinically occultbreast cancer at the national level.

MAMMOGRAPHIC APPEARANCE OF BREAST CANCER

Breast cancer has numerous clinical and imagingpresentations. The classic mammographic appear-ance of infiltrating breast cancer is an irregularmass, often with ill-defined or spiculated mar-gins. In addition to a discrete mass, the invasivetumor can also present as a subtle asymmetricdensity or an architectural distortion. Clusteredpleomorphic calcifications are the common pre-sentation of in situ carcinoma that may or may notbe associated with invasive disease.

Secondary signs of malignancy, often asso-ciated with advanced stages of breast cancer,are detectable clinically and radiographically asareas of skin thickening or dimpling, nippleretraction, and axillary adenopathy. Diffuseskin thickening and breast edema manifested asincreased mammographic density are associ-ated with lymphangitic spread of cancer involv-ing the dermal lymphatics with inflammatorycancer. The underlying primary tumor is oftenobscured by the diffuse breast edema. Thesefindings must be differentiated from mastitis.Typically, the diagnosis of inflammatory canceris made clinically and by biopsy. Isolated nip-ple and areolar thickening can occur in patientswith Paget’s disease of the nipple.

Mammographic Analysis of Masses

The mammographic mass is a space-occupyinglesion seen in two different projections. Mam-mographic analysis of the mass is based on its

shape, margins, and density. Round or ovalshaped masses are typically associated with abenign etiology, most commonly a cyst orfibroadenoma. Increasing lobulations, irregularshapes, and spiculations increase the probabil-ity of malignancy.

Assessment of the lesion margin addsimportant distinguishing information. Cir-cumscribed lesions with sharp, distinct mar-gins are almost always benign (Figure 3–1).Poorly defined margins reflect the irregularinterface of the cancer cells invading the sur-rounding breast tissue. Due to superimposednormal fibroglandular tissue, lesion marginscan occasionally be obscured and difficult toaccurately assess. Spot compression viewsand ultrasonography allow for further charac-terization of lesion margins, particularly indense breast tissue.

Screening and Diagnostic Imaging 43

Figure 3–1. A, Mammographic image reveals the circum-scribed and gently lobulated margins of this fibroadenoma. B,Corresponding ultrasound shows the solid nature of this benigntumor as evidenced by the homogenous hypoechoic internalarchitecture.

A

B

Most breast cancers are mammographicallydense (more radio-opaque) relative to an equalvolume of normal fibroglandular tissue. Thepresence of radiolucent fat within a lesion ischaracteristic of a benign etiology. Fat contain-ing lesions include hamartoma (fibroadeno-lipoma), lipoma, galactocele, fat necrosis, andlymph nodes.

Sonographic Evaluation of Masses

Sonograpraphy is an excellent method for dis-tinguishing a simple cyst from a solid, circum-scribed tumor. On sonography, the simple cysthas smooth, rounded, or oval margins, containsno internal echoes, and has a sharply definedposterior wall with posterior acoustic enhance-

ment (Figure 3–2). The complex cystic struc-ture containing internal echoes or an intracysticmass requires further evaluation so as to notoverlook pathology such as papillary carci-noma or a necrotic neoplasm (see Figure 3–8).Thorough sonographic evaluation of all lesionmargins and internal architecture is necessaryto detect subtle signs of malignancy. Subtlemargin irregularity and internal heterogeneitymay be the only findings to suggest a malignantprocess. Sonographic features associated withbreast malignancy include marked hypo-echogenicity, irregular margins, and shadow-ing.9 Malignant lesion margins visualized withultrasound are often poorly defined, with anangulated, microlobulated, or branching pattern.

Imaging Specific Types of Infiltrating Breast Cancer

Approximately 85 percent of breast carcinomasarise from ductal structures, with the remaining15 percent arising from lobular structures.Infiltrating ductal carcinoma accounts for thelargest group of breast cancers, representing 65to 80 percent of cases.10 The classic mammo-graphic presentation of infiltrating ductal carci-noma is a high-density mass with spiculatedmargins (Figure 3–3A). Sonographically, thislesion is typically seen as a shadowing, hypo-echoic mass with irregular margins (Figure3–3B). The presentation of infiltrating ductal car-

44 BREAST CANCER

Figure 3–2. Sonographic depiction of a simple cyst. Note thelack of internal echoes and increased echogenicity deep tothe fluid (posterior acoustic enhancement).

Figure 3–3. A, Classic mammographic appearance of infiltrating ductal carcinoma demonstrating irregular, spiculated margins.B, Sonographic visualization of the same infiltrating ductal carcinoma illustrating a hypoechogenic mass with irregular margins.

A B

cinoma, however, can mimic a benign lesion withpartially circumscribed margins (Figure 3–4).

Infiltrating lobular carcinoma is the secondmost common type of invasive breast cancer,representing approximately 15 percent ofcases.10 It has a higher rate of multicentricityand bilaterality than infiltrating ductal carci-noma.11 Infiltrating lobular carcinoma is knownfor its insidious nature, delaying clinical andmammographic diagnosis. The subtle nature ofinfiltrating lobular carcinoma is thought to bedue to its pattern of single-file cellular infiltra-tion and lack of associated desmoplastic reac-tion.12 This tumor often presents as an evolvingasymmetric density, or, less often, a spiculatedmass on mammography (Figure 3–5A).13,14

Despite its elusive appearance on mammography,

infiltrating lobular carcinoma appears sono-graphically indistinguishable from infiltratingductal carcinoma (Figure 3–5B).3

Medullary, colloid, and papillary carcino-mas often present as partially circumscribedmammographic masses. Medullary carci-noma, accounting for approximately six per-cent of breast carcinomas, typically presentsin patients < 50 years of age and can be mis-taken for a fibroadenoma (Figure 3–6). Theslow growing colloid carcinoma, also knownas mucinous carcinoma, comprises only twopercent of all breast cancers and is more com-mon in older females (Figure 3–7). Papillarycarcinoma represents less than one percent ofall breast cancers and is associated with spon-taneous serosanguinous nipple discharge. An


Figure 3–5. A, The arrow identifies a vague, developing asymmetric density on this spot compression view. B, Despite the sub-tle mammographic appearance, the ultrasound image clearly visualizes this infiltrating lobular carcinoma revealing the markedlyjagged margins.

A B

Figure 3–4. A, Mammographic spot view of a 0.8 cm infiltrating ductal carcinoma with microlobulated and partially ill-definedmargins. B, Sonographic image of the same infiltrating ductal carcinoma demonstrating the marked irregularity of the tumor mar-gins (arrows), despite the small size.

A B

intracystic mass or intraductal lesion depictedby sonography raises concern for a papillaryneoplasm (Figure 3–8).

Tubular carcinoma accounts for less thantwo percent of breast cancers. Due to its veryslow growth, this tumor is typically small at thetime of detection. Tubular carcinoma often pre-sents as a small spiculated mass on mammog-

raphy, indistinguishable from infiltrating ductalcarcinoma (Figure 3–9). Tubular carcinoma canbe confused or associated with a radial scar(sclerosing papillomatosis), a benign entity(Figure 3–10).

Phyllodes tumor, once termed “cystosarcomaphyllodes,” comprises less than one percent ofbreast tumors. Approximately ten percent of

46 BREAST CANCER

Figure 3–7. A, This partially circumscribed mammographicnodule represents a colloid (mucinous) carcinoma. B, The het-erogeneous hypoechoic and lobulated appearance of this col-loid carcinoma is easily visualized with ultrasound.

A

B

Figure 3–6. A, Mammographic image of a 6 cm circum-scribed medullary carcinoma. B, Ultrasound reveals the het-erogeneous hypoechoic nature of this medullary carcinoma.

A

B

phyllodes tumors are malignant. This tumor canpresent as a rapidly growing palpable mass.Breast imaging usually shows a large rounded orlobulated circumscribed mass. Cystic spaces canbe seen by sonography (Figure 3–11).

Metastasis to the breast, although uncom-mon, can occur from a variety of primary malig-nancies, including melanoma, lymphoma, lungcancer, and contralateral breast carcinoma.Mammographically, the metastatic lesions tendto be round and lack spiculations (Figure 3–12).

BREAST CALCIFICATIONS

The presence of suspicious microcalcificationson a mammogram can make possible the earlydiagnosis of clinically occult breast cancer. Sincethe description of calcifications on radiographsof breast cancer by Leborgne in 195115 there havebeen substantial improvements in the mammo-

graphic detail of this finding as well as greaterawareness of its importance. Current mammo-graphic techniques can detect calcification in asmany as 50 percent of all breast cancers.16

Screening studies have shown that 90 percent ofall cases of nonpalpable ductal carcinoma in situ(DCIS)17 and 70 percent of minimal carcino-mas18 were detected on the basis of microcalci-fications. Many women, however, have someform of calcification in their breasts, the greatmajority of which are benign. Thus, we are chal-lenged to both detect and analyze calcificationsseen on mammograms to accurately diagnosebreast cancer without incurring consequences ofa false positive or false negative study.

Detection

Nowhere in medical imaging are fine detailimages as vital as they are in the detection and


Figure 3–8. A, Papillary carcinoma with the typical appearance of a partially circumscribed mass. B, Sonographic image of thesame papillary neoplasm reveals the intracystic mass (arrow).

BA

Figure 3–9. A and B, Mammographic and sonographic images demonstrating the irregular shape and spiculated margins of this0.5 cm tubular carcinoma (arrow).

A B

evaluation of breast microcalcifications, whichmay be as small as 0.1 mm. It is necessary tooptimize the technique in all stages of produc-tion of the mammographic image to yield highcontrast, high resolution, and motion-freefilms. Viewing conditions, including the rou-tine use of a hand magnifying lens to system-atically search each film, are also important.Magnification radiography in the study ofbreast calcifications, although at the cost of ahigher radiation dose (2´), provides greaterresolution than that achieved with a hand lens.

Analysis

Due to the frequency of calcifications on mam-mograms, careful analysis is needed to recognizeclearly benign calcifications and allow follow-up

of low suspicion calcifications. While some cal-cifications have classically benign or malignantfeatures that allow the mammographer to easilycharacterize them for appropriate action, there isa sizable intermediate or indeterminate groupthat requires thorough analysis to assess the like-lihood of malignancy. The characteristics mostuseful in evaluating calcifications include size,number, form, distribution, and location.

Size

Calcifications associated with malignancy areoften as small as 0.1 to 0.3 mm in diameter andusually < 0.5 mm. However, larger granularforms, up to 2 mm, and longer fine linear formsof calcification, may occasionally be seen. Theindividual calcifications in cancer often vary in

48 BREAST CANCER

Figure 3–10. A and B, The imaging appearance of this benign radial scar (arrow) mimics the tubular carcinoma in Figure 3–9.

BA

Figure 3–11. A, This large phyllodes tumor encompasses much of the mammographically visualized breast tissue. B, Sonogra-phy confirms the solid nature of this mass, heterogeneous and hypoechoic.

A B

size. In a group of mixed-size calcifications,the degree of suspicion should relate to thesmallest forms.

Number

The presence of a focal group of five heteroge-neous microcalcifications in a volume of 1 cm3

of tissue has been accepted as suspicious.19

Biopsy of fewer calcifications may be per-formed if new, pleomorphic, or fine linear orbranching calcifications have developed since aprior mammogram. Generally, the greater thenumber of suspicious calcifications grouped ina small area, the higher the chance of malig-nancy.20,21 Magnification studies are used tomore accurately determine the number of calci-fications present. The pathologist invariablyfinds more calcifications than are visible on themammogram.

Shape

Small round to oval dense punctate calcifica-tions located in cystically dilated acini are con-sidered benign lobular calcifications. Malig-nant calcifications are typically ductal inorigin, forming in ductal cellular secretions ornecrotic cellular debris. Fine linear and branch-ing ductal calcifications or pleomorphic calci-fications grouped to form a cast of the duct aremost typical of malignancy, often comedo car-

cinoma, but are not the most common presenta-tion. Irregular granular forms are more fre-quently seen, often differing in size and varyingfrom jagged “fractured crystal” shapes to roundpunctate dots similar to lobular calcifications.It is this overlap of benign and malignantshapes of granular calcifications that results inthe large indeterminate group of microcalcifi-cations requiring biopsy. Magnification studiesare invaluable in characterizing these calcifica-tions, allowing elimination of some typicallylobular forms from consideration for biopsy.

Distribution

Bilateral diffusely scattered calcifications arealmost always benign and are often associatedwith adenosis. However, the calcifications ofadenosis or sclerosing adenosis may be focaland indistinguishable from malignancy.Malignant calcifications are usually found asa focal cluster involving a small area of onebreast but can be more extensive, presentingas one or several clusters in the distribution ofthe ductal system of one lobe, or virtually anentire breast. While some benign masses con-tain coarse calcifications, the presence of fineor pleomorphic calcifications associated witha mass increases the likelihood of malignancyand may suggest a related extensive intraduc-tal component.


Figure 3–12. A, Bilateral craniocaudal (CC) images demonstrating numerous round masses in this patient with multiple myelomametastasis. B, Sonographic image of the largest metastatic lesion in the right breast, depicting the heterogeneous hypoechoic nature.

BA

Location

Calcifications must be proven to be within thebreast to accurately evaluate their significance.Mimics of breast parenchymal calcificationinclude skin calcium, artifacts, and pseudocal-cifications.

Benign Calcifications

Analysis of breast calcifications by an experi-enced mammographer will allow accurate diag-nosis of characteristically benign calcifications.The various types are described below.

Skin Calcification and Pseudocalcification

Skin calcification is commonly related to seba-ceous glands and appears as lucent, centeredrings in a peripheral location (Figure 3–13A).Skin calcification may be punctate or irregular,and may appear to lie within the breastparenchyma on standard views. Therefore, atangential view skin localization study (Figure3–13B) may be necessary to prove a cutaneouslocation. Calcium in warts, moles, scars, anddermal lesions as well as pseudocalcificationsdue to tattoos, talc, deodorant, or film artifactscan be misleading (Figure 3–14).

Vascular Calcifications

Calcification caused by atherosclerosis in arter-ial walls is usually easy to recognize due to thetypical continuous linear tubular pattern (Figure3–15A). Early changes of short segment calcifi-cations appearing as discontinuous deposits inone wall may have a granular or fine linearappearance that can arouse suspicion (Figure3–15B). Magnification views in alternate pro-jections will usually allow a correct diagnosis.

Calcium in Cysts

Thin, curvilinear calcifications defining themargin of a mass are seen with cyst wall calci-

50 BREAST CANCER

Figure 3–13. Skin calcification. A, Lucent centered sebaceous gland calcification (arrow) appears intramammary on standardprojection. B, Tangential view confirms location of calcium in the skin (arrow).

BA

Figure 3–14. Pseudocalcification. Talc powder in molesbeneath breast mimics parenchymal calcification.

fications. Intracystic calcium particles sus-pended in fluid, known as “milk of calcium,”may appear in multiple tiny cysts or a singlelarger cyst. This diagnosis is best proven with a90-degree lateral film showing a meniscus orteacup shape of layered calcium in a cyst (Fig-ure 3–16). These calcifications may be difficultto see when viewed en face in the CC view.

Fibroadenoma

Calcifications appear in fibroadenomas as aresult of involution which may be due to myx-oid degeneration, hyalinization, or infarction.Early calcifications may occur in the peripheryof the mass and progress to large, geographicareas of calcium the appearance of which hasbeen compared to popcorn (Figure 3–17).Eventually, the soft-tissue component may becompletely replaced by a dense conglomerateof calcifications. However, when this classicpattern is not followed, an involuted fibroade-noma may appear as fine pleomorphic calcifi-cations without a visible mass and biopsy maybe required for diagnosis.

Secretory Calcification

Inspissated ductal secretions in normal ordilated ducts may calcify to form solid, coarse,and linear ductal casts involving one or more

ducts diffusely and often bilaterally. Large,tubular periductal calcifications can appear inplasma-cell mastitis (Figure 3–18). Increaseddensity of the subareolar parenchyma may befound. These large, rod-like secretory calcifica-tions are usually easily differentiated frommalignant calcifications by their large size andgreater length.

Fat Necrosis

Calcifications due to fat necrosis are often seenas fine-rim calcifications surrounding a lucentcenter, varying in size from a few millimeters to


Figure 3–15. Vascular calcification. A, Typical tubular arterial calcification. B, Irregular linear calcification, due to incomplete arte-rial wall involvement, can appear suspicious.

BA

Figure 3–16. Milk of calcium. Layered sedimented calcium inmicrocysts appears curvilinear or teacup-shaped on horizon-tal beam lateral film.

several centimeters. Small ring forms, usually< 5 mm in diameter, are often idiopathic (Figure3–19A). Dystrophic calcifications depositedafter trauma, hemorrhage, surgical biopsy, andradiation may appear as larger and less regularcalcifications surrounding an oil cyst (Figure3–19B). It is important to note that this type ofcalcium may appear several years after alumpectomy and breast radiation. In its earlystages, it can be difficult to differentiate fromrecurrent malignant calcifications.

Lobular Calcifications:Adenosis/Sclerosing Adenosis

Lobular calcifications form in the acini in asso-ciation with such entities as adenosis, scleros-ing adenosis, atypical lobular hyperplasia, andcystic hyperplasia. Characteristically, these cal-cifications are small, dense, and round (Figure3–20A). If a lobule is distorted by surroundingsclerosis, the individual forms may be moreirregular. The distribution of calcifications inadenosis and sclerosing adenosis is often bilat-eral, diffuse, and inhomogeneous due to vari-able involvement of individual lobules. Thesecharacteristic findings indicate a benignprocess. Alternatively, the calcifications can bemore focal, presenting as unilateral looselygrouped calcifications, regional calcifications,or a solitary small cluster (Figure 3–20B). In

these situations, careful analysis of the calcifi-cations with magnification views may allowperiodic follow-up, but biopsy will often benecessary to exclude carcinoma.

Malignant and Indeterminate Calcifications

Microcalcifications are present in as many as50 percent of all breast cancers and in an evenhigher percentage of stage 0 and stage 1 breastcancers.16–18 The presence of clustered micro-calcifications may be the only indication ofearly preinvasive malignancy. Mammographic

52 BREAST CANCER

Figure 3–17. Fibroadenoma. A, Few early coarse peripheral calcifications. B, Classic popcorn calcification. C, Large dense cal-cification nearly replaces mass.

A B C

Figure 3–18. Secretory calcification. Large, solid rod-like andtubular calcifications appear in a ductal orientation.

detection of microcalcifications in patientswith DCIS accounts for this entity rising froma small percentage of lesions found at biopsy tothe current rate of 20 to 40 percent for biopsiesfor clinically occult lesions.22 Stomper and col-leagues,23 in a group of 100 patients withDCIS, reported that 84 percent of cases pre-sented with microcalcifications, either alone(72%) or as calcifications associated with asoft-tissue density (12%).

Classic malignant calcifications are typicallyassociated with comedo carcinoma but are alsopresent in other histologic subtypes of DCIS.Characteristic malignant calcifications occur as

fine, pleomorphic, linear, and branching calcifi-cations (Figure 3–21) or multiple irregular gran-ules forming castings arranged in a ductal distri-bution (Figure 3–22). The extent of involvementmay vary from < 1 cm to an entire lobule or evena whole breast (Figure 3–23). Holland observeda significant discrepancy between the estimatedmammographic and actual histopathologicextent of DCIS.24 This discrepancy is most pro-nounced for low grade DCIS. Mammographyunderestimates the histopathologic extent by 16percent for high grade DCIS and 47 percent forlow grade DCIS.24

Clustered irregular granular calcifications,


Figure 3–20. Lobular calcification. A, Punctate, round, scattered calcifications (were bilateral) due to adenosis. B, Small groupof round clustered calcifications (arrow), likely acinar, in a dilated lobule.

Figure 3–19. Fat necrosis. A, Dense round lucent-centered calcifications caused by idiopathic fat necrosis. B, Postoperative oilcysts with thin eggshell (white arrow) and course rim calcification (black arrow).

BA

A B

not clearly ductal in distribution, or mixedforms of granular and casting calcifications, area more common presentation of DCIS (65%)compared to the classic pure casting and linearforms (35%).23 These granular calcifications areseen more frequently in low-grade, noncomedocarcinoma, although there is enough overlapthat one cannot reliably subtype DCIS based onthe mammographic morphology. It is also thisgroup, because of the variability of the granular

calcifications (Figure 3–24), that may have themost similarity to benign forms of calcium andthus require the greatest scrutiny. Invasivebreast cancer associated with DCIS involving

54 BREAST CANCER

Figure 3–21. Linear and branching calcifications typical ofcomedo type ductal carcinoma in situ.

Figure 3–22. Granular calcifications forming ductal casts(arrows) in comedo type ductal carcinoma in situ.

Figure 3–23. Distribution of microcalcifications in ductal car-cinoma in situ. A, focal (arrow); B, segmental; C, diffuse (wholebreast).

A

B

C

25 percent or more of the area of the tumor isclassified as an extensive intraductal component(EIC) (Figure 3–25). Magnification mammog-raphy is helpful in defining the extent of the in


Figure 3–25. Extensive intraductal component. Calcificationswithin a tumor mass and extention into surrounding tissue.

situ component and defining the margins of theresection. Magnification mammography isneeded to accurately determine the number,shape, and distribution of calcifications. How-ever, even after careful study, clustered granularcalcifications are often considered indetermi-nate and biopsy is performed, resulting in ayield for malignancy of 20 to 33 per-cent.20,21,25,26

GALACTOGRAPHY

Galactography or ductography may be used toevaluate patients presenting with spontaneousisolated bloody or clear nipple discharge.Numerous studies document a 10 to 15 percentincidence of carcinoma in women with sponta-neous unilateral discharge from a singleduct.27–29 The incidence of carcinoma inpatients with bloody versus serous discharge issimilar.27 Other types of discharge, includinggreen, yellow, or milky discharges, have notbeen associated with carcinoma.12

Galactography is not indicated in pregnant orlactating women or when the discharge occursfrom multiple bilateral ducts. Galactography

Figure 3–24. Variability of calcifications in non-comedo duc-tal carcinoma in situ. A, round, punctate granules; B, pleomor-phic granules (arrow) which vary in size; V, “fractured crystal,”highly irregular granules (arrow).

A

B

C

should not be performed on a patient with activemastitis because it may worsen the inflammation.

Evaluation of women with bloody or serousnipple discharge should begin with mammogra-phy. If the mammogram is unrevealing, ductog-raphy can be performed by painlessly cannulat-ing the discharging duct and gently injectingradiographic contrast material. Postinjectionmammographic images reveal intraductal fill-ing defects or abrupt duct termination whenpathology is present (Figure 3–26). The benignintraductal papilloma is the most commoncause of spontaneous serosanguinous nippledischarge. Benign duct ectasia may also causenipple discharge.

Prior to surgical excision of the ductallesion, preoperative galactography can be per-formed with a mixture of iodinated contrastmaterial and methylene blue dye to enableintraoperative localization of the involved duct.It has been suggested that this technique canallow a more accurate and limited resection.27

EVALUATION OF THE CONSERVATIVELY TREATED BREAST

Mammography is an essential tool for monitor-ing conservatively treated breast cancer patients.Recognizing the distinctions between mammo-graphic appearance of the expected postsurgical,

postradiation developments and that of recurrentcarcinoma is critical for patient care.

Postexcision Mammography

Magnification mammography is useful aftersurgery to ensure complete excision of themalignant lesion. If the targeted lesion and alltumor-related calcifications are not clearlyincluded on the specimen radiograph, or ifthere is discordance between the pathologyresults and the preoperative diagnosis, magnifi-cations mammography can reveal the retainedprimary lesion or residual malignant calcifica-tions. Postoperative magnification mammogra-phy is useful prior to radiation to ensure com-plete excision of the calcium-containing tumor.Unfortunately, mammography cannot defini-tively predict the histologic extent of tumors.

Mammography performed within the firstfew weeks after tumorectomy is often limitedby the patient’s discomfort, breast edema, post-surgical architectural distortion, and the pres-ence of postoperative fluid collections (ie, sero-mas and hematomas) (Figure 3–27A). Post-excision changes frequently result in increaseddensity that can obscure subtle residual malig-nant calcifications (Figure 3–27B).

Although these alterations regress and stabi-lize with time, they are accentuated and pro-longed by subsequent radiation therapy.30,31

Postradiation Mammography and Long-Term Follow-up

Mammography is important for long-term mon-itoring to evaluate for recurrent disease or newlesions in either breast. This should commencewith a post-treatment baseline mammogrambeing performed within 3 to 9 months followingtumor excision and completion of radiationtherapy. Standard views may be supplementedwith special projections to fully define post-therapy changes. Magnification mammographyis particularly important when evaluating thebreast for retained or recurrent malignant

56 BREAST CANCER

Figure 3–26. Galactographic image demonstrating theabrupt duct termination due to an intraductal filling defect(arrow), in this case, a benign intraductal papilloma.

microcalcifications. Subsequent annual or morefrequent mammograms should be obtained asindicated by clinical or radiographic evaluation.Comparison of the post-treatment mammo-grams to preceding studies is necessary to accu-rately assess radiographic changes followingcompletion of therapy. Among the most com-mon post-therapy changes are breast edema,skin thickening, postoperative fluid collections,scarring, fat necrosis, and calcifications.

The mammographic findings of skin thick-ening, irregular breast parenchyma, and breastedema following surgery and radiation are mostprominent on the post-treatment baseline studyand typically diminish over 2 to 3 years follow-ing conservative therapy.30,32 Once mammo-graphic stability of the breast has been estab-lished, any increase in the architecturaldistortion or enlargement of the dense scar atthe surgical site suggests the presence of recur-rent tumor. Interrupted lymphatic drainage afterextensive axillary node dissection may producechronic breast edema. Recurrent edema witherythema may be a manifestation of infection(mastitis), inflammatory breast carcinoma, orrecurrent breast cancer with lymphatic involve-ment. Postoperative fluid collections, such ashematomas or seromas at the lumpectomy site,present mammographically as high-density ovalmasses that may have ill-defined or spiculatedmargins (see Figure 3–27). These diminish insize as the fluid is resorbed over a period of 6 to18 months.30,31 If the fluid collection is enlarg-ing or an abscess is suspected, ultrasonographycan be used to evaluate the process further andto guide diagnostic needle aspiration.

Coarse, benign, dystrophic calcificationscan develop several years after radiation andsurgery. These calcifications frequently repre-sent fat necrosis or occasionally calcifiedsuture material in the surgical bed. Sometimesthe developing benign calcifications have anindeterminate appearance necessitating tissuesampling. Occasionally fat necrosis can presentas an irregular mass-like lesion that may simu-late tumor recurrence.

Recurrent Breast Cancer

The mammographic indications of tumor recur-rence at the surgical site frequently developbetween 2 to 3 years following conservativebreast surgery.33 The development of increased


Figure 3–27. Postoperative fluid collection obscuring resid-ual malignant microcalcifications. A, The large high densityoval mass with ill-defined margins represents the postsurgi-cal fluid collection. B, Magnification view 2 months latershows postsurgical scar with resorption of the seroma andadjacent subtle residual malignant calcifications (arrows).

A

B

skin thickening, progressive architectural distor-tion, enlargement of the surgical scar, a new soft-tissue mass, or new pleomorphic microcalcifica-tions raise the suspicion of recurrent disease(Figures 3–28 and 3–29). Calcifications associ-ated with recurrent DCIS frequently resemblethe mammographic appearance of the originaltumor.34 Contrast-enhanced MR imaging mayhelp distinguish scar tissue from recurrent dis-ease.35 Computed tomography (CT) scanningcan be used to assess extensive local breast dis-ease (Figure 3–30), revealing direct chest-wallinvasion or tumor recurrence in the chest wall.Computed tomography is also useful in assess-ing regional and distant metastatic disease.

EVALUATION OF THE AUGMENTED BREAST

The presence of breast implants causes technicalproblems that impair the ability to detect breastcancer by mammography. The radiopaqueimplant blocks x-ray transmission, which limitsthe imaging of breast tissue. Implants compressbreast tissue against the skin which can obscurea significant amount of anterior breast tissue onconventional mammographic images. Implantsalso diminish the compressibility of the breast,particularly if there is capsular contracture.

Improved visualization of the anterior breasttissue is provided by implant displacement

58 BREAST CANCER

Figure 3–28. Recurrent infiltrating ductal carcinoma. Sequential views of the right breast showA, the spiculated primary tumor (arrow) in the lateral breast; B, post-lumpectomy changes; andC, recurrent invasive disease (arrow) in the surgical bed.

A B C

Figure 3–29. Recurrent DCIS and invasive ductal carcinoma developing 3 years after breast conservation therapy. A, Image on12/96 shows architectural distortion at the lumpectomy site (arrow). B, Spot magnification view demonstrates faint microcalcifica-tions (curved arrow) adjacent to enlargement of the surgical scar representing DCIS associated with invasive tumor (straight arrow).

A B

views (Figure 3–31). This modified positioningtechnique supplements the standard views inwomen with cosmetic augmentation. Unfortu-nately, breast tissue near the chest wall is notcompletely imaged with either standard ormodified views.

Breast cancer has the same mammographicfeatures in women with or without implants

(Figure 3–32). Lesions are usually more con-spicuous on the implant displacement views(see Figure 3–31). Patients with cosmetic aug-mentation or reconstruction with implants maydevelop parenchymal scarring that should notbe confused with a malignant process. Dys-trophic calcifications and calcifications associ-ated with the fibrous capsule surrounding theimplant can occur but are usually clearly benignin appearance.

Ultrasonography of the augmented breastassists in the evaluation of palpable and mam-mographically detected masses (Figure 3–33).Ultrasound can identify a palpable implant valveand distinguish a focal implant herniation orcontained implant rupture from a breastparenchymal abnormality, eliminating the need


Figure 3–30. Recurrent infiltrating ductal carcinoma after leftmastectomy and radiation therapy. CT scan shows dermalinvolvement (short white arrows), left chest wall and axillaryrecurrence with direct invasion into the mediastinum and sub-pleural space (long white arrow). Confluent pathologic medi-astinal adenopathy (black arrows) is present.Tumor surroundsthe left axillary clips.

Figure 3–31. Multicentric DCIS in a woman with subpectoralimplants. Implant displacement and standard MLO view (pho-tographed back to back for comparison) show multiple groupsof pleomorphic microcalcifications (arrows). The malignantcalcifications are more conspicuous on the implant displace-ment image.

Figure 3–32. Multifocal invasive disease with EIC adjacent toimplant. Mammographic image reveals three irregular masses(arrows) with bridging spicules and pleomorphic malignantcalcifications in the right axillary tail and inferior axilla.

Figure 3–33. Stromal fibrosis presenting as an indeterminatemass. A, Implant displacement view shows the oval lesion(arrow) with partially defined margins located superficial to theimplant. B, Ultrasound demonstrates the oval hypoechoic massdirectly subjacent to the dermis. The anterior aspect of theintact hypoechoic implant is seen deep to the breast tissue.

B

A

for tissue sampling.30,36 If a suspicious or inde-terminate parenchymal abnormality is con-firmed, ultrasound guided fine-needle aspirationor core biopsy limits the risk of implant rupture.

Magnetic resonance imaging is widely usedto evaluate prosthetic implant integrity and canaid in differentiating an implant complicationfrom an intramammary lesion.37,38 Dynamic,contrast-enhanced MR imaging shows promisefor improved detection and monitoring ofbreast cancer in certain women with cosmeticbreast augmentation or breast reconstructionusing prosthetic implants.36

MAGNETIC RESONANCE IMAGING OF THE BREAST

High–spatial–resolution MR imaging of thebreast is evolving as an important adjunctivediagnostic tool for the detection, characteriza-tion, staging, and monitoring of breast cancer.Contrast-enhanced MR imaging may allowmore accurate preoperative evaluation of pri-mary malignant breast lesions that may beunderestimated or not seen on mammography orby ultrasonography. Magnetic resonance imag-ing has revealed unsuspected multifocal, multi-centric, diffuse, and bilateral disease in patientswith a solitary mammographic lesion.39–41

The sensitivity of contrast-enhanced MRimaging approaches 100 percent in the detec-tion of invasive breast cancer when compared tomammography and physical examination.35,39,41

The specificity of breast MR imaging rangesfrom 37 to 97 percent.39,41,42 This wide range isattributed to the overlap in contrast enhance-ment of benign and malignant lesions. Higherspecificity for breast MR imaging can beachieved using a dynamic contrast-enhancedtechnique with three-dimensional imaging. Ondynamic contrast-enhanced studies, malignantlesions typically exhibit rapid enhancement,whereas benign lesions show slower or noenhancement. False-positive enhancing lesionsinclude fibroadenomas, sclerosing adenosis,radial scars, mastitis, atypical hyperplasia, lobu-lar neoplasia and normal breast tissue duringvarious phases of the menstrual cycle.35–37

Magnetic resonance imaging is also capa-ble of demonstrating mammographically andclinically occult in situ carcinoma.39–46 High-resolution contrast-enhanced dynamic MRimaging may enable more accurate evaluationof tumors in dense fibroglandular tissue andassist in screening of high-risk patients.35,39–41 Itmay also be useful in differentiating recurrentcarcinoma from scarring and in evaluating theresponse to neoadjuvant chemotherapy.39

Limitations of MR imaging include the highcost of the examination, limited availability of

60 BREAST CANCER

dedicated imaging equipment, and significantoverlap in the enhancement patterns of benignand malignant lesions. Additional large multi-institutional studies will help define the clinicalusefulness and cost effectiveness of MR imag-ing in the assessment of breast cancer.

NUCLEAR MEDICINE TECHNIQUESFOR BREAST IMAGING

Scintigraphic imaging of the breast frequentlyuses the single-gamma radiotracer technetium-99m (Tc-99m) sestamibi. Studies have indicated95 to 97 percent sensitivity for breast tumors > 1cm.47 The sensitivity, however, is poor for small,nonpalpable or medially located lesions, rangingfrom 26 percent for lesions < 0.5 cm to 56 per-cent for lesions between 0.5 cm and 1 cm.47 Thespecificity for Tc-99m sestamibi imaging rangesfrom 73 to 90 percent.47,48

Breast cancer imaging has also been per-formed using positron emission tomography(PET) imaging of a structural analog of glu-cose, fluorodeoxyglucose (FDG). The sensitiv-ity and specificity for PET scanning of thebreast range from 70 to 90 percent and 85 to 95percent, respectively.47,48 As with scintigraphicimaging, lesions < 1 cm are not reliablydetected with PET imaging. Improvements inthe ability to detect small lesions will be neces-sary before clinical utility of scintigraphy andPET can be proven.

COMPUTER-AIDED DIAGNOSIS

Computer-aided detection and diagnosis(CAD) is an evolving technology that functionsas a “second reader” of the mammographicfilms.49 Current CAD requires digitizing themammography films to allow computer analy-sis. The computer program identifies areas ofthe breast that match certain prescribed tissuepatterns, densities, or calcifications for furtherreview by the interpreting radiologist. Limita-tions of CAD include the increased film han-dling time and high start-up costs. Prospective

clinical studies are needed to determine theefficacy of the CAD methods.

CONCLUSION

Increased public awareness of the importanceof breast cancer screening and continued tech-nical advances in breast imaging will enhancepatient care by allowing early detection, stag-ing, and monitoring of the disease. Recognitionof the diverse imaging presentations of breastcancer is crucial for early diagnosis and propermanagement.

REFERENCES

1. Breast cancer screening for women ages 40–49.NIH Consens Statement. 1997;Jan 21–23;15(1):1–35.

2. Feig SA, D’Orsi CJ, Hendrick RE, et al. AmericanCollege of Radiology Guidelines for BreastCancer Screening. Am J Roentgenol 1998;171:29–33.

3. Kopans DB. Breast imaging. Philadelphia: Lip-pincott-Raven Publishers; 1998.

4. Huynh PT, Jarolimek, AM, Daye S. The false-negative mammogram. Radiographics 1998;18:1137–54.

5. American College of Radiology (ACR). 1998Standards. Reston (VA): American College ofRadiology; 1998. p. 317.

6. American College of Radiology (ACR). Breastimaging reporting and data system (BIRAD-STM). 3rd ed. Reston (VA): American Collegeof Radiology;1998.

7. Orel SG, Kay N, Reynolds C, et al. BI-RADS cat-egorization as a predictor of malignancy. Radi-ology 1999;211:845–50.

8. Federal Register, Part II, Department of Healthand Human Services, 1997, October 28.

9. Stavros AT, Thickman D, Rapp CL, et al. Solidbreast nodules: use of sonography to distin-guish between benign and malignant lesions.Radiology 1995;196:123–34.

10. Adler DD. Imaging evaluation of spiculatedmasses. In: Friedrich M, Sickles EA, editors.Radiological diagnosis of breast diseases.Berlin: Springer-Verlag; 1997.

11. Fisher ER. What is early breast cancer? In: ZanderJ, Baltzer J, editors. Early breast cancer; histo-pathology, diagnosis, treatment. New York:Springer;1985.


12. Bassett LW, Jackson VP, Johan R, et al. Diagnosisof diseases of the breast. Philadelphia: W. B.Saunders Company; 1997.

13. Mendelson EB, Harris KM, Doshi N, et al. Infiltrat-ing lobular carcinoma: mammographic patternswith pathologic correlation. Am J Roentgenol1985;153:265–71.

14. Newstead GM, Baute PB, Toth HK. Invasive lob-ular and ductal carcinoma: mammographicfindings and stage at diagnosis. Radiology1992;184:623–7.

15. Leborgne R. Diagnosis of tumor of the breast bysimple roentgenography: calcification in carci-noma. Am J Roentgenol 1951;65:1–11.

16. Miller RR, Davis R, Stacey AJ. The detection andsignificance of calcification in the breast. Aradiological and pathological study. Br JRadiol 1976;49:12–26.

17. Feig SA, Shaber GS, Patchefsky A. Analysis ofclinically occult and mammographically occultbreast tumors. Am J Roentgenol 1977;128:403–8.

18. Moskowitz M. The predictive value of certainmammographic signs in screening for breastcancer. Cancer 1983;51:1007–11.

19. Sickles EA. Mammographic features of early breastcancer. Am J Roentgenol 1984;143:461–4.

20. Egan RL, McSweeney MD, Sewell CW. Intra-mammary calcification without an associatedmass in benign and malignant disease. Radiol-ogy 1980;137:1–7.

21. Muir BB, Lamb J, Anderson TS, Kirkpatrick AE.Microcalcification and its relationship to can-cer of the breast: experience in a screeningclinic. Clin Radiol 1983;34:193–200.

22. Rebner M, Rajic V. Noninvasive breast cancer.Radiology 1994;190:623–31.

23. Stomper PC, Connolly JL, Meyer JE, Harris JR.Clinically occult ductal carcinoma in situdetected with mammography: analysis of 100cases with radiographic-pathologic correlation.Radiology 1989;172:235–41.

24. Holland R, Hednriks JH, Verbeek AL, et al.Extent, distribution and mammographic/histo-logical correlations of breast ductal carcinomain situ. Lancet 1990;335:519–22.

25. Meyer JE, Kopans DB, Stomper PC, Lindfors KK.Occult breast abnormalities: percutaneous pre-operative needle localization. Radiology 1984;150:335–7.

26. Feig SA. Mammographic evaluation of calcifica-tions. In: Kopans DB, Mendelson EB, editors.Syllabus: a categorical course in breast imag-ing. Radiol Soc North Am; 1995.

27. Van Zee KJ, Perez GO, Minnard E, et al. Preoper-ative galactography increases the diagnosticyield of major duct excision for nipple dis-charge. Cancer 1998;82:1874–80.

28. Tabar L, Dean PB, Pentek Z. Galactography: thediagnostic procedure of choice for nipple dis-charge. Radiology 1983;149:31–8.

29. Leis HP. Management of nipple discharge. WorldJ Surg 1989;13:736–42.

30. Mendelson EB. Evaluation of the postoperativebreast. Radiol Clin North Am 1992;30:107–38.

31. Kopans DB. The altered breast: pregnancy, lactation,biopsy, mastectomy, radiation, and implants. In:Kopans DB. Breast imaging. 2nd ed. Philadel-phia: Lippincott-Raven; 1997. p. 445–96.

32. Brenner RJ, Pfaff JM. Mammographic featuresafter conservative therapy for malignant breastdisease: serial findings standardized by regres-sion analysis. Am J Roentgenol 1996;167:171–8.

33. Mendelson EB, Tobin CE. Imaging the breastafter surgery and radiation therapy. In: Syl-labus: a categorical course in breast imaging.Radiological Society of North America; 1995.p. 175–84.

34. Liberman L, Van Zee KJ, Dershaw DD, et al.Mammographic features of local recurrence inwomen who have undergone breast-conservingtherapy for ductal carcinoma in situ. Am JRoentgenol 1997;168:489–93.

35. Viehweg P, Paprosch I, Strassinopoulou M, et al.Contrast-enhanced magnetic resonance imag-ing of the breast: interpretation guidelines. TopMagn Reson Imaging 1998;9:17–43.

36. Brenner RJ. Tumor detection in the augmentedbreast. In: Gorczyca DP, Brenner RJ, editors. Theaugmented breast: radiologic and clinical per-spectives. New York: Thieme; 1997. p. 154–69.

37. Gorczyca DP. Magnetic resonance imaging of thefailing implant. In: Gorczyca DP, Brenner RJ,editors. The augmented breast: radiologic andclinical perspectives. New York: Thieme; 1997.p. 121–43.

38. Reynolds HE. Evaluation of the augmented breast.Radiol Clin North Am 1995;33:1131–45.

39. Orel SG, Schnall MD, Powel, et al. Staging of sus-pected breast cancer: effect of MR imaging andMR-guided biopsy. Radiology 1995;196:115–22.

40. Mumtaz H, Hall-Craggs M, Davidson T, et al.Staging of symptomatic primary breast cancerwith MR imaging. Am J Roentgenol 1997;169:417–24.

41. Nunes LW, Schnall MD, Orel SG, et al. Breast MR

62 BREAST CANCER

imaging: interpretation model. Radiology1997;202:833–41.

42. Hulka CA, Smith BL, Sgroi DC, et al. Benign andmalignant breast lesions: differentiation withecho-planar MR imaging. Radiology 1995;197:33–8.

43. Orel SG, Reynolds C, Schnall MD, et al. Breastcarcinoma: MR imaging before re-excisionalbiopsy. Radiology 1997;205:429–36.

44. Stompre PC, Herman S, Klippenstein DL, et al.Suspect breast lesions: findings at dynamicgadolinium-enhanced MR imaging correlatedwith mammographic and pathologic features.Radiology 1995;197:387–95.

45. Muller-Schimpfle M, Ohmenhauser K, Stoll P, et al.

Menstrual cycle and age: influence on parenchy-mal contrast medium enhancement in MR imag-ing of the breast. Radiology 1997;203:145–9.

46. Soderstrom CE, Harms SE, Copit DS, et al.Three-dimensional RODEO breast MR imag-ing of lesions containing ductal carcinoma insitu. Radiology 1996;201:427–32.

47. Williams MB, Pisano ED, Schnall MD, et al.Future directions in imaging of breast diseases.Radiology 1998;206:297–300.

48. Wahl RL. Nuclear medicine techniques in breastimaging. Sem Ultrasound CT MRI 1996;17:494–505.

49. Feig SA, Yaffe MJ. Digital mammography. Radi-ographics 1998;18:893–901.


65

The recent increase in the detection of nonpal-pable breast abnormalities requiring furtherevaluation is thought to be the direct result ofmore favorable participation in mammographyscreening. Appropriate diagnostic work-upwill lead to a relative increase in lesions thatare of sufficient risk to warrant a biopsy. Infact, it has been estimated that approximately1.2 million breast biopsies are performed peryear in the United States. Unfortunately, anaverage positive predictive value for mammog-raphy of 20 percent (range 15 to 35%) willyield a significant number of biopsies per-formed for benign disease.1–4 If 5 women areidentified on the mammogram to have a lesionrequiring biopsy, only 1 of these 5 women willbe found to have breast cancer. Therefore, iftraditional methods for histologic confirmationare used, all 5 women would proceed to theoperating room for an open surgical biopsyafter first having had wire localization in theradiology suite. Image-guided percutaneousbreast biopsy, an effective alternative, hasrecently gained favor. Image-guided percuta-neous breast biopsy would provide a secondarylevel of screening for these five women in a lessinvasive, cost-effective manner as well as a his-tologic diagnosis without sacrificing accu-racy.5–9 The patient with breast cancer may thenproceed to definitive surgical management andthe other four women with a benign diagnosismay be placed in an appropriate follow-up pro-tocol. It is with this concept in mind that wereview the state-of-the-art in image-guided per-cutaneous breast biopsy.

NEEDLE LOCALIZATION BREAST BIOPSY

The gold standard with which image-guidedpercutaneous breast biopsy is compared is theneedle or wire localization open surgical breastbiopsy. However, this traditional managementof a suspicious nonpalpable breast abnormalityis not without its own error rate. The inability tosuccessfully remove the appropriate lesionranges from 0.5 to 17 percent.10–15 Some of thereasons given for unsuccessful biopsies include(1) poor radiologic placement of the localiza-tion wire, (2) preoperative and intraoperativedislodgment of the wire, (3) surgical inaccu-racy and inadequacy in excising the appropriatetissue, (4) failure to obtain a specimen radio-graph, and (5) the pathologist missing the focusof disease when searching through a larger tis-sue sample provided by the surgeon.

The needle localization and open surgicalbreast biopsy is typically more invasive.Although a surgeon may discount the impor-tance of a scar on the breast, women frequentlyhave a great concern over even a one- to two-inch scar, especially on the superior aspect of thebreast. The possibility of altered breast shapeassociated with tissue removal is also important.This fear is thought to be responsible for womenfailing to participate in recommended screening.

In addition to cutaneous scarring, parenchy-mal scarring may also complicate future mam-mographic follow-up.6 Kopans has suggestedthat significant parenchymal scarring is rarelyassociated with a properly performed needle

4

Image-Directed Breast BiopsyRICHARD E. FINE, MD

localization breast biopsy.12 However, surgeonsare frequently faced with mammographicreports indicating architectural distortion at thesite of a prior biopsy that might mimic thechanges associated with a malignancy.

Despite the potential advantages of image-guided percutaneous breast biopsy, there arestill reasons why a standard needle localizationopen surgical biopsy may be chosen for histo-logic diagnosis. Some patients desire completesurgical removal of a breast abnormality andwill not be satisfied with a “sampling proce-dure.” Certain facilities and insurance plans donot provide access to facilities where image-guided procedures (ie, stereotactic) are per-formed. There are also certain lesion character-istics, patient characteristics, and potentialpathologic entities that may render image-guided breast biopsy difficult or inappropriate.

The essentials for a properly performed wirelocalization breast biopsy include accurate local-ization, a comfortable, confident patient, andappropriate surgical planning and tech-nique.10,12–15 The radiologist most often localizesthe lesion with orthogonal mammography.12,16

Increasingly, stereotaxis and ultrasonography areused to identify the location of a nonpalpablelesion.17,18 Whichever technique is used, it isimportant to have the wire within 1 cm of thelesion for the localization to be considered accu-rate and to limit the potential for error.19

The majority of these procedures are per-formed under local anesthesia with or withoutintravenous sedation.12,14,20 Appropriate instilla-tion of local anesthesia for both image-guidedpercutaneous biopsies and open surgical proce-dures is important for gaining the patient’s con-fidence in the earliest aspect of the procedure.This will ensure a calmer patient through theremainder of the procedure. A skin wheal israised using a small (27- or 30-gauge) needlewith 1 percent Xylocaine ™ (lidocaine). Deeperlocal anesthetic should be placed (25-gauge nee-dle) as a block around the potential dissectionsite. Despite the added cost, some have foundsedation or general anesthesia to lessen the dis-

comfort while improving the technical ease andsuccess of removing the lesion with a needlelocalization and open surgical breast biopsy.10

While planning the incision site, cosmesisshould be taken into consideration withoutignoring cancer surgery principles.18 If a lesionhas a relatively low probability of malignancyand is within a reasonable distance from thenipple-areolar complex, a circumareolar inci-sion should be considered. Regardless of thewire insertion site, localization mammographyshould be used to estimate the location of thelesion within the breast.21 The breast incisiondoes not necessarily require inclusion of theguide wire insertion site. In addition to usinglocalization mammography, familiarity withthe localization wire lengths and inherentmarkings may aid in a more accurate estimationof lesion location (Figures 4–1A to 4–1C).19–21

Once the lesion location is determined, theincision is planned to avoid tunneling of a moresuspicious lesion through benign breast tis-sue.22 The incision is carried straight downthrough the subcutaneous layer without thedevelopment of any flaps until the body of thewire is encountered and can be brought into theconfines of the biopsy cavity. This is designedto maintain more support during closure and toavoid potential indentation.18,19 On the basis ofthe relationship of the lesion to the tip of thewire, the excision is performed.

If the needle localization biopsy is being per-formed as a follow-up to an abnormal lesion onimage-guided percutaneous breast biopsy or apresumed malignancy, it is then performed as aneedle localization lumpectomy.23 Marginassessment becomes crucial to the success of theprocedure. A technique of assisting the patholo-gist with margin assessment involves intraopera-tive inking of the margins by the surgeon usingthe Davidson™ multicolor inking system. Theanterior, 12, 3, 6, and 9 o’clock positions as wellas the deep margins may be marked with corre-sponding colors (red, green, blue, black, yellow,and orange), using a cotton-tipped applicator(Figure 4–2). The specimen is then dipped into 3

66 BREAST CANCER

percent acetic acid (vinegar) to set the colors.Subsequently, the specimen is sent to the radiol-ogy department for a specimen radiograph andthen on to the pathology department. The speci-men is sent dry so as to allow further stabiliza-tion of the ink prior to pathology sectioning.

When the probability of malignancy is highor already confirmed, a technique to ensureadequate margins may be instituted, especiallywhen the procedure is for a potential multifocalprocess, such as ductal carcinoma in situ(DCIS). Once the main biopsy/lumpectomyspecimen is removed, random additional mar-gins may be taken from the walls of the biopsycavity (12, 3, 6, 9 o’clock, and deep positions).The thickness of the additional margins isapproximately 5 to 10 mm. The multicolor ink-ing system is used to mark the side of the spec-imen representing the new margin of resection.

If there is focal margin involvement of the mainspecimen, the resection is felt to be adequate, ifthe additional margins are clear.

Image-Directed Breast Biopsy 67

Figure 4–1. A, Based on the wire length, angle of insertion,and the localization mammograms, the position of the wire tipand lesion are determined and the incision planned. B, Theincision is carried down through subcutaneous tissue, thewire is brought into the confines of the cavity and then a 2-0 silk suture is used for traction, and dissection is carriedout around the course and tip of the wire. C, The specimenradiograph maybe magnified to determine the completenessof excision of microcalcifications.A

B C

Figure 4–2. The Davidson™ multicolor inking system is usedto identify margins of resection.

After hemostasis is obtained with electro-cautery, the wound is closed. Closure consistsonly of reapproximating the subcuticular anddermal layers. There is neither draining norreconstruction of the deep aspect of the biopsycavity. If postoperative radiation therapy is aprobability for the patient, small hemoclips areplaced in the walls and base of the cavity. Thiswill assist the radiation oncologist in planningtherapy (especially the boost therapy requiredfor close margins or used as a standard treat-ment in many centers). A clear, waterproof,Tegaderm™ dressing is applied, which allowsthe patient to shower or bathe in the immediatepostoperative period.

The patient returns to the office during thefollowing week for wound assessment, to dis-cuss the pathology, and to plan future follow-up. A baseline mammogram of the biopsiedbreast is obtained after 6 months to look for anyparenchymal scarring and to evaluate forappropriate and adequate biopsy. If the preop-erative mammogram contained suspicious cal-cifications at the site of the biopsy, a baselinemammogram should be obtained prior to theinitiation of radiotherapy.

IMAGE-GUIDED BREAST BIOPSY

The physician will have several concerns aboutinstituting an image-guided breast biopsy pro-gram. Will patients accept sampling rather thanexcision? What will be the false-negative rate?Will we maintain the proper indications, andwho should perform the image-guided breastbiopsy—radiologists or surgeons?

While the physicians deal with these concerns,the patient is exposed to media headlines such as“The Needle Replaces the Knife.” It does not takea very sophisticated patient-consumer to under-stand that a needle is less invasive than a knife.In addition to the media deluge, patients are alsoexposed to corporate-driven advertisement ofnew breast biopsy devices. Acceptance of a newtechnology, in the face of physician reluctance,is now being influenced by outside sources.

Stereotactic Breast Biopsy: Equipment

Percutaneous breast biopsy for nonpalpable dis-ease requires imaging. The two main imagingmodalities are stereotaxis and ultrasonography.These modalities are complementary to oneanother and therefore knowledge of both isrequired to provide the physician with the fullrange of options. Stereotactic breast biopsy isperformed using specialized mammographyequipment. The equipment obtains stereo mam-mogram images of a lesion within the breast and then relies on computerized triangulation of the targeted lesion to calculate the three-dimensional position of this lesion.5,24 There aretwo main types of stereotactic equipment, theupright add-on and the dedicated prone.25–27

Add-on stereotactic equipment uses standardupright mammography with an attachable plat-form to perform targeting and biopsy (Figure4–3). Add-on stereotactic units provide theadvantage of maximum use of the equipment,which has dual capabilities: screening or diag-nostic mammography and stereotactic breastbiopsy. This can provide considerable cost sav-ings by avoiding not only dedicated equipmentbut also dedicated space within a breast diag-nostic facility. Despite these potential advan-

68 BREAST CANCER

Figure 4–3. The StereoLoc™ is the upright add-on stereotacticguidance system used with the Lorad™ mammography unit.

tages, add-on stereotactic breast biopsy unitstraditionally have been less popular than dedi-cated prone stereotactic tables. Because of theupright patient position and patient visualiza-tion of the procedure, there is the potential foran increase in syncopal episodes.5,28 In addition,the upright units provide minimal work spaceand limited access to the breast. Dedicated pronestereotactic tables are more costly and requirededicated space for performing stereotacticbreast biopsy. These disadvantages appear to be

outweighed by several advantages. The proneposition allows gravity to aid the technologist inreaching more posterior lesions. A greatlyenhanced work space underneath the table andout of the patient’s view allows for the additionof more advanced breast biopsy devices.

There are two dedicated prone systems avail-able, the Mammotest/Mammovision™ (FischerImaging, Denver, Colorado) and the LoradStereoguide™ (Lorad, a division of Trex Medical,Danbury, Connecticut) (Figures 4–4A and 4–4B).


Figure 4–4. A, Fischer Mammotest/Mammovision™. B, Lorad Stereoguide™.

A

B

The Stereoguide™ table has, until recently, alsobeen distributed as the ABBI™ table by the U.S.Surgical Corporation. The Fischer Mammotest/Mammovision™ is a unidirectional patient-positioning table, with the aperture for thepatient’s breast located at one end of the table.Access around the breast is approximately 180ºto 210º with rotation of the C-arm and mam-mography tube head. Special software featuresincrease the angle of access to the breast up to240º with “target on scout” and close to 360ºwith the “lateral arm.” The Lorad Stereoguide™

is a bidirectional patient-positioning table withthe aperture for the breast in the center of thetable and a foot extension at either end. Withthe facility to rotate the patient 180º and theaddition of the 180º rotation of the C-arm, thereis a true 360º access to the breast.

Specimen Acquisition Devices

The stereotactic biopsy equipment is amenableto the addition of a number of different devicesthat are used in specimen acquisition. Theseinclude fine-needle aspiration (FNA), auto-mated Tru-cut needle core, and vacuum-assisted

biopsy devices (Mammotome™, MIBB™) (Fig-ure 4–5). Breast biopsy acquisition devices(ABBI™, Site-Select™) have larger cannulatype tools and are designed for an image-guided excision of tissue as a substitute for tra-ditional surgical biopsy.

Fine-needle aspiration using stereotacticguidance was the first minimally invasivebiopsy technique used for nonpalpable lesionsand is preferred by some to this day. In 1989,Lancet published a classic article from theKarolinski Institute, which evaluated thestereotactic fine needle biopsy of 2,594 mam-mographically detected, nonpalpable lesionsfrom 1983 to 1987.29 Of the 2,005 (77.3%)cases judged to be benign, only 1 turned out tobe malignant 14 months later. Of the 576 cases(21.9%) selected for needle localization fol-lowed by open breast biopsy on the basis ofcytologic and/or mammographic interpretation,cancer was identified in 429 (75.7%). Dowlat-shahi and colleagues published 528 cases ofstereotactic FNA, in corroboration with theUniversity of Kiel from the Federal Republic ofGermany.17 In their report stereotactic guidancewith 23-gauge FNA had a sensitivity of 95 per-

70 BREAST CANCER

Figure 4–5. Prone stereotactic tables can perform fine-needle aspiration, needle core biopsy, vacuum-assisted biopsy, and wire localization procedures.

cent and accuracy of 92 percent. Furthermore,this article confirmed the accuracy of stereo-tactic localization by imaging the tip of the nee-dle within 2-mm of the center of the lesion in96 percent of the cases.

Acceptance of FNA as a standard techniquefor the performance of stereotactic biopsy hadlimited success, especially in the United States.Fine-needle aspiration has long been recog-nized to have several potential pitfalls. Thisincludes insufficient sampling as high as 38percent, a low ranging sensitivity rangingbetween 68 and 93 percent, and specificityvarying between 88 and 100 percent.30 Thebroad range of sensitivity and specificity isdependent on a number of factors, including thetype of lesion to be sampled, the individual per-forming the aspiration, and the individual inter-preting the cytologic specimen. Finally, cytol-ogy rarely provides a specific benign diagnosis.

Automated Tru-cut ™ Type Biopsy

In the late 1980s, Parker (Radiology ImagingAssociates, Englewood, Colorado) and othersbegan working with the automated Tru-cutBiopsy™ instrument (Bard Urologic, Coving-ton, Georgia) designed for biopsy of theprostate. He combined this technology on theprone Fischer™ stereotactic table for perform-ing large-core stereotactic breast biopsy. In1991, Parker published a case series of 102patients, where every patient had a stereotactic-guided, large-core needle biopsy, followed bytraditional surgical excision of the lesion.5

There was agreement of histologic results in 98cases (96%). One cancer missed with a corewas determined to be a very difficult lesion tolocalize because of its posterior position. Thisarticle set the stage for the standardization ofstereotactic biopsy techniques, which is stillbeing followed today. In addition to the use ofdedicated prone stereotactic equipment, he alsoadvocated the use of “long throw” biopsy nee-dle devices. The longer excursion of the innerand outer sheaths of the needle provided a con-

sistently larger tissue sample. The standard useof the 14-gauge needle eliminated the issue ofinsufficient sampling. Several different gaugeneedles for automated Tru-cut™ biopsy havebeen evaluated. The lower rate of insufficientsampling and increased sensitivity, withoutincreased complications, have led to the mini-mum 14-gauge size becoming the standard.31,32

Another principle identified to increase theaccuracy is the routine use of pre- and postfirestereotactic imaging. Prefire stereo imagesassess the appropriate alignment of the needleto the lesion, and postfire stereo images docu-ment the penetration of the needle through thelesion. Tru-cut™-type core-needle biopsy withstereotactic localization was demonstrated byParker and others to be a cost-effective proce-dure, which was less invasive and reducedpatient anxiety. It has a lower false-negativerate when compared with FNA.5–8 Furthermore,the need for cytologic expertise is avoided,which is important in the community settingwhere expert cytopathologist interpretationmay not be available.

Initial experience with stereotactic breastbiopsy was based on film screen technologyuntil 1993, when the charged coupled device(CCD) camera replaced the film cassetteimage receptor. This facilitated digital recre-ation of the breast lesion on a computer moni-tor. Several advantages for digital image acqui-sition were immediately recognized. As thesoftware improved, image acquisition timedecreased. This significantly reduced proce-dure time, which, in turn, increased patientcomfort. A more comfortable patient was lesslikely to move and as a result the breast andtarget lesion remained stationary for improvedtargeting accuracy. A critical differencebetween digital image acquisition and filmscreen image acquisition involves postprocess-ing of the image. Once a film is developed, thecharacteristics of the lesion on the film cannotbe changed. When a digital image is acquired,the lesion can be enhanced, magnified, andeven inverted to appear black on a white back-


ground. These postprocessing features areespecially helpful when evaluating small clus-ters of microcalcifications. An additionalbonus for patients included lower radiationwith each procedure, related to a narrow fieldof view. Digital imaging for stereotactic biopsyis not full field (ie, the entire breast is notimaged). The area of the breast imaged is lim-ited to a square area of 5 ´ 5 cm. The truelearning curve in performing stereotacticbreast biopsy revolves around this narrow fieldof view. The lesion for which the biopsy is tobe performed requires prior identification on ahigh-quality, film-screen diagnostic mammo-gram. The physician must recognize and trans-fer the appearance of the lesion to a digitalimage. The digital image is at least four timesmagnified, is dependent on monitor resolution,and the lesion is not seen in association withthe full anatomic image of the breast.

Performing Stereotactic Breast Biopsy

After review of the mammogram, the generalapproach to the breast is chosen. The shortestskin-to-lesion distance and the ability to visual-ize the lesion are both factors in choosing theapproach.33 Also important is assuring enoughtissue beyond the lesion to account for theexcursion (“throw”) of the biopsy needle dur-ing sampling. A lesion in the lower inner aspectof the breast may be approached from the medi-olateral position. A lesion in the most inferiorsix o’clock position of the breast may beapproached from the mediolateral or laterome-dial position on the Fischer Mammotest™ table.The “lateral arm” may be attached to access thelesion through the noncompressed portion ofthe breast. The six o’clock lesion may also beapproached from the caudal cranial position (ie,from below) on the Lorad Stereoguide™ table.The mammography technologist is responsiblefor positioning the patient for the desiredapproach. Other responsibilities of the technol-ogist include calibration and maintenance ofthe equipment and quality assurance.

The first digital image to be taken is thezero-degree scout image. No matter whatapproach to the breast is taken, the scout imageis perpendicular to the plane of the skin. Oncethe appropriate lesion is identified, it is posi-tioned in the middle third of the scout image,from left to the right. A set of stereo images areobtained by rotating the mammography tubehead to a +15° and –15° to yield an arc of sepa-ration between the two stereo images of 30°(Figure 4–6). When targeting for an automatedTru-cut instrument, a central target is chosen onthe lesion in image number one, and a corre-sponding target is chosen on the lesion in imagenumber two. Additional targets or offsets arechosen around the center of the lesion, forexample, at 12, 3, 6, and 9 o’clock positions(Figure 4–7). With the appropriate targetsentered into the system, the software determinesthe horizontal, parallax shift of the lesion fromstereo image number one to stereo image num-ber two. The vertical coordinate is unchanged. Atrigonometric formula then calculates the hori-zontal, vertical, and depth coordinates of thelesion’s center within the breast.

With the three-dimensional coordinates ofthe lesion calculated, the puncture device orstage, which houses the biopsy instrumenta-tion, is motor driven to the calculated horizon-tal and vertical positions. The biopsy device isadvanced toward the skin, and the site forinsertion is identified. A local anesthetic isinjected to raise an appropriate skin wheal.Additional local anesthetic may be judiciouslyinjected into the deeper aspects of the breast.This is done in a radial manner to avoid plac-ing too much local anesthetic in any one areafor fear of moving the targeted lesion, or plac-ing too much local anesthetic in front of thelesion for fear of limiting its visualization. An11-blade scalpel is used to make a vertical-ori-ented skin incision of 2 to 4 mm in length. Theneedle is advanced into the breast to a positionseveral millimeters short of the center of thecalculated depth of the lesion. On the basis ofthe mechanics of the biopsy instruments in use

72 BREAST CANCER

(dead space at needle tip, sample notch size,excursion or throw of instrument), a precalcu-lated pull-back is determined by each biopsyinstrument manufacturer. The depth chosen isthe calculated depth minus this predeterminedpull-back depth. By using the pull-back depth,the sampling notch will be positioned formore adequate, fuller sampling of the lesion(Figure 4–8).

Prefire stereo digital images are acquired toassess the alignment of the needle tip in rela-

tionship to the lesion to be biopsied. Theremust be symmetry of alignment between thetwo stereo images. If the alignment is satisfac-tory, the automated Tru-cut instrument is firedand the first sample from the center of thelesion is obtained. Prior to removing the needlefrom the breast, it is important to documentsymmetrical penetration of the needle throughthe center of the lesion by taking stereo postfireimages of the lesion (Figures 4–9A, 4–9B). Theneedle is withdrawn from the breast, the speci-


Figure 4–6. Viewing the stereotactic tables from above. The diagram illustrates the movement of the tube head for acquiringstereo images and the resultant z-value determination.

Figure 4–7. Digital stereo images in reverse video with central targets and multiple offsets on theFischer™ system.

men is placed on moistened gauze, and thebiopsy instrument is recocked and reinsertedinto the breast through the same skin incision toacquire the next target or offset sample. Thesame process is repeated until all samples areobtained. The average number of needle corebiopsy samples for a nodular density is five orsix (Figure 4–10). When biopsies of microcal-cifications are performed, a greater sampling isrequired. Even in open biopsy surgical litera-ture, pathologic assessment has identified atyp-ical hyperplasia and DCIS at a “distance” fromthe targeted microcalcifications.34 Liebermanfrom Memorial Sloan Kettering Hospital dis-cussed the issue of how many core biopsy spec-imens are needed.35 In her report, biopsies of145 lesions were performed; 92 were nodulardensities and 53 were microcalcifications. Fivecores yielded diagnosis in 99 percent of thebiopsies for nodular densities. Five coresyielded a diagnosis in only 87 percent of thepatients presenting with microcalcifications;six or more cores yielded a diagnosis in 92 per-cent of the cases. When the targeted sampleshave been obtained, a set of postprocedure dig-

ital images are acquired. The purpose of theseimages is to document the removal of themicrocalcifications and to verify the presenceof residual calcifications. A specimen radi-ograph of the tissue documents the inclusion ofmicrocalcifications within the core sam-ples.36,37 Frequently, only a few calcificationsare evident on specimen radiography. Increas-ingly, the literature has begun to support con-cern over insufficient sampling of core biopsyfor microcalcifications. Israel and Finereported a series of 500 consecutive core-nee-dle biopsies with a sensitivity of 97.8 percentand a false-negative rate of 1.5 percent. How-ever, upgrading of diagnosis on open surgicalexcision was evident in 33 percent of the cases,where atypical ductal hyperplasia was identi-fied on core-needle biopsy, and DCIS was iden-tified on excision.38 The presentation for theupgrading of diagnosis was microcalcifica-tions, where the average core sampling wasbetween 9 and 12 samples. Not surprisingly,atypical ductal hyperplasia diagnosed at stereo-tactic core-needle biopsy has been called anindication for open surgical biopsy.39,40 Consis-

74 BREAST CANCER

Figure 4–8. Positioning the needle at the lesion center will result in failure to sample the front halfof the lesion. With a 5-mm pull-back to the front of the lesion, the entire length of the lesion will besampled.

tent with the reporting of others, Libermanreported a series of 25 cases of atypical ductalhyperplasia identified on stereotactic core-nee-dle biopsy, with significant upgrading (52%) tocarcinoma on open surgical excision.41

Vacuum-Assisted Biopsy Devices

A solution to decrease the upgrading of diagno-sis is to increase tissue sampling. Tissue sam-pling can be increased by increasing the numberof samples taken, increasing the size of the tis-sue sample taken, and/or the manner in whichthe tissue samples are obtained. A vacuum-assisted biopsy satisfies these requirements.25,42

Not only is the sample size larger, but the sam-ples may be taken in a circumferencial, contigu-ous manner to allow reconstruction of the histo-logic architecture of the area of the biopsy. TheMammotome™ consists of a driver, which maybe housed on either of the dedicated pronestereotactic systems. Inside the driver, there is abiopsy probe that consists of three parts. Theprobe body has a sampling notch with severalholes connected to a vacuum. The second partof the probe is a rotating cutter lumen that fitsinside the probe body and is rotated by a circu-lar gear mechanism in the driver. The third andlast component of the probe is the knockout pin,which fits inside the cutter lumen and isattached to the rear vacuum to assist in retriev-ing tissue samples from the probe body. Theprobe is manufactured in two sizes: 14- and 11-

gauge.43 The Mammotome™ system is ideallysuited for taking biopsy specimens of microcal-cifications under stereotactic guidance. Oncethe calcifications have been imaged stereotacti-cally, they are targeted on the computer monitor.By changing the position of the sampling notch,multiple tissue samples can be obtained usingonly a single target probe insertion. The targetmay be centered in the cluster of microcalcifi-cations or preferentially placed at the cluster’sedge so that the overlying probe will notobscure a small cluster of microcalcificationswhen digital images are obtained. When thethree-dimensional coordinates are calculated bythe stereotactic system software, the driver andprobe are positioned at the corresponding hori-zontal and vertical coordinates. The driver andprobe are advanced toward the breast, wherelocal anesthetic is injected. The probe is insertedthrough a vertical skin incision to the appropriate


Figure 4–9. A, Prefire alignment; the relation of the tip of the needle to the lesion must be symmetrical. B, Postfire alignment;the relation of the tip of the needle to the lesion must be symmetrical.

A B

Figure 4–10. 14-gauge needle core biopsysample with histologic confirmation of fibroadenoma.

depth (Figure 4–11). Prefire stereo images assessthe alignment of the probe with the microcalcifi-cations. The driver is designed with a spring-loaded mechanism to permit automated advance-ment of the probe through the breast tissue. Thesampling notch may be positioned within thebreast by the automated forward movement ofthe probe, or it can be manually aligned with thelesion by taking the driver to the appropriatedepth with the probe already in its full excur-sion. The alignment of the sampling notch with

the lesion is confirmed with postfire (align-ment) stereo images (Figure 4–12). A vacuumpulls breast tissue into the sampling notch, andthe cutter is advanced across the samplingnotch, cutting the tissue free from inside thebreast. The tissue sample is removed from thespecimen retrieval chamber. The entire processis then repeated. The number of biopsy samplestaken is based on the size of the lesion and thevolume of tissue desired. On average, approxi-mately 12 to 16 tissue samples are obtained fora small cluster of microcalcifications, fre-quently resulting in removal of the entire mam-mographic evidence of the lesion. The vacuumassistance provides several advantages. It elimi-nates the need for pinpoint accuracy requiredwith automated Tru-cut biopsy instruments andfacilitates removal of multiple tissue sampleswithout removal of the biopsy probe.42 In addi-tion, the notch may be positioned for specificdirectional sampling, on the basis of the align-ment of the probe notch with the lesion. Thelarger tissue samples provide a greater chance

76 BREAST CANCER

Figure 4–11. Local anesthetic injection, skin incision with an11-blade scalpel followed by insertion of vacuum-assistedprobe into the skin, in preparation for biopsy.

Figure 4–12. The vacuum assisted probe may be used fordirectional sampling. In this example, the majority of sampleswill be taken between 3:00 and 9:00 through the 6:00 position.

of lesion removal and a greater percentage ofpositive specimen radiographs (Figure 4–13).Burbank and Jackman have demonstrated thatthe improved accuracy with the directional vac-uum-assisted biopsy device decreases theupgrading of diagnosis seen with core-needlebiopsy.44,45 Burbank showed that the 14-gaugedevice provided no upgrading of atypical ductalhyperplasia to carcinoma at open biopsy. Jack-man compared 14-gauge vacuum-assisteddevice to 14-gauge core-needle biopsy andillustrated a reduction of upgrading of atypicalductal hyperplasia from 48 to 18 percent. Post-procedure imaging yields a well-defined air-contrast cavity in the majority of cases. The 11-gauge core probe also allows a marker clip to beplaced in the wall of the biopsy cavity to assistin the localization of the area in the future whenall evidence of the lesion has been eliminated.46

This marker clip and/or the residual cavity maybe localized if the diagnosis requires furthersurgical management. The complication rate forthe device is less than 1 percent, which is com-parable to core-needle biopsy.7,42 Ecchymosis ofthe skin at the insertion site is common as wellas intraparenchymal hemorrhage localized tothe biopsy cavity, but clinically significanthematomas that interfere or complicate subse-quent surgery or follow-up are rare.

DIFFICULTIES IN STEREOTACTICBREAST BIOPSIES

It is important for the physician to anticipatethat some patients and some lesions will be dif-ficult for obtaining biopsy specimens. Certainlesion characteristics, such as low-density nod-ules, faint or nonclustered microcalcifications,or vague asymmetric densities, may be difficultto visualize with digital imaging despite post-processing features. The position of certainlesions, such as those that are very superficial,those against the chest wall, or those in the axil-lary tail of the breast, may require innovativepositioning by the experienced technologist.However, some lesions may be inaccessible. It

is essential that the physician be able to recog-nize and correct for targeting errors (Figures4–14A to 4–14C). Certain patient characteris-tics will interfere with the success of a stereo-tactic breast biopsy. Patients with neurologic ormusculoskeletal conditions may not toleratepositioning on the stereotactic table. Patientsthat are coughing because of an acute orchronic respiratory condition will increasebreast motion and lesion movement, which mayinterfere with accurate targeting. Patients witha high level of anxiety, especially those suffer-ing from claustrophobia or agoraphobia, mayrequire sedation. As any biopsy has the poten-tial for bleeding complications, those patientswith a history of bleeding abnormalities or whoare taking anticoagulants will require correc-tion prior to biopsy. The small or ptotic breastcreates one of the most common difficulties instereotactic breast biopsy. A breast that flattensto a marginal thickness in compression maylead to “stroke margin problems.” The strokemargin is defined as the distance in millimetersfrom the postfired biopsy needle/probe to theback of the breast or the rear image receptor.The stroke margin must be greater than zero onthe Lorad Stereoguide™ system or greater thana positive 4 mm on the Fischer Mammotest™

table (Figure 4–15). When a breast is very thin,or the lesion is more posteriorly positioned, anegative stroke margin may be encountered.This situation will result in the biopsy needle orprobe striking the rear image receptor andpiercing the back of the patient’s breast skin.


Figure 4–13. Postprocedure images after six samples withthe vacuum-assisted device illustrated the air-contrast cavityand the majority of calcifications removed.

78 BREAST CANCER

Figure 4–14. A, Correct pre- and postfire needle/probe alignment with the lesionresult in favorable tissue sampling. B, An “X” targeting error will result in the needle/probe being off to one side of the lesion. A “Y” targeting error will result in the needle/probe being above or below the lesion. C, A “Z” axis targeting error will illustrate theneedle/probe being too far in front or too far past the lesion. All three targeting errorsbecome less important with the vacuum-assisted device.

A

B

C


Several mechanisms for correcting stroke mar-gins are available. The most commonlyemployed is pulling back the prefire position ofthe needle/probe several millimeters until thecalculated stroke margin is adequate (Figure4–16). Other methods for dealing with strokemargin difficulties include, but are not limitedto, taking a different approach to the breastlesion, using a shorter “throw” biopsy instru-

ment, manual insertion of the Mammotome™

probe, use of the lateral arm, and implementa-tion of a double-paddle technique.

BREAST ULTRASONOGRAPHY

Breast ultrasonography is an effective diagnostictool when used in the proper clinical setting.Appropriate indications for breast ultrasonog-

Figure 4–15. With an adequate stroke margin, the back of the breast and image receptor areprotected. Ignoring a negative stroke margin will result in patient discomfort and damage to theimage receptor.

Figure 4–16. The most common method for correcting for an inadequate strokemargin involves reducing the depth of the needle/probe by “x” mm.

80 BREAST CANCER

raphy have been somewhat controversial. Alower comfort level with hands-on breast ultra-sonography and a heavy reliance on ultrasonog-raphers has, until recently, minimized the impor-tance of breast ultrasonography. The availabilityof computer-enhanced imaging and high-fre-quency transducers have allowed the indicationsfor breast ultrasonography to move well beyonddistinguishing cystic from solid lesions.47,48

Diagnostic breast ultrasonography may beused to evaluate palpable abnormalities, espe-cially in a radiographically dense breast. Mam-mographically indeterminate lesions will notonly be evaluated for cystic versus solid nature,but the sonographic characteristics may assistin distinguishing a lesion’s benign or malignantnature. Diagnostic breast ultrasonography willassist the clinician in evaluating the patient thatis pregnant or lactating. Postoperative follow-up for both benign and malignant disease maybe enhanced with ultrasonography, includingmonitoring and management of seromas andhematomas. The chest wall and the conservedbreast can be assessed for recurrent disease.The axilla may be scanned for preoperativestaging. Associated with every diagnostic indi-cation is the ability to use ultrasonography toguide interventional procedures.

The goals of diagnostic breast ultrasonog-raphy are to reduce the number of benign biop-sies by recognizing simple cysts and areas offibroglandular tissue, which may clinically pre-

sent as a “thickening” or a mammographic asym-metry. When a focal abnormality is identified, anultrasound-guided percutaneous biopsy will pro-vide an efficient and cost-effective diagnosis.

Breast ultrasonography can easily be inte-grated into the surgeon’s practice as a directextension of his clinical breast evaluation. Thepatient in the supine position with the ipsilateralarm raised, much like the clinical breast exami-nation, is ready for ultrasound examination(Figure 4–17). The 7.5- to 10-MHz linear arraytransducer is used for a targeted diagnosticultrasound examination of a clinical abnormal-ity. Whole breast ultrasound is usually done fortwo reasons:

1. In the patient with a suspected breast cancerto look for other areas of involvement;

2. High-risk patient with a clinically difficultbreast and normal or non-specific mammo-gram to look for an occult lesion.

The scanning techniques commonly usedinclude the radial scan, transverse sweepingscan, and tangential scan of the nipple. Theaxilla may also be scanned, especially in associ-ation with a known malignancy. The normalsonographic anatomy of the female breast thatcan be routinely imaged includes skin, subcuta-neous fat, breast parenchyma, Cooper’s liga-ment, retromammary fat, pectoral muscle andfascia, ribs, and pleura.49 Diagnosis of ultra-sound abnormalities is dependent on familiaritywith the normal sonographic breast anatomy andthe recognition of patterns of different breasttypes (Figure 4–18).

Indications for Intervention

A symptomatic or enlarging cyst, whether pal-pable or nonpalpable, is a common indicationfor intervention. Typically, a symptomatic, pal-pable cyst is aspirated by inserting a needleunder palpation guidance and withdrawingfluid until the lesion is no longer palpable.Ultrasound guidance for aspiration may be justas appropriate for the palpable lesion as it is

Figure 4–17. The patient is positioned supine, with the ipsi-lateral arm raised for ultrasound scanning. The position maybe altered by rolling the patient and/or propping the patientwith a pillow to allow scanning through the thinner portionsof the breast.

Figure 4–18. Normal breast anatomy.

Figure 4–19. A, This anechoic, smooth-walled lesion withreverberation artifact is being prepared for aspiration, usingan attachable needle guide. The lesion is aligned with thepuncture lines on the ultrasound monitor. B, Postcyst aspira-tion documenting complete resolution.

A

B


necessary for the nonpalpable lesion.50,51 Ultra-sound guidance may assist in getting the needleinto a thick-walled cyst, documenting any cystwall irregularities and confirming total cystcollapse (Figures 4–19A, 4–19B).

Cysts not meeting the criteria for a simplecyst require aspiration. A lesion with a mixedinternal echo pattern and posterior enhancementsuggests the presence of fluid versus a solidlesion, and an aspiration may be attempted priorto a core-needle biopsy.52 The aspiration of thick,paste-like material, frequently found with mam-mary duct ectasia, might require a local anes-thetic and the use of a larger-gauge (18 g) needle.

The presence of a nonpalpable, solid lesionis an indication for an ultrasound-guided core-needle biopsy to obtain a histologic diagnosis.53

Once again, ultrasound guidance is advanta-geous for guiding a core-needle biopsy for thepalpable solid lesion as well as the nonpalpablelesion. Ultrasound visualization and documen-tation of the needle within the biopsy lesionenhances the accuracy. For the solid lesion withsmooth margins, a homogeneous internal echopattern and ellipsoid shape, such as a fibroade-noma, core-needle biopsy will achieve a spe-cific benign diagnosis. For the more suspiciouslesion with jagged edges, nonhomogeneous andirregular shadowing considered to be suspi-cious for carcinoma, the diagnosis may be

obtained in a cost-effective, efficient manner inthe office setting.

Ultrasound-guided aspiration and/or biopsywill assist in the management of postoperativecomplications, such as seromas or hematomas.With the increasing use of breast conservationtherapy, the architectural distortion identified atthe lumpectomy site may be ideally suited forfurther evaluation with image-guided breastbiopsy. Ultrasound-guided FNA of clinical

adenopathy may assist in staging or evaluationof recurrent disease. With image-guided aspira-tion and drainage, frequently a part of the sur-gical management of intracavitary abscesses, itonly makes sense that a more conservativeapproach should be considered for breastabscess. Ultrasound guidance is advantageousfor the aspiration of pus or the insertion of acatheter for drainage and also to monitor theresolving abscess.

Technique for Ultrasound Guidance

Whether guiding a 25-gauge needle for the aspi-ration of a simple cyst, a needle for the insertionof a local anesthetic, a wire for localization, or a14-gauge needle for core-needle biopsy, theprinciples for ultrasound-guided biopsy remainthe same. It must be remembered that an ultra-sound image visualized on the ultrasound mon-itor represents a “slice” of the breast of approx-imately 1.5 mm in thickness. When scanningbrings the lesion requiring intervention intofocus, we once again are only seeing a 1.5-mm

thick slice of this lesion in its location within thebreast. This ultrasound plane is parallel to thelong axis of the ultrasound transducer. There-fore, the intervening instrument must be guidedalong the long axis of the transducer. When thetip of the needle is seen advancing toward thelesion and penetrates that lesion, we know thatthe needle is in the same 1.5-mm plane as theslice of the lesion being visualized and thereforewithin the lesion (Figure 4–20).54–56

To aspirate an enlarging or symptomaticcyst, one simply wipes the skin with an alcoholpreparation and inserts a small needle (25-, 22-,20-gauge); feedback is immediate when fluid isseen in the syringe. When using a larger-gaugeneedle for aspiration of a complex cyst, a localanesthetic may be injected with a small-gaugeneedle (30-, 27-, 25-gauge) and aspiration per-formed with an 18-gauge needle. Cytologicevaluation of the aspirated fluid is reserved forbloody fluid or lack of cyst resolution.55

Ultrasound core-needle biopsy for histo-logic diagnosis requires more planning. Whenthe lesion requiring a core-needle biopsy isidentified, the skin is marked at the edge of thetransducer for the proposed insertion site of theneedle. The optimal insertion site and approachto the lesion is the shortest skin-to-lesion dis-tance. For best cosmesis, care should be takento avoid placing the scar of the insertion sitenear the inner portion of the breast. At thispoint, the breast is prepared with an appropriateantiseptic solution. The transducer is likewiseprepared, or a sterile sleeve may be placed overthe transducer. Sterile ultrasound gel is avail-able in individual packets. The local anestheticis then injected under direct ultrasound visual-ization. The skin is anesthetized as well as thetrack leading to the lesion. In addition, the localanesthetic is applied above, below, and to theopposite side of the lesion. The use of ultra-sound visualization limits obscuring the lesionwith the administration of too much local anes-thesia in any one area. A small skin incision ismade at the transducer edge using an 11-bladescalpel. The core biopsy needle is then inserted

82 BREAST CANCER

Figure 4–20. The ultrasound plane is approximately 1 to 1.5 mm thick. The biopsy needle must remain within this thinultrasound plane at the same time that the lesion is main-tained within that same plane to confirm accurate biopsy.

into the breast, and the lesion is approachedwith ultrasound guidance. The needle may beguided freehand or directed by an attachableneedle guide that is available from certain ultra-sound manufactures.49,53,54 The attachable nee-dle guides assist the neophyte in maintainingthe proper needle alignment within the narrowultrasound plane to ensure accurate lesion sam-pling. Visualization is maximized by keepingthe needle at a more shallow angle and there-fore parallel to the sole of the transducer. Theshallow angle of the needle will also minimizethe chance of penetrating the chest wall andsubsequently the pleura. The freehand methodof biopsy allows greater flexibility in the needleinsertion site and angle of approach of the nee-dle to the breast lesion. With either method, theneedle is observed and the information docu-

mented as it approaches the lesion. When theposition of the needle is confirmed at the frontof the lesion, the needle/gun combination isfired.53,57,58 There are several needle/gun com-binations available, some of which are dispos-able and others that use disposable needleswithin a more permanent housing (gun). Themechanism of tissue acquisition is similar withthe automated movement of an inner sheaththat contains a sampling notch followed imme-diately by an outer sheath that cuts the tissuefree. This action is accomplished under directultrasound visualization (Figure 4–21). Theneedle is then withdrawn, and the specimen istransferred onto a moistened telfa pad. Addi-tional tissue samples are taken in the samemanner through the existing incision until thelesion is adequately sampled. Approximately


Figure 4–21. A, With the ultrasound transducer and breaststabilized with the nondominant hand, the biopsy instrumentis advanced forward through a small skin incision maintain-ing orientation of the needle in the long axis of the trans-ducer. B, Documenting alignment and penetration under real-time ultrasound guidance.A

B

three to five tissue samples are obtained,depending on the quality of the samples and theconfirmation of the needle penetrating thelesion. When adequate sampling is achieved,manual compression is applied. The incisionmay be approximated with a Steri-strip™, and aTegaderm™ dressing is applied (Figure 4–22).

The accuracy of ultrasound core-needlebiopsy has been widely documented.53,54,56,59

Staren reviewed his series of more than 1,000consecutive diagnostic ultrasound scans. Ofthese, 210 patients underwent ultrasound-guided core biopsy of nonpalpable, mammo-graphically detected lesions using a 14-gaugeneedle. Symptomatic cysts underwent ultra-sound-guided FNA. Lesions characterized asfibroadenoma, indeterminate, or suspiciousunderwent ultrasound-guided FNA and/or core-needle biopsy. There were no false positivesand the false-negative rate was 3.6 percent.Small lesions were noted with ultrasound char-acteristics that warranted biopsy on diagnosticgrounds alone. Staren concluded that ultra-sound-guided aspiration and/or biopsy couldaccurately diagnose nonpalpable, mammo-graphically detected breast masses.

SUMMARY

When patients are referred with mammographicabnormalities requiring further evaluation, themammographic lesion is evaluated to determineif the work-up is complete. When the abnormal-ity is determined to require a biopsy, options arepresented to the patient. The options presentedshould include traditional, open surgical biopsyand percutaneous, image-guided breast biopsy.The option of image-guided breast biopsy mustinclude a discussion of monitoring and follow-up. If this type of breast biopsy is acceptable,then the physician must choose the most appro-priate method of imaging to guide the biopsy.

Microcalcifications, which cannot be visu-alized with the current ultrasound technology,will require stereotactic guidance. Certainnodular densities, architectural distortions, andasymmetric densities without ultrasound find-ings will also be amenable to stereotacticbiopsy.45,53,54 When both mammography andultrasonography visualize a lesion, such as asolid nodular density, ultrasonography is thepreferable method for image guidance.52,53,55

The real-time nature of ultrasound imagingprovides increased accuracy and is more costeffective. In addition, ultrasound-guided biop-sies are more comfortable for the patient. Thepatient may lie supine, and local anesthetic maybe injected under direct visualization. This con-trasts with stereotactic breast biopsy, where thepatient lies prone on a stereotactic table withher breast in compression for the entire proce-dure and her neck hyperextended. Also, theability to guide the injection under direct visu-alization resolves concerns associated with theliberal use of local anesthesia and the potentialfor obscuring or moving a lesion undergoingstereotactic core-needle biopsy.

The vast majority of nonpalpable lesionsrecommended for biopsy are evaluated withpercutaneous image-guided breast biopsy. If abenign diagnosis is obtained, no further work-up is recommended, and the patient is placed ina follow-up protocol. A specific benign diagno-

84 BREAST CANCER

Figure 4–22. The skin incision is re-approximated using aSteri-strip™ after manual compression has been applied forhemostasis. A dressing may then be applied.

sis (fibroadenoma) requires only a return toroutine screening. With microcalcifications ornodular densities with a less specific benigndiagnosis, short-term mammography in 4 to 6months is recommended.31

The obvious indication to proceed withopen surgical excision is an image-guidedbiopsy for malignancy and atypical hyperpla-sia. Medical judgment or lack of pathologic andradiologic diagnostic concordance would alsobe sufficient cause for further intervention. Theissue of pathologic concordance with a suspi-cious mammographic lesion has fueled adebate over the indication to perform an image-guided breast biopsy on a highly suspiciouslesion. Histologic confirmation assists inpatient planning and allows wider excision forclear margins at the first surgical setting.Image-guided breast biopsy of a suspiciouslesion may bypass open biopsy altogether forthose patients that require a mastectomy andare not candidates for breast conservation. His-tologic confirmation of an obvious cancer withimage-guided technology leaves a tumor in situto aid in the successful performance of the sen-tinel lymph node biopsy procedure.60 Concernsover potential added cost with image-guidedbiopsy are exaggerated.

As technology advances and the approach tobreast cancer treatment evolves, additionalindications for image-guided biopsy of suspi-cious lesions will emerge. For many, image-guided percutaneous breast biopsy has perma-nently altered the management of nonpalpablebreast disease. Image-guided percutaneousbreast biopsy will provide the stage for achiev-ing nonoperative histologic diagnosis and thepotential for future therapeutic modalities.

REFERENCES

1. Kopans DS. The positive predictive value of mam-mography. AJR Am J Roentgenol 1992;158:521–6.

2. Sailors DM, Crabtree JD, Land RL, et al. Needlelocalization for non-palpable breast lesions.Am Surg 1994;60:186–9.

3. Wilhelm NC, DeParedes ES, Pope RT. The chang-ing mammogram: a primary indication for nee-dle localization biopsy. Arch Surg 1986;121:1311.

4. Miller RS, Adelman RW, Espinosa MH, et al. Theearly detection of nonpalpable breast carci-noma with needle localization. Experiencewith 500 patients in a community hospital. AmSurg 1992;58:193–8.

5. Parker SH, Lovin JD, Jobe WE, et al. Nonpalpablebreast lesions: stereotactic automated large-core biopsies. Radiology 1991;180:403–7.

6. Elvecrog EL, Lechner MC, Nelson MT. Nonpal-pable breast lesions: correlation of stereotaxiclarge-core needle biopsy and surgical biopsyresults. Radiology 1993;188:453–5.

7. Parker SH, Burbank F, Jackman RJ, et al. Percuta-neous large-core breast biopsy: a multi-institu-tional study. Radiology 1994;193:359–64.

8. Dershaw DD, Morris EA, Liberman L, et al. Non-diagnostic stereotaxic core breast biopsy:results of rebiopsy. Radiology 1996;198:323–5.

9. Liberman LL, Fahs MC, Dershaw DD, et al.Impact of stereotaxic core breast biopsy oncost of diagnosis. Radiology 1995;195:633–7.

10. Tinnemans JGM, Wobbes T, Hendricks JHCL, etal. Localization and excision of nonpalpablebreast lesions. A surgical evaluation of threemethods. Arch Surg 1987;122:802–6.

11. Norton LW, Zeligman BE, Pearlman MD. Accu-racy and cost of needle localization breastbiopsy. Arch Surg 1988;123:947–50.

12. Kopans DB, Meyer JE, Lindfors KK, McCarthyKA. Spring-hookwire breast lesion localizer:use with rigid compression mammographicsystems. Radiology 1985;157:537–8.

13. Bigelow R, Smith R, Goodman PA, Wilson GS.Needle localization of non-palpable breastlesions. Arch Surg 1985;120:565–9.

14. Landercasper J, Gunderson SB, Gunderson AL, etal. Needle localization and biopsy of nonpal-pable lesions of the breast. Surg GynecolObstet 1987;164:477–81.

15. Homer MJ, Smith TJ, Marchant DJ. Outpatientneedle localization and biopsy for nonpalpablebreast lesions. JAMA 1984;252:2452–4.

16. Feig SA. Localization of clinically occult breastlesions. Radiol Clin North Am 1983;21:155–71.

17. Dowlatshahi K, Gent HJ, Schmidt R, et al. Non-palpable breast tumors: diagnosis with stereo-taxic localization and fine-needle aspiration.Radiology 1989;170:427–33.


18. Schwartz GF, Goldberg BB, Riften MD, D’OrazioSE. Ultrasonography: an alternative to x-rayguided needle localization of nonpalpablebreast masses. Surgery 1988;104:870–3.

19. Kopans DB. Breast Imaging. Philadelphia: Lippincott Williams & Wilkins; 1998.

20. Wilhelm MC, Wanebo HJ. Technique and guide-lines for needle localization biopsy of nonpal-pable lesions of the breast. Surg GynecolObstet 1988;176:439–41.

21. Swann CA, Kopans DB, McCarthy KA, et al.Practical solutions to problems of triangulationand preoperative localization of breast lesions.Radiology 1987;163:577–9.

22. Leeming R, Madden M, Levy L. An improvedtechnique for needle localization biopsies of thebreast. Surg Gynecol Obstet 1993;177:85–7.

23. Fisher B. Reappraisal of breast biopsy promptedby the use of lumpectomy. JAMA 1985;253:3585–8.

24. Fine RE, Boyd BA. Stereotactic breast biopsy: apractical approach. Am Surg 1996;62:96–102.

25. Lovin JD, Parker SH, Leuthke JM, Hopper KD.Stereotactic percutaneous breast core biopsy,technical adaptation, and initial experience.Breast Dis 1990;176:741–7.

26. Parker SH, Lovin JD, Jobe WE, et al. Stereotacticbreast biopsy with a biopsy gun. Radiology1990;176:741–7.

27. Caines JS, McPhee MD, Konak GP, Wright BA.Stereotactic needle core biopsy of breastlesions using a regular mammographic tablewith an adapable stereotaxic device. AJR Am JRoentgenol 1994;163:317–21.

28. Parker SH, Burbank F. State of the art: a practicalapproach to minimally invasive breast biopsy.Radiology 1996;200:11–20.

29. Azavedo E, Svane G, Auer G. Stereotactic fine-needle biopsy in 2594 mammographicallydetected non-palpable breast lesions. Lancet1989;171:373–6.

30. Schmidt RA. Stereotactic breast biopsy. Cancer JClin 1994;44:172–91.

31. Parker SH. When is a core really a core? Radiol-ogy 1992;185:641–2.

32. Dowlatshahi K, Yaremko ML, Kluskens LF, JokichPM. Nonpalpable breast lesions: findings ofstereotaxtic needle-core biopsy and fine-needleaspiration cytology. Radiology 1991;181:745–50.

33. Soo MS. Imaging-guided core biopsies in thebreast. South Med J 1998;91:994–1000.

34. Tocino I, Gaargia B, Carter D. Surgical biopsyfindings in patients with atypical hyperplasia

diagnosed by stereotactic core needle biopsy.Ann Surg Oncol 1996;3(5):482–8.

35. Lieberman LL, Dershaw DD, Rosen PR, et al.Stereotactic 14-gauge breast biopsy: how manycore biopsy specimens are needed? Radiology1994;192:793–5.

36. Lieberman LL, Evans WP, Dershaw DD, et al.Radiography of microcalcifications in stereo-taxic mammary core biopsy specimens. Radi-ology 1994;190:223–5.

37. Meyer JE, Lester SC, Grenna TH, White FV.Occult breast calcifications sampled with large-core biopsy: confirmation with radiography ofthe specimen. Radiology 1993;188:581–2.

38. Israel PZ, Fine RE. Stereotactic needle corebiopsy for occult breast lesions: a minimallyinvasive alternative. Am Surg 1995;61:87–91.

39. Jackman RJ, Nowels KWW, Shepard MJ, et al.Stereotaxic large-core needle biopsy of 450non-palpable breast lesions with surgical cor-relation in lesions with cancer or atypicalhyperplasia. Radiology 1994;193:91–5.

40. Liberman L, Dershaw DD, Rosen PP, et al. Stereo-taxic core biopsy of breast carcinoma: accu-racy of predicting invasion. Radiology 1995;194:379–81.

41. Liberman L, Cohen MA, Dershaw DD, et al.Atypical ductal hyperplasia diagnosed atstereotaxic core biopsy of breast lesions: anindication for surgical biopsy. AJR Am JRoentgenol 1995;164:1111–3.

42. Burbank F, Parker SH, Fogerty TJ. Stereotacticbreast biopsy: improved tissue harvesting withthe mammotome. Am Surg 1996;62:738–44.

43. Berg WA, Kerbs TL, Campassi C, et al. Evalua-tion of 14 and 11-gauge directional vacuum-assisted biopsy probes and 14-gauge biopsyguns in a breast parenchymal model. Radiol-ogy 1997;205:203–8.

44. Burbank F. Stereotactic breast biopsy of atypicalductal hyperplasia and ductal carcinoma in insitu lesions: improved accuracy with direc-tional vacuum-assisted biopsy. Radiology1997;202:843–7.

45. Jackman RJ, Burbank SH, Parker SH, et al. Atyp-ical ductal hyperplasia diagnosed at stereotac-tic breast biopsy: improved reliability with 14-gauge, directional vacuum-assisted biopsy.Radiology 1997;204:485–8.

46. Burbank F, Forcier N. Tissue marking clip forstereotactic breast biopsy: initial placementaccuracy, long-term stability and usefulness asa guide for wire localization. Radiology 1997;205:407–15.

86 BREAST CANCER

47. Dempsy PJ. Breast sonography: historical per-spectives, clinical application and image inter-pretation. Ultrasound Q 1988;6:69.

48. McSweeney MB, Murphy CH. Whole breast sono-graphy. Radiol Clin North Am 1985;23:157–67.

49. Leucht W. Teaching atlas of breast ultrasound.New York: Thieme; 1992.

50. Kopans DB, Meyer JE, Lindfors KK, et al. Breastsonography to guide cyst aspiration and wirelocalization of occult solid lesions. AJR Am JRoentgenol 1984;143:489–92.

51. Meyer JE, Christian RL, Frenna TH, et al. Image-guided aspiration of solitary occult breast“cysts.” Arch Surg 1992;127:433–35.

52. Jackson VP. The role of ultrasound in breast imag-ing. Radiology 1990;177:305–11.

53. Parker SH, Stavros AS. Interventional breastultrasound. In: Parker SH, Jobe WE, editors.Percutaneous breast biopsy. New York: Raven;1993. p. 129.

54. Fornage BD. Interventional ultrasound of thebreast. In: McGahan JP, editor. Interventional

ultrasound. Baltimore: Williams & Wilkins;1990. p. 71.

55. Staren ED. Surgical office-based ultrasound of thebreast. Am Surg 1995;61:619–26.

56. Fornage BD, Coan JD, David CZ. Ultrasound-guided needle biopsy of the breast and otherinterventional procedures. Radiol Clin NorthAm 1992;30:167–85.

57. Staren ED. Ultrasound-guided biopsy of nonpalpable breast masses by surgeons. Ann SurgOncol 1996;3:476–82.

58. Parker SH. Needle selection. In: Parker SH, JobeWE, editors. Percutaneous breast biopsy. NewYork: Raven; 1993. p. 7.

59. Parker SH, Jobe WE, Dennis MA, et al. Ultra-sound-guided automated large-core breastbiopsy. Radiology 1993;187:507–11.

60. Pijpers R, Meijer S, Hoekstra OS, et al. Impactof lymphoscintigraphy on sentinel node iden-tification with technitium-99m colloidal albu-min in breast cancer. J Nucl Med 1997;38:366–8.


89

The basic classification of malignant breast dis-eases has remained relatively unchanged sincethe most recent WHO revision in 1982.1 Theseconditions can be broadly divided into epithelialand nonepithelial lesions, with separation of theformer into in situ and invasive tumors (Table5–1). Although recent studies have shed newlight onto our understanding of the basic biol-ogy and natural history of breast cancer, this tra-ditional classification still retains its relevancefor clinicians involved in the diagnosis andtreatment of malignant breast disease.

DUCTAL CARCINOMA IN SITU

Since, by definition, ductal carcinoma in situ(DCIS) is an atypical proliferation of cells con-fined by an intact basement membrane to theductolobular system of the breast, it cannotcause serious morbidity unless it becomes inva-sive. Thus, the major goal of any pathologicevaluation of a patient with DCIS should be todetermine the level of risk of subsequent inva-sion so that optimal treatment can be offeredand possible over- or undertreatment avoided.

In the premammographic era, pure DCIS wasmost often seen as a mass lesion of high-grade,comedo type and usually treated, appropriately,by mastectomy. In the current clinical setting,however, the vast majority of DCIS present as amammographic abnormality or may be entirelyincidental to the lesion seen on radiography. Asthe pattern of the disease has shifted over the

5

Histopathology of Malignant Breast DiseaseROBERT A. GOLDSCHMIDT, MD

years from the bulky mass with a high risk ofinvasion to minute foci of questionable clinicalsignificance, numerous studies have been under-taken to identify prognostic factors and optimizetherapy for the individual patient.

The most important change in our conceptof DCIS was from the monolithic view of a sin-gle disease highly likely to invade if leftuntreated to the realization that DCIS repre-sents a nonobligate precursor with variable riskof progression, depending on a combination offactors. These factors include histologic pattern(and by extension histologic grade), lesion size,margin status, and ancillary studies such as pro-liferation markers and c-erbB-b2.

Classification and Grading of Ductal Carcinoma In Situ

Although the traditional classification of DCISbased on architectural pattern is now recog-

Table 5–1. MALIGNANT DISEASES OF THE BREAST:EPITHELIAL TUMORS

In Situ Lesions Invasive Lesions

Lobular carcinoma in situ (LCIS) Invasive lobularDuctal carcinoma in situ (DCIS) Ductal

comedo no special type (NST)micropapillary tubularcribriform mucinoussolid medullarypapillary invasive

cribriformpapillary

nized to be limited in terms of prognostic value,it remains in use by many pathologists and mer-its review for that reason. In addition, beforeprognosis can be assessed, the pathologist mustestablish a diagnosis by applying criteria to sep-arate DCIS from other atypical or nonatypicalproliferations. These criteria are almost purelyarchitectural and include the solid, cribriform,micropapillary, papillary, and intraductal come-docarcinoma variants.

Comedo DCIS is the only type likely to pre-sent as a palpable mass and accounts for themajority of cases of DCIS diagnosed before theadvent of mammography. As will be seen, it isalso the most likely to be high grade and assuch, the most likely to be associated with con-current or subsequent invasion. The histologicfeatures of comedo DCIS are solid growth withcentral necrosis, often with calcification (Fig-ure 5–1). Marked nuclear atypia is often seen,but a recent consensus conference on the clas-sification of DCIS2 allows use of the comedodesignation in the absence of high nucleargrade. Marked fibrosis and elastosis of the sur-rounding stroma are frequently present, as is anassociated periductal lymphocytic infiltrate.Not infrequently, a question of possiblemicroinvasion arises as small nests of tumorcells are trapped in the fibroinflammatory reac-tion encircling the affected ducts. Diagnostic

criteria for microinvasion are not well estab-lished, nor is its clinical significance, but atpresent it is probably best to limit such a diag-nosis to those cases where single tumor cellsare clearly evident outside an affected duct.

Micropapillary DCIS, sometimes referredto as “clinging carcinoma,” may vary in appear-ance from a relatively flat proliferation withshort projections to a pattern of long, slenderepithelial fronds lacking fibrovascular cores(Figure 5–2). Peripheral spaces formed by so-called “Roman bridges” are common and leadsome to group micropapillary and cribriformlesions together. Key features of micropapillarycarcinoma distinguishing it from proliferationsthat appear similar but are benign aremonomorphism and lack of polarity. Myoep-ithelial cells should be absent from their usualperipheral location, and the atypical cellsshould be homogeneous in appearance.Micropapillary carcinoma is usually composedof cells with low-grade nuclei, although caseswith high nuclear grade are not rare. Centrallylocated necrotic debris and microcalcificationmay be present, especially in cases with high-grade nuclei. There is some evidence thatmicropapillary DCIS may be more likely toinvolve multiple quadrants than other forms ofDCIS.3 Further studies will be necessary toestablish the clinical significance of this finding.

90 BREAST CANCER

Figure 5–1. DCIS, comedo type, with central necrosis. Thiswould qualify as a high-grade lesion based on nuclear mor-phology (original magnification x400).

Figure 5–2. Micropapillary DCIS. Left, dilated ducts with pap-illary tufting of the epithelium (original magnification x100).Right, the micropapillary structures show stratification and lossof polarity (original magnification x200). This is a low-gradelesion based on nuclear grade and lack of necrosis.

Figure 5–4. Solid DCIS. There is a monotonous proliferationof cells filling the duct (x200). Lack of necrosis and non high-grade nuclear features qualify this as a low-grade lesion.

Figure 5–3. Cribriform DCIS with central necrosis (x400).This lesion is intermediate grade based on necrosis and low-grade nuclear features.

Cribriform DCIS describes a lesion charac-terized by the formation of secondary microlu-mens. These lumens tend to be round and uni-formly distributed, although some variability isacceptable. The classic term used to describethe monomorphous nature of the cells outliningthe spaces is “rigid bridges,” referring to thelack of stretched or elongated cells separatingglandular spaces (Figure 5–3). Nuclear mor-phology is typically low grade, although high-grade variants exist. Likewise, necrosis is com-monly encountered in cribriform DCIS andmay be so prominent as to mimic comedoDCIS. Fortunately, as will be seen in the dis-cussion on grading, this distinction is obviatedin current classification systems.

Solid DCIS, as the name implies, consists ofa solid proliferation of neoplastic cells filling aductal structure (Figure 5–4). Nuclear grademay vary widely, and spotty necrosis may beencountered. As in the other forms of DCIS, themonomorphous nature of the cell population isthe hallmark of the process.

Papillary DCIS is the least common of thewell-described variants and exhibits prominentpapillary features with fibrovascular cores butno myoepithelial cell layer (Figure 5–5). As inother variants, a monomorphous cytologicappearance is essential to the diagnosis.

As previously alluded to, DCIS grade has alargely supplanted pattern as the most impor-tant guide to clinical behavior and treatment.There have been a variety of different systemsproposed, but all include some assessment ofnuclear features combined with other factors,typically necrosis or cell polarity, with separa-tion of cases into either two or three grades.Two recent studies4,5 have compared a numberof different grading systems for their interob-server reproducibility. Although none of thesystems demonstrated a high degree of agree-ment among reviewing pathologists, both stud-ies found the Van Nuys system of Silversteinand Lagios6 to be the most reproducible. Thisscheme relies on the distinction of three nucleargrades based on size, texture and nucleoli, and

on the presence or absence of comedo-typenecrosis. Using these parameters, tumors canbe divided into three groups. Group 1 (low-grade) includes those tumors with either low-or intermediate-grade nuclei and no necrosis.Group 2 (intermediate-grade) describes thosetumors with low- or intermediate-grade nucleiand comedo necrosis, whereas Group 3 (high-grade) encompasses all tumors with high-gradenuclei regardless of necrosis.

Tumor grade has emerged as a significantprognostic factor for risk of recurrence, althoughother pathologic features may also be important.The Van Nuys grading scheme is part of a prog-nostic index for DCIS that includes tumor grade,tumor size, and margin width. Evaluation of thelatter two factors in DCIS can be problematic,

Histopathology of Malignant Breast Disease 91

however. Since DCIS only rarely forms a grosslyvisible mass, measurement of lesion size is typi-cally done from microscopic slides. If, as is oftenthe case, tumor is present on more than oneslide, the pathologist must be able to reconstructthe specimen to estimate size. This requires thatthe sections be submitted in an orderly fashion topermit reconstruction. Even so, it is sometimesdifficult to know how to report lesion size whensmall foci of DCIS are scattered throughout alumpectomy specimen, and such data are notreadily available in existing clinical studies.

If tumor size assessment is occasionally aproblem, margin width determination is evenmore of one. The most common approachinvolves the application of colored ink(s) to thesurface of a specimen that has been oriented bythe surgeon. The specimen is then submitted formicroscopic examination in serial transversesection and the shortest distance between tumorand ink reported as the margin width. Since thismethod can only examine a tiny fraction of theactual surface area of the specimen, it is a crudemeasurement at best. Some workers recom-mend an alternate method in which sections areshaved tangentially from the surface of thespecimen to permit wider sampling of the mar-gins. Yet another technique advocated by somesurgeons is the separate removal of shaved mar-gins from the biopsy cavity after the specimenhas been resected. Ultimately, selection of a

method of margin examination will rely on theexperience and preference of the pathologistand surgeon, at least until better clinical studiesare available. Since most “recurrences” ofDCIS probably represent persistence followingincomplete removal, the issue of margins is notof trivial importance. Routine specimen mam-mography is often helpful in guiding the patho-logic sampling by identifying areas of suspi-cious calcification near resection margins.

A recent study published by Silverstein andLagios7 uses a retrospective analysis of DCISpatients to demonstrate that margin width is theonly significant risk factor for local recurrence8.In their series, there was no difference in recur-rence rate between patients who received post-operative radiation and those treated by lumpec-tomy only, so long as there was an uninvolvedmargin width of 10mm or more in all directions.In both the radiated and non-radiated groups,the risk of recurrence after 8-year follow-up was3–4%. Tumor size, nuclear grade, and presenceof comedonecrosis did not alter relative risk inthese cases. For margin widths of 1-10mm,therecurrence rates were higher (12–20%), butthere was still no statistically significant benefitto radiation therapy in this group. Only whenfinal margin width was < 1mm did postopera-tive radiation show a significant reduction inrecurrence rates.

Given the above considerations, the pathol-ogy report in cases of DCIS must include alarge amount of data. Many institutions havefound that the use of a template form ensuresall vital data is present. Such a form would typ-ically include features such as nuclear grade,pattern, presence of necrosis, distance to mar-gin, lesion size, presence of calcifications, andany other parameters deemed to be important.This type of systematic reporting scheme hasthe added advantage of making any retrospec-tive clinical studies much easier to perform.

LOBULAR CARCINOMA IN SITU

Lobular carcinoma in situ (LCIS) is, by defini-tion, a microscopic process and as such is

92 BREAST CANCER

Figure 5–5. Papillary DCIS. Left, an intraductal proliferationof papillary structures with fibrovascular cores (original mag-nification x100). Right, high power shows nuclear atypia andstratification (original magnification x200).

Figure 5–6. Lobular carcinoma in situ. The acini are filledand distended by a monotonous proliferation of small cellswith bland nuclear features (original magnification x100).

almost always an incidental finding seen inassociation with some other gross or mammo-graphic abnormality. In its most classic form,described by Foote and Stewart in 1941,8 it isreadily identifiable by the general pathologist.The lobular acini are filled and distended by apoorly cohesive proliferation of cells withround, rather bland nuclei, scant cytoplasm, andinconspicuous nucleoli (Figure 5–6). Problemsin interpretation of these lesions can arise, how-ever, from variations in qualitative and quanti-tative aspects of the neoplastic process. Forexample, the cytologic features of the neoplas-tic cells may demonstrate more variability, withpleomorphic nuclei and abundant cytoplasm. Insuch cases, it is often difficult to determinewhether one is dealing with lobular carcinomain situ or so-called “lobular cancerization” by anin situ duct carcinoma. Since there is no reli-able marker to distinguish the origin of such alesion as either lobular or ductal, the distinctionmust ultimately be made by the pathologist’sassessment of the light microscopic features.

The situation may be further complicated bythe frequent association of typical ductal carci-noma in situ and LCIS in the same biopsy. Theidentification, when present, of intracytoplas-mic lumens and associated mucin globules inthe atypical cells is helpful. The finding of suchcells in a proliferative lesion involving the ter-minal duct/lobular unit is compelling evidencefor lobular carcinoma in situ.

Another common feature, seen in up to 75percent of patients with LCIS, is a pattern of so-called pagetoid spread of atypical cells alongsmall ductules and occasional larger ducts. Typ-ically, this manifests as a proliferation of atypicalcells lining the duct, with an overlying layer ofintact ductal epithelium. In the atrophic breast ofthe postmenopausal patient, the latter patternmay be the only evidence of LCIS. Some expertsdisagree as to whether this appearance repre-sents true ductal involvement rather than an“unfolding” of the lobule yielding a pseudoduc-tular appearance. Recognition of the process,however, is more important than the terminol-ogy. The important distinction is between true

LCIS and ductal carcinoma in situ (DCIS)involving lobular units, since the therapeuticimplications of these lesions may be quite differ-ent. In most cases of in situ duct carcinomainvolving the terminal duct/lobular unit in a sec-ondary fashion, there is at least some retentionof features suggestive of ductal origin. The pres-ence of large pleomorphic nuclei and/or sec-ondary lumens in such a proliferation wouldfavor DCIS over LCIS. Likewise, as previouslymentioned, identification of intracytoplasmiclumens with mucin droplets strongly suggestsLCIS. Occasionally, a case will defy classifica-tion, even after exhaustive examination of thespecimen. If there are foci of coexistent invasivecarcinoma or unequivocal DCIS, the distinctionbecomes one of academic interest only, as treat-ment would be dictated by those lesions. With abiopsy containing only an in situ lesion of inde-terminate etiology, however, communicationbetween the pathologist, surgeon, oncologist,and patient is essential for optimal care.

INVASIVE BREAST CARCINOMA

Invasive (infiltrating) breast carcinoma can bebroadly subdivided into ductal and lobular cat-egories, with a number of recognized variantsof each. Although current evidence suggeststhat the majority of invasive cancers arise fromcells of the terminal duct lobular unit, theirwide variation in appearance and clinical pre-


sentation continues to make subtyping a usefulexercise until some better method of predictingbehavior becomes available.

Infiltrating Lobular Carcinoma

Infiltrating lobular carcinoma (ILC) is gener-ally considered to account for up to 16 percent9

of invasive breast cancers, depending on thestudy population and the rigidity of diagnosticcriteria. It may present as a scirrhous massgrossly and mammographically indistinguish-able from infiltrating ductal carcinoma,although it is often more insidious, with onlyvague gross findings and occasionally negativemammographic appearance. In contrast to inva-sive ductal carcinoma, invasive lobular cancerpresents as a less distinct tumor that is lessapparent on physical examination and mam-mography. As a result, the microscopic extentof disease is often much greater than grosslyappreciated, and clear lumpectomy margins aresomewhat more difficult to achieve.10 The clas-sical microscopic description, generally cred-ited to Foote and Stewart,11 is of a diffuselyinfiltrative tumor composed of cells with small,round nuclei with minimal pleomorphism ormitotic activity. Intracytoplasmic lumina yield-ing a signet-ring appearance are often presentbut are not pathognomonic since they may beseen in ductal carcinomas also. Linear files(“Indian files”) of infiltrating tumor cells arethe most characteristic pattern of invasion,

often swirling around native ductal structures ina so-called “targetoid” fashion (Figure 5–7).

There are a number of variants of infiltrat-ing lobular carcinoma that have been described,and many tumors show mixtures of two or moretypes. Most of these variants consist of cellswith the same cytologic features as the classicaltype but different patterns of growth such asalveolar, solid, or tubulolobular. While it isunclear whether identification of these variantshas clinical significance, the pleomorphic typedoes merit separate distinction. This variantconsists of cells which infiltrate in the samemanner as classical ILC but have high-gradenuclei. Several studies have suggested a moreaggressive behavior for these tumors.12,13

Infiltrating Ductal Carcinoma

Invasive breast cancers that do not exhibit thefeatures described above for lobular variantsare considered to be ductal in origin. While thisdistinction is somewhat arbitrary, it is firmlyembedded in the literature of breast cancer andserves as a useful tool for the recognition andsubclassification of malignant breast disease.This large group of tumors accounts for themajority (85 to 95%) of invasive breast cancercases and can be broadly divided into those of“no special type” or “not otherwise specified”(NST, NOS) and “special type” tumors of dis-tinctive appearance and behavior.

Various studies place the percentage of NSTbreast carcinomas at 50 to 75 percent of allinvasive breast cancers.13 The tumors withinthis large group vary widely in appearance andoften contain minor components of special typehistology. An appreciation of the cytologic fea-tures of the tumor cells and of the architecturalpattern of the invasive process is useful whenstudying breast cancer. As will be discussedlater, these factors also form the basis for grad-ing NST tumors. In regard to cytology, thenuclei of NST neoplasms can vary from smalland rather bland in appearance to those exhibit-ing marked enlargement and pleomorphism.Mitotic activity can likewise range from mini-

94 BREAST CANCER

Figure 5–7. Infiltrating lobular carcinoma. The tumor cellsinfiltrate in typical linear files (original magnification x200).

Figure 5–9. Mucinous carcinoma. Left, clusters of tumorcells float in a pool of extracellular mucin (original magnifica-tion x100). Right, in this case, the cells exhibit minimalnuclear atypia (original magnification x200).

Figure 5–8. Tubular carcinoma. Left, the tumor consists of ahaphazard arrangement of small tubular structures (originalmagnification x100). Right, the tumor glands are lined bycells with low-grade nuclei (original magnification x400).

mal to brisk and generally follows nucleargrade. Architectural patterns are typicallydescribed on the basis of degree of gland for-mation, which may be quite prominent or com-pletely absent. In addition, both cytologic andarchitectural features may vary widely within asingle tumor. Other features, such as extensivenecrosis or widespread DCIS, may also haveprognostic importance and should be noted.

Ductal carcinomas of special type include agroup of tumors distinguished from NSTtumors by their unique histologic appearanceand often, less aggressive behavior. Within thisgroup are the tubular, mucinous, medullary,invasive cribriform, papillary, and metaplasticvariants. Tubular carcinoma has become themost commonly diagnosed special type tumorsince the advent of mammography, due to itssmall size and lack of clinical symptoms. Themajority of tubular carcinomas are < 1.0 cm indiameter, accounting for 7 to 21 percent ofmammographically detected lesions in variousstudies.14,15 Precisely defining tubular carci-noma is elusive, particularly regarding theextent of tubule formation required to make thediagnosis. The basic requirement is the pres-ence of tubular structures lined by a single layerof epithelial cells of low-nuclear grade. Thetubule lumens are rounded and/or angulated,and mitosis is rare. The tumor stroma is quitecharacteristic, consisting of a cellular desmo-plastic reaction, often with a central fibrousscar from which the tubules radiate (Figure5–8). Low-grade DCIS is a frequent accompa-niment of tubular carcinoma. As previouslyalluded to, tumors that are not purely tubularare somewhat controversial. Most authors willaccept 90 percent tubular architecture as a min-imum criterion for the diagnosis of tubular car-cinoma, and many use 75 percent as a cut-off.Lesser degrees of tubular differentiation are gen-erally reported as “tubular features” in an NSTtumor. The distinction is not trivial, as severalstudies have shown prognostic differencesrelated to varying degrees of tubule formation.16

Mucinous (colloid) carcinoma of the breastaccounts for one to three percent of invasive

breast cancer and has a distinctive gross andmicroscopic appearance. Such tumors tend to besoft, well-defined, rounded masses with a glis-tening mucoid surface. The tumor cells showminimal pleomorphism and form tight clusterswhich float in pools of extracellular mucin (Fig-ure 5–9). As with tubular carcinoma, 90 percentmucinous morphology is the generally acceptedminimum for designation as mucinous carci-noma, with its associated favorable prognosis.

Medullary carcinoma is the most controver-sial of the special-type tumors, both in terms ofhistologic criteria for diagnosis and subsequentbehavior. Poor intraobserver reproducibility, anddisagreement over the diagnostic requirements


for medullary carcinoma, likely account for thereported frequency range of 2 to 10 percent.17 Inour own practice, medullary carcinoma is a dis-tinctly unusual variant. Disclaimers aside, theclassic description of medullary is of a soft,fleshy, circumscribed tumor with a homoge-neous appearance. Microscopically, the tumorcells have large, often bizarre nuclei, abundantmitoses, and grow in syncytial sheets. Equallyimportant for the diagnosis is the presence of asignificant lymphoplasmacytic infiltrate thatusually occupies narrow bands of fibrovascularstroma within the tumor and may also surroundthe periphery of the tumor. Lastly, microscopiccircumscription is essential. The tumor musthave a smooth, pushing margin without infiltra-tion of surrounding breast tissue or fat by tumorcells (Figure 5–10). Foci of DCIS surroundingthe tumor, however, are not uncommon andshould not preclude a diagnosis of medullarycarcinoma. It is important to adhere to strictdiagnostic criteria because the less aggressivebehavior ascribed to medullary carcinomabelies its high-grade appearance. Attempts todistinguish an intermediate variant18 (atypicalmedullary carcinoma) exhibiting some, but notall, of the classic features and having an inter-mediate prognosis remain controversial.

Invasive cribriform carcinoma is a relativelyinfrequent variant in pure form, although com-bination with tubular carcinoma is not rare. The

cribriform designation comes from its resem-blance to cribriform DCIS, from which it mayoccasionally be difficult to distinguish. The his-tologic appearance is of infiltrating sheets andnests of cells of low-nuclear grade with the typ-ical punched-out spaces seen in its in situnamesake (Figure 5–11). The 90 percent rulefor designation as a special-type tumor gener-ally applies for combinations of invasive cribri-form carcinoma and NST tumors, while tumorscomposed entirely of tubular and invasive crib-riform structures in any proportion still qualifyas special type.

Grading of Invasive Breast Carcinoma

Grading schemes for invasive breast carcinomaoriginated with the work of Greenhough19 60years ago and have not changed substantiallysince that time. It is a testament to the power ofhistological grading of breast cancer that virtu-ally all of the numerous methods and variationsproposed over the years correlate to somedegree with clinical behavior. While no singlesystem has achieved universal acceptance, alluse some combination of nuclear features,mitotic activity, and gland formation. One ofthe more commonly used methods, and thatadopted by the WHO, is the Scarff-Bloom-Richardson system,20 which assigns a score ofone to three points for each of the three above-

96 BREAST CANCER

Figure 5–10. Medullary carcinoma. Left, the tumor is sharplydemarcated from the surrounding fat (original magnificationx100). Right, the tumor cells are highly pleomorphic withnumerous mitoses and a dense lymphoid infiltrate (originalmagnification x400).

Figure 5–11. Invasive cribriform carcinoma. Nests of tumorwith a cribriform configuration infiltrate the stroma (originalmagnification x200).

mentioned parameters. Based on the point total,tumors are assigned to grade I (three to fivepoints), grade II (six to seven points), or gradeIII (eight to nine points) (Figures 5–12 to5–14). Further refinement of this method byElston and Ellis21 defined specific criteria fornuclear grade, architectural pattern, and mitoticrate, and has resulted in a system showing bothstrong correlation with outcome and reasonablygood interobserver reproducibility.

SPECIMEN HANDLING AND REPORTING

As the surgical approach to breast cancer haschanged, so has the pathologist’s approach tospecimen handling. Many lumpectomy speci-mens arrive in the pathology lab with a diagno-sis already established by fine-needle or stereo-tactic core biopsy. If the specimen is oriented bythe surgeon using sutures or some other method,the margins can then be optimally assessed onpermanent section. If the lesion was nonpalpa-ble and removed using stereotactic localization,radiography of the specimen is essential to con-firm the adequacy of excision. Hormone recep-tor assays and other ancillary studies no longerrequire fresh tumor tissue, thus obviating theneed to perform a frozen section, with its atten-

dant problems. Frozen section is entirely appro-priate, however, when confirmation of invasivetumor is necessary prior to concurrent axillarydissection or when the surgeon requires intraop-erative margin assessment. The popularity ofsentinel node techniques using frozen section todetermine whether to proceed with axillary dis-section is also increasing.

In general, all lumpectomy and re-excisionspecimens must be assessed, at a minimum, fortumor size, tumor type, and margin width. Thisrequires input from the surgeon as to specimenorientation, and some method for identifying


Figure 5–12. Invasive ductal carcinoma, no special type.Left, the tumor makes some glandular structures (originalmagnification x200). Right, the nuclei are moderatelyenlarged with inconspicuous nuclei and no mitoses (originalmagnification x400). In the Scarff-Bloom-Richardson system,this tumor would score 1 point for architecture, 2 points fornuclei, and 1 point for mitosis, making it grade I.

Figure 5–13. Invasive ductal carcinoma, no special type.Left, the prominent gland formation scores 1 point for archi-tecture (original magnification x100). Right, high nucleargrade and moderate mitotic activity score 3 and 2 pointsrespectively (original magnification x400). The total score of6 points makes this a grade II tumor.

Figure 5–14. Invasive ductal carcinoma, no special type.Left, this tumor shows no gland formation (original magnifi-cation x100). Right, the nuclei are high grade and shownumerous mitoses (original magnification x400). The totalscore of 9 points makes this tumor grade III.

margins on the histologic sections. Typically,this involves the application of colored inks tothe surfaces of the specimen prior to sectioning.Adequate fixation prior to processing is ofutmost importance in achieving optimal resultsfor histologic examination. Since most breastspecimens are fairly fatty, overnight formalinfixation is often necessary. The attendant delayin reporting is more than offset by the quality ofthe information that can be gleaned from highquality histologic sections.

Because of the large amount of pathologicdata entering into breast cancer prognosis andtreatment strategy, many pathologists find ithelpful to utilize some sort of standardizedreporting format as part of the pathologyreport. As described above for DCIS, we use abreast cancer worksheet to describe relevantfeatures of invasive tumors. Features such astumor size, grade, subtype, and margin statusare described, and this data is included as partof the final report. The design of such a formatshould include input from clinicians to ensurethey receive all information required for opti-mal patient management as well as any datathat may be helpful in future retrospective stud-ies. Standardized forms have been developedby the Association of Directors of SurgicalPathology (ADSP) to serve this purpose.

REFERENCES

1. Azzopardi JG, Chepick OF, Hartmann WH. TheWorld Health Organization histological typingof breast tumors. 2nd ed. Am J Clin Pathol1982;78:806–16.

2. The Consensus Conference Committee. Consensusconference on the classification of ductal car-cinoma in situ. Cancer 1997;80:1798–1802.

3. Bellamy COC, McDonald C, Salter DM, et al.Noninvasive ductal carcinoma of the breast:the relevance of histologic categorization. HumPathol 1993;24:16–23.

4. Douglas-Jones AG, Gupta SK, Attanoos RL, et al.A critical appraisal of six modern classifica-tions of ductal carcinoma in situ of the breast(DCIS): correlation with grade of associatedinvasive tumor. Histopathology 1996;29:397–409.

5. European Commission Working Group on Breast

Screening Pathology. Consistency achieved by23 European pathologists in categorizing duc-tal carcinoma in situ of the breast using fiveclassifications. Hum Pathol 1998;29:1056–62.

6. Silverstein MJ, Lagios MD, Craig PH, et al. Aprognostic index for ductal carcinoma in situ ofthe breast. Cancer 1996;77:2267–74.

7. Silverstein MJ, Lagios MD, Groshen S, WaismanJR, Lewinsky BS, et al. The influence of mar-gin width on local control of ductal carcinomain situ of the breast. NEJM 1999;340:1455–61.

7. Foote F, Stewart F. Lobular carcinoma in situ: arare form of mammary carcinoma. Am J Pathol1941;17:491–6.

8. Elston CW, Ellis IO. Systemic pathology, Vol.13.The breast. Edinburgh: Churchill Livingstone;1998.

9. Yeatman TJ, Cantor AB, Smith TJ, et al. Tumorbiology of infiltrating lobular carcinoma. AnnSurg 1995;222:549–61.

10. Foote FW, Stewart F. A histologic classification ofcarcinoma in the breast. Surgery 1946;19:74–9.

11. Eusibi V, Magalhaes F, Azzopardi JG. Pleo-morphic lobular carcinoma of the breast: anaggressive tumor showing apocrine differentia-tion. Hum Pathol 1992;23:655–62.

12. Weidner N, Semple JP. Pleomorphic variant ofinvasive lobular carcinoma of the breast. HumPathol 1992;23:1167–71.

13. Rosen PP. Breast pathology. Philadelphia: Lippin-cott-Raven; 1997.

14. Cooper HS, Patchefsky AS, Krall RA. Tubular car-cinoma of the breast. Cancer 1978;42:2334–42.

15. Parl FF, Richardson LD. The histologic and bio-logic spectrum of tubular carcinoma of thebreast. Hum Pathol 1983;14:694–8.

16. Ridolfi R, Rosen P, Port A, et al. Medullary carci-noma of the breast. A clinicopathologic studywith 10 year follow-up. Cancer 1977;40:1365–85.

17. Wargotz ES, Silverberg SG. Medullary carcinomaof the breast. A clinicopathologic study withappraisal of current diagnostic criteria. HumPathol 1988;19:1340–46.

18. Greenhough RB. Varying degrees of malignancyin cancer of the breast. J Cancer Res 1925;9:452–63.

19. Bloom HJG, Richardson WW. Histological grad-ing and prognosis in breast cancer. A study of1409 cases of which 359 have been followedfor 15 years. Br J Cancer 1957;11:359–77.

20. Elston CW, Ellis IO. Pathological prognostic fac-tors in breast cancer. I. The value of histologi-cal grade in breast cancer: experience from alarge study with long-term follow-up. Histo-pathol 1991;19:403–10.

98 BREAST CANCER

99

This chapter reviews clinical and pathologic fea-tures of uncommon breast malignancies. Themajority of the data used in the course of writingthe chapter was obtained from small studies ofspecific tumor subtypes, or has been gleanedfrom larger studies that included several types ofmore common breast cancers. Unfortunately,there is often insufficient information availableto draw absolute conclusions regarding therapyand prognosis.

Much of the data cited was collected prior tothe widespread use of breast conservation. Forthis reason, the vast majority of patients studiedwere treated using mastectomy. The reliance onmastectomy has resulted in a lack of informationregarding the natural history and radiosensitivityof many of the tumors presented. Therefore, therisk of local recurrence for patients with rarebreast malignancies opting for breast conserva-tion is unclear. There is, however, no reason tosuspect a significant difference in the risk of localrecurrence in this group of patients compared topatients with more common types of breast can-cer. Except where specifically indicated, the clin-ically appropriate use of breast conservationshould be considered in the informed treatmentof patients with rare breast malignancies.

PAPILLARY CARCINOMA

Papillary carcinoma accounts for 1 to 2 percentof newly diagnosed breast cancers in women,

6

Unusual Breast PathologyDAVID R. BRENIN, MDHANINA HIBSHOOSH, MDDAVID W. KINNE, MD

and a slightly higher proportion in men.1,2,3 Itoccurs in both an invasive and noninvasiveform. The World Health Organization (WHO)defined papillary carcinoma as follows: “A rarecarcinoma whose invasive pattern is predomi-nantly in the form of papillary structures. Thesame architecture is usually displayed in themetastases. Frequently, foci of intraductal pap-illary growth are recognizable.”4 Further, theWHO classification states that “papillary carci-noma arising, and limited to a mammary cyst,is [to be] referred to as noninvasive intracysticcarcinoma.”4 Invasive carcinoma, however, maybe associated with an intracystic carcinoma.5

Papillary carcinoma occurs most frequentlyin the central portion of the breast, and is asso-ciated with a malignant nipple discharge in 22to 34 percent of patients.1,6 The mean age ofdiagnosis for papillary carcinoma, 63 to 67years, is older compared to the more commontypes of breast cancer.1,2,7 The tumors tend togrow slowly, not infrequently being present formore than 1 year prior to patients seeking treat-ment. On physical exam, papillary carcinomasare well-circumscribed, and often lobulated.There may be a bloody nipple discharge. Theaverage clinical size is 2 to 3 cm.3 Clinicallyenlarged axillary lymph nodes are not uncom-mon in patients with larger tumors containingareas of hemorrhagic necrosis. Mammographi-cally, papillary carcinomas typically have sharpmargins and are rounded or lobulated. Breast

ultrasound may reveal a solid component in acyst that otherwise appears benign.

The appearance of the gross tumor varieswith the proportion of the cystic component.Some fibrosis may be present. The cut surface ofthe tumor is typically described as tan or gray,and areas of focal hemorrhage and necrosis arenot uncommon.3 Larger tumors may form a largecyst containing partially clotted blood and tumorfragments. Microscopically, the tumors form apredominately frond-like pattern (Figures 6–1and 6–2). Cystic areas may be present but are nota prerequisite for diagnosis. Distinguishingbetween benign and malignant papillary tumorscan be challenging. Kraus and Neubecker,8

Lefkowitz and colleagues,7 and Rosen3 haveattempted to delineate guidelines for diagnosis.Various immunohistochemical markers havebeen evaluated but have proved to be of littlehelp. Analysis of DNA content, however, hasdemonstrated significant differences betweenpapillary carcinoma and benign lesions.9,10

Papillary carcinoma has a favorable progno-sis. Noninvasive papillary carcinoma is a variantof ductal carcinoma in situ (DCIS), and is asso-ciated with a less than one percent rate of axil-lary metastasis.3,7 To date, there is no significantbody of data addressing the use of radiationtherapy in the treatment of this lesion. As is thecase with DCIS, there are no prospective trialscomparing mastectomy to breast preservation

with whole breast irradiation in patients withnoninvasive papillary carcinoma. The use ofbreast conservation, however, appears reason-able for these patients, as there is no reason tosuspect a significant difference in the risk oflocal recurrence compared to patients with morecommon types of noninvasive breast cancer.The low rate of axillary metastasis observedmakes elimination of axillary dissection appro-priate in patients with noninvasive papillary car-cinoma and a clinically negative axilla.

Even less data exists to aid in treatmentselection for patients with invasive papillarycarcinoma. Fisher and colleagues reported on35 patients with invasive papillary cancer.2 Ofthe 22 patients who underwent axillary dissec-tion, 32 percent were found to have axillarymetastases. Of patients with axillary metastases,only two (nine percent) had four or more lymphnodes involved. Life-table plots calculated byFisher and colleagues showed a favorable prog-nosis comparable to patients with tubular can-cers. At 5-year follow-up, only one patient haddied of papillary carcinoma. Recurrences, whenthey do occur, are typically “late,” coming > 5years after the initial diagnosis.3 The majority ofreports concerning the treatment of invasivepapillary carcinoma have addressed patientswhose primary therapy consisted of mastectomywith or without axillary dissection. When clini-cally appropriate, the use of breast conservation,

100 BREAST CANCER

Figure 6–1. Papillary carcinoma demonstrating frond-likepattern (original magnification x400).

Figure 6–2. Intracystic papillary carcinoma. Note solid cystwall on periphery (original magnification x400).

whole breast irradiation, and axillary dissectionor sentinel node biopsy is a reasonable option.

METAPLASTIC MAMMARY CARCINOMA

Metaplastic mammary carcinoma refers to aclassic breast carcinoma containing a variablecomponent exhibiting a nonglandular growthpattern. These tumors constitute fewer than onepercent of breast cancers.11,12 The metaplasticchanges typically manifest as squamous cells,spindle cells, and/or as areas of heterologousmesenchymal growth showing cartilaginous orosseous differentiation. The histologic diversityobserved in metaplastic mammary carcinomahas led to various subdesignations includingspindle-cell carcinoma, carcinoma with osseousmetaplasia, carcinoma with pseudosarcomatousmetaplasia, squamous cell carcinoma with pseu-dosarcomatous stroma, and carcinosarcoma. Thehistogenesis of these carcinomas is assumed tobe of ductal origin. Results derived from ultra-structural and immunohistochemical studiessuggest that metaplastic mammary carcinomasoriginate from undifferentiated multipotentialcells.13 Tavassoli suggested that myoepithelialcells are the cell of origin.11 Although the num-ber of reported cases is small, all of the subtypesappear to have a similar prognosis14 and will bepresented as a single group in this chapter.

Metaplastic mammary carcinoma typicallypresents as a mass. Skin changes and fixationto underlying tissues have been reported.15 Thegross appearance of the tumor varies with sub-type, but most are described as hard with well-circumscribed borders. Cystic degenerationmay occur when there is an extensive squamousmetaplastic component.14 Histologically, meta-plastic carcinoma is divided broadly intotumors showing squamous and/or heterologous(cartilaginous or osseous) (Figure 6–3) or pseu-dosarcomatous differentiation. The formerappears to be more frequent; however, mixedand transition forms are common. The extentand degree of differentiation varies widely. The

histology of carcinoma at metastatic sites maynot be predicted by the extent and subtype seenin the breast.

The number of reported patients with meta-plastic mammary carcinomas is insufficient todraw accurate conclusions concerning therapyand prognosis. The majority of the data has beenobtained from small studies of specific tumorsubtypes. Most patients underwent mastectomy.Rosen and Ernsberger reported that in four ofseven patients treated with excisional biopsyalone, disease locally recurred between 1 and 3.5years after diagnosis.16 There is no informationconcerning the responsiveness of metaplasticcarcinoma to radiation or chemotherapy.15 Whencompared to the more common histologies,metaplastic mammary carcinoma has a low rateof axillary lymph node involvement.17–19 Distantfailure, however, is common, with an overall 5-year survival rate reported to be 44 percent.13

Given the low rate of axillary metastasis and a lack of prognostic information gained by axillary staging in these patients, elimination ofaxillary lymphadenectomy or the use of sentinellymph node biopsy alone may be appropriate.

APOCRINE CARCINOMA

Apocrine carcinoma of the breast reportedlyaccounts for 0.4 percent of new mammarymalignancies.20,21 This tumor derives its name

Unusual Breast Pathology 101

Figure 6–3. Metaplastic carcinoma with cartilagenous differ-entiation (original magnification x400).

from the apocrine glands normally present inskin. Apocrine carcinomas of the breast, how-ever, do not originate from apocrine glands ofthe skin. Apocrine carcinomas appear to arisefrom the apocrine metaplasia commonly foundin excised breast tissue.3 The histologic similar-ity of apocrine metaplasia commonly found inexcised breast tissue, rare carcinomas withapocrine differentiation, and apocrine glands ofthe skin is due to their common embryologicalderivation from the epidermis.

Apocrine carcinoma presents in a fashionsimilar to other, more common breast cancers.The reported age range of affected patients isfrom 19 to 86 yrs.22–25 Most patients with infil-trating apocrine carcinoma of the breast presentwith a palpable mass.24,25 Abati and colleaguesfound that approximately one-third of both theintraductal and invasive lesions were detectedmammographically.24 Infiltrating apocrine car-cinomas are hard on palpation. Grossly, thelesions are typically gray to white with infiltrat-ing borders.3 Some tumors are cystic or have amedullary appearance.3 Microscopically, thecytoplasm is markedly eosinophilic and may begranular or homogeneous. The cellular architec-ture of both intraductal and invasive apocrinecarcinomas is similar to that seen with morecommon mammary carcinomas. The distinctionbetween atypical apocrine hyperplasia and apoc-rine intraductal carcinoma can be difficult.26–28

The prognosis for patients with apocrinecarcinoma of the breast is generally consideredto be analogous to patients with similarlystaged ductal carcinomas.23–25 Abati and col-leagues identified a 15 percent local recurrencerate in 20 patients with intraductal apocrinecarcinomas treated by biopsy alone, but norecurrences in two patients treated withlumpectomy and irradiation.24 The majority ofreported patients with invasive apocrine carci-noma have been treated with mastectomy andsome form of axillary dissection. The radiosen-sitivity of these lesions has yet to be deter-mined, but the use of breast conserving therapymay be appropriate.

ADENOID CYSTIC CARCINOMA

Adenoid cystic carcinoma of the breast, alsoknown as cylindroma, is a rare neoplasmaccounting for < 0.1 percent of mammary car-cinomas.29,15 First described in the breast byGeschickter30 in 1945 and again by Foote andStewart5 in 1946, its characteristic histopatho-logic appearance is identical to like-namedtumors arising from the salivary glands. Ade-noid cystic carcinomas typically present in thesixth or seventh decade of life. The characteris-tic presentation is that of a 2 to 3 cm movabletumor which may be tender or painful.31 Thelesions tend to be centrally located in thebreast32 and may exhibit skin changes whensuperficial. These tumors have a gray to paleyellow cut surface with well-defined margins.Larger lesions have been found to undergo cys-tic degeneration.31,33 Adenoid cystic carcino-mas of the breast have marked histological het-erogeneity, making diagnosis by needle biopsyproblematic. Examination of many microscopicfields may be required before the classic cylin-dromatous and/or cribriform growth pattern isidentified. Ro and colleagues divided adenoidcystic carcinomas of the breast into three histo-logic grades based on the proportion of solidgrowth to the overall tumor size.34 Tumors withno solid component were classified as grade I,those with < 30 percent solid component weregrade II, and tumors consisting of > 30 percentsolid component were grade III. Ro noted thattumors with a solid component were morelikely to be larger, recur, or metastasize.

Adenoid cystic carcinoma of the breast hasan excellent prognosis. The rate of axillarymetastasis is low, less than one percent.31,34 Dis-tant metastasis is rare.31 When systemic recur-rence does occur, it is typically pulmonary.Metastases to bone,34 liver,34 brain,35 and kid-ney36 have also been reported. Distant metas-tases typically occurr in patients who had nega-tive axillary dissections.31 There is noprospective data to support one therapeuticmodality over another in the treatment of this

102 BREAST CANCER

disease. Reported data on prognosis has beengathered mostly from patients treated with mod-ified radical or radical mastectomy. The use ofbreast conservation, however, appears reason-able for these patients, as there is no reason tosuspect a significant difference in the risk oflocal recurrence compared to patients with morecommon types of breast cancer.37 The low rateof axillary metastasis observed, combined witha lack of prognostic information gained fromaxillary staging, makes elimination of axillarydissection appropriate in patients with adenoidcystic carcinoma and a clinically negative axilla.

SQUAMOUS CELL CARCINOMA

Squamous cell carcinoma of the breast is anextremely rare form of metaplastic carcinomaconsisting of a lesion entirely, or nearly entirely,composed of keratinizing squamous cell carci-noma. Typically, lesions composed of > 90 per-cent keratinizing squamous carcinoma havebeen placed in this group. One must be carefulto exclude a metastatic squamous cell carci-noma or skin carcinoma involving the breastprior to accepting squamous cell carcinoma as aprimary breast tumor. The usual precursor ofthis cancer is thought to be squamous meta-plasia, which occurs in a wide variety of set-tings including fibroadenoma, cystic lesions,phyllodes tumors, gynecomastia, mammaryduct hyperplasia, papillomatosis, subareolarabscesses, and areas of inflammation.3 In somecases squamous cell carcinoma may represent avariant of metaplastic carcinoma in which theadenocarcinomatous component has been over-grown by the squamous component.15

The mean age at diagnosis of patients withsquamous cell carcinoma of the breast is simi-lar to that seen with more common breast cancers.38,39 The lesions are usually palpable,and fixation to the chest wall as well as skininvolvement have been observed. Calcifica-tions may be seen on mammography.40 Grossly,the tumors frequently undergo cystic degenera-tion producing a cavity filled with necrotic

squamous debris. Microscopically, squamouscell carcinomas of the breast resemble similartumors arising in other sites. Keratin pearls andkeratohyaline granules may be present.3

As squamous cell carcinoma is a very rarelesion, information on prognosis and treatmentis limited. The majority of patients reported inthe literature have been treated by mastectomywith axillary dissection. Radiosensitivity ofthis tumor has not been defined.

SECRETORY CARCINOMA

Secretory carcinoma is a rare tumor affectingboth adults and children. In 1966, McDivitt andStewart described a series of seven youngpatients with this tumor, referring to it as juvenilecarcinoma.41 It soon became apparent, however,that most patients found to have this tumor werenot juveniles. Tavassoli and Norris, reporting ona series of 19 patients, found the median age atthe time of diagnosis to be 25 years, with sixpatients being > 30 years of age.42 As it becameobvious that the majority of patients with“juvenile carcinoma of the breast” were adults,it was redesignated “secretory carcinoma.”

Secretory carcinoma has been described inpatients from the first to the eighth decade oflife. Typically, these lesions are palpable andpresent as painless, well-circumscribed masses.Grossly, secretory carcinomas are white to grayor tan to yellow in color and may be lobulated.3

The margins are usually well-circumscribedand rarely infiltrative. Microscopically, thecells are filled with secretory material which ispale pink or amphophilic when stained withhematoxylin and eosin (Figure 6–4).3,41,42

Secretory carcinoma is considered a low-grade carcinoma with an excellent prognosis.Axillary metastasis has been identified inapproximately 20 percent of cases, but very fewpatients have been reported to have distantmetastasis.42–44 There is, however, a risk of latelocal recurrence.43,45 Wide local excision is pre-ferred in children, with an attempt to preservethe breast bud. In adults, breast conservation is


appropriate. Axillary lymphadenectomy shouldbe performed selectively based on physicalexamination or with the identification ofmetastases on sentinel node biopsy. Radiosen-sitivity of this tumor has not been defined, andthe majority of reported patients treated usingbreast conservation have not received postoper-ative radiation therapy.45

CARCINOMA OF THE BREAST WITHENDOCRINE DIFFERENTIATION

Rarely, tumors of the breast may undergoendocrine metaplasia and have the ability to pro-duce ectopic hormones such as human chorionicgonadotropin (HCG), calcitonin, adrenocorti-cotropin, and epinephrine. Endocrine differenti-ation may arise in the setting of ductal carcinomain situ, small-cell undifferentiated carcinoma,mucinous carcinoma, lobular carcinoma, andinfiltrating ductal carcinoma.3,46 Rarely, themicroscopic architecture of a breast cancer withendocrine differentiation may mimic the histo-logic structure of nonmammary tissue that con-tains the ectopic hormone being produced.

The clinical presentation of patients withcarcinomas of the breast with endocrine differ-entiation is similar to patients with more com-mon mammary neoplasms. Systemic symptomsattributable to the ectopic hormone producedare absent in nearly all cases. However, rarereports of systemic manifestations ascribed toectopic hormones do exist.47–49 Most of these

tumors are palpable.3 Grossly, there are no fea-tures specifically associated with endocrine dif-ferentiation. Microscopically, most carcinomasof the breast with endocrine differentiation con-tain argyrophilic cytoplasmic granules. Rarely,choriocarcinomatous differentiation may occur,resulting in tumors that are microscopicallysimilar to syncytiotrophoblast and cytotro-phoblast and are strongly reactive for HCG.3

There is general agreement that patientswith carcinomas of the breast with endocrinedifferentiation have a similar prognosis to thosewith like-staged more common mammary cancers.3,46 Treatment selection should be basedon conventional clinical and pathologic criteria.

PHYLLODES TUMOR

First characterized by Muller50 in 1838, phyl-lodes tumors are fibroepithelial neoplasmsaccounting for approximately 0.3 to 0.5 percentof breast tumors in women.51,52 This tumor’sother name, cystosarcoma phyllodes, is usedless often and considered by some as misleadingfor a lesion that is more often benign thanmalignant. These tumors may be locally aggres-sive but have minimal capability for metastasis.Phyllodes tumors have been the subject of manyreports, but their optimal management has yet tobe clearly defined.

Patients typically present with firm, discrete,mobile masses often clinically indistinguishablefrom fibroadenomas. Palpable axillary lymphnodes may be present in as many as 20 percentof patients but they are infrequently involved bytumor.3,53 The median age at presentation in themajority of published series is the fourth or fifthdecade, with a range of 10 to 86 years.3,53,54 It isimportant to note, however, that the mean age ofpresentation for patients with a fibroadenoma,approximately 30 years, is significantly lowerthan that for phyllodes tumors.55 The occurrenceof phyllodes tumors in patients < 30 years israre.3 Bernstein and colleagues identified race-specific differences in the incidence and meanage of diagnosis of patients in Los Angeles

104 BREAST CANCER

Figure 6–4. Secretory carcinoma (original magnification x400).

County with phyllodes tumors.56 The averageannual age-adjusted incidence rate for all racial-ethnic groups combined was 2.1 per 1 millionwomen in the population. Latina whites had thehighest incidence rate (2.8 per 1 million popula-tion) followed by non-Latina whites, Asians, andAfrican Americans, respectively.56 Bernstein andcolleagues found that the mean age of diagnosisfor non-Latina whites was 53.7 years, for Latinawhites 45.8 years, African Americans 48.7 years,and Asians 32.9 years. Clinically, tumors thatexhibit rapid growth, are > 4 cm, or previouslystable tumors that suddenly increase in size,should arouse suspicion. Mammographically,these lesions are smooth and lobulated. Withlocally invasive disease, margin irregularity maybe present. Ultrasound typically reveals a solidmass with no posterior shadowing. Cysts may bepresent within the lesion.

Grossly, phyllodes tumors are well circum-scribed and firm. On sectioning, the tumors aregray to tan and bulging (Figure 6–5). Focal cys-tic necrosis may be present. Microscopically,phyllodes tumors can be difficult to differenti-ate from benign cellular fibroadenomas. Typi-cally, an increased cellularity of the stromalcomponent is present. Long epithelial-linedclefts (intracanalicular pattern) are a prominentfeature and may help differentiate these tumorsfrom fibroadenomas (Figure 6–6). Mixoidchanges may be present. Some degree ofepithelial hyperplasia is common, with as manyas 10 percent of tumors containing squamousmetaplasia.57 Histologically, these tumors aredivided into three groups: benign, low-grademalignant (borderline), and high-grade malig-nant (Table 6–1). Benign phyllodes tumors arecharacterized by having < 1 mitosis per 10high-power fields (HPF). The stromal expan-sion is uniform throughout the lesion, and thecellularity is modest in extent with mild cellu-lar atypia. Low-grade malignant tumors typi-cally have microscopically invasive borders,moderate heterogeneously distributed stromalexpansion, and < 5 mitoses per 10 HPF. High-grade malignant phyllodes tumors have marked

hypercellular stromal overgrowth. Cellularpleomorphism is common, with typically > 5mitoses per 10 HPF.3 Although the proportionof patients with clinically enlarged axillarylymph nodes approaches 20 percent,3,54 the rateof pathologically confirmed axillary metastasisis well under 5 percent.53,58,59

The likelihood that a phyllodes tumor willmetastasize and/or locally recur depends on itshistologic classification. Histologically, benignlesions have a local recurrence rate of six to tenpercent.54,60,61 and minimal risk of systemicmetastasis.3,62 Low-grade malignant phyllodestumors locally recur in 25 percent to 32 percentof cases3,57 and have a reported incidence ofdistant metastasis of under 5 percent.3 Tumorswith a malignant histologic classification havea high rate of local recurrence and a 25 percentrisk of systemic metastasis.3,57,62


Figure 6–5. Cut surface of phyllodes tumor.

Figure 6–6. Phyllodes tumor (original magnification x100).

Primary therapy of phyllodes tumors isaimed at reducing the risk of local recurrence.These tumors must be excised to clear surgicalmargins. The magnitude of the negative mar-gin must be dictated by the histologic featuresof the tumor and the size of the breast. Exci-sion with negative margins up to 2 cm hasbeen suggested by some authors.3,15,53,54,62

Mastectomy may be indicated if the lesioncannot be completely excised in a cosmeti-cally acceptable wide local excision. Axillarydissection is not required. Clinically suspi-cious nodes are invariably hyperplastic andshould be individually biopsied. Locallyrecurrent phyllodes tumor does not mandatemastectomy. Complete excision to wide nega-tive margins is acceptable.15

The role of radiotherapy in the treatment ofpatients with phyllodes tumor remains unclear.There are multiple reports of insensitivity toradiation when used for palliation.63–65 Twoinvestigators, however, describe the use of post-operative whole breast irradiation followingbreast-preserving surgery for phyllodestumor.66,67 Local recurrence rates were notreported in these studies.

PRIMARY BREAST LYMPHOMA

Primary lymphoma of the breast is a rare dis-ease, accounting for < 1.0 percent of all breastmalignancies.68–70 The origin of this tumorremains unclear. Several investigators havesuggested that mucosa-associated lymphoid tis-sue (MALT) may play a role in its develop-ment.71–73 In 1972, Wiseman and Liao defined

the lesion and established the following crite-ria: (1) a close association between breast tissueand the infiltrating lymphoma, (2) no history ofextramammary lymphoma, and (3) the breastmust be the primary clinical site.

Patients with primary breast lymphoma typ-ically present with a palpable, sometimes ten-der, breast mass.74–76 Rapid growth of the tumoris common.75,76 Diffuse infiltration, skinchanges, and clinically palpable axillary nodeshave been described.75–77 Lymphoma “B” typesymptoms are rare.3,74–76,78 Bilateral involve-ment has been reported in up to 25 percent ofpatients.74,78,79 The majority of patients presentin their sixth decade of life, but a bimodal agedistribution with peaks in the mid-30s and mid-60s has been reported.3,74–76,78–80 Mammog-raphy and ultrasound of patients with primarybreast lymphoma demonstrates a solitary massin the majority of cases. The imaging charac-teristics of this lesion are nonpathognomonic.81

Grossly, these tumors have a gray-white cutsurface, are well circumscribed, fleshy, andmay be nodular.3 The majority of primarybreast lymphomas are classified as, mixed, orlarge cell with diffuse architecture and a B-cellphenotype.3,75,76,78,79 T-cell tumors are rare.74,78

In some cases, the linear arrangement of lym-phoma cells in the stroma may mimic invasivelobular carcinoma. Immunostains for epithelialand lymphoid markers may be required to dif-ferentiate between the two.3

Patients with primary breast lymphomamust be staged in a manner similar to otherlymphoma patients. Local excision followed by radiation therapy provides excellent localcontrol.70,75,82,83 Negative margins are notrequired. Most patients who fail therapy willrecur at distant sites or in the other breast.3 Asprimary therapy appears to have little impact onsurvival, radical surgery is rarely required inthe treatment of patients with primary breastlymphoma. Patients with stage I disease andthose with histologically low-grade tumorshave the most favorable prognosis.69 Systemictherapy should be considered in all cases.3

106 BREAST CANCER

Table 6–1. DIFFERENTIATION OF BENIGN,BORDERLINE, AND MALIGNANT PHYLLODES TUMORS

Benign Borderline Malignant

Borders Pushing Mostly pushing Infiltrating

Atypia Slight Moderate Marked

Mitoses/10HPF <1 1–4 ³ 5

Stromal overgrowth Rare Occasional Frequent

HPF = high-power field

BREAST SARCOMA

First described in 1828 by Chelius,84 breast sar-coma accounts for less than one percent of allbreast malignancies,85,86 with an annual inci-dence in the United States of approximately17.5 new cases per 1 million women.87 The rar-ity of this lesion has resulted in reports onbreast sarcoma typically addressing a heteroge-neous group of tumors, including malignantphyllodes tumors. Mammary sarcomas shouldbe limited to tumors arising from interlobularmesenchymal elements comprising the sup-porting stroma.3 These tumors include liposar-coma, leiomyosarcoma, osteogenic sarcoma,chondrosarcoma, malignant fibrous histiocy-toma, fibrosarcoma, rhabdomyosarcoma, pri-mary angiosarcoma, and hemangiopericytoma.Also included in this category, but discussedseparately in this review, are postradiotherapyangiosarcomas. Phyllodes tumors, which arisefrom intralobular and periductal stroma, shouldbe excluded.

Typically, breast sarcomas present as painless,mobile, well-circumscribed breast masses.3,15,88,89

There is commonly a history of rapid growth ina pre-existing mass.15,89 Skin involvement andnipple changes have been reported but are infre-quent.88,90 Enlarged axillary lymph nodes may bepalpable, but pathologically confirmed axillarymetastases are rare. The mean age at the time ofpresentation in three recent reports ranged from44 to 55 years (range 16 to 87 years).89–91 Mam-mographically, these lesions typically appear aswell-circumscribed, dense masses. Rarely, thepresence of osseous trabeculae within anosteogenic sarcoma may be noted.92

Grossly, the majority of breast sarcomas arewell circumscribed. The tumors usually grow asexpansile masses, compressing surrounding tis-sue as they enlarge. The margin of a liposar-coma may be multinodular and infiltrative. Thecut surface is typically yellow, gray or white incolor. There may be a whorled texture as well asareas of necrosis. Gelatinous areas are fre-quently noted in liposarcomas. Histologically,

sarcomas of the breast are similar to their morecommon counterparts occurring in other areasof the body (Figure 6–7). In the breast, however,metaplastic carcinoma must be excluded priorto establishing a diagnosis of mammary sar-coma.3 Diagnosis of this lesion requires exten-sive sampling to rule out the presence of in situor invasive carcinoma. Immunohistochemicalstudies for epithelial markers may be useful indifficult cases. Axillary lymph node involve-ment is exceedingly uncommon.3 Gutman andcolleagues identified axillary nodal metastasesonly in the context of disseminated disease.89

Breast sarcomas should be treated similarlyto sarcomas occurring elsewhere in the body.Surgery is the mainstay of treatment. Wide localexcision with histologically negative margins isrequired.15,89,93 Mastectomy may be necessaryto ensure complete excision of larger tumors.Neoadjuvant chemoradiation should be consid-ered in patients with large tumors. No staging ortherapeutic role for axillary lymphadenectomyhas been demonstrated.89 Axillary lymph nodedissection need be performed only if requiredfor complete excision of the tumor.

Gutman and colleagues, reporting on 60cases, found a median disease-free survival of17.7 months and median overall survival of 67months (median follow-up of 120 months).89

Local failure was reported in 19 patients. Thosesuffering local failure typically did so withinthe first 24 months. Patients with sarcomas < 5


Figure 6–7. High grade sarcoma (original magnification x100).

cm in size were found to have a significantlybetter prognosis.89 Pollard and colleagues iden-tified an overall 5-year mortality of 64 percentwith a local recurrence rate of 44 percent in the25 patients in their series.90 The role of adjuvantradiotherapy has yet to be defined.89

POSTRADIOTHERAPYANGIOSARCOMA

The occurrence of sarcoma following radiother-apy has been well described.94,95 The wide-spread acceptance of breast preservation in thetreatment of breast cancer may have an unfore-seen secondary result: an increase in the numberof patients at risk for developing post irradiationsarcoma. Postradiotherapy sarcoma was definedby Cahan and colleagues in 1948 as sarcomadeveloping in a previously irradiated field aftera latency period of several years.96 Angiosar-coma, osteosarcoma, malignant fibrous histio-cytoma, and fibrosarcoma have all beenreported in the irradiated breast. Angiosarcomahas been the topic of many recent reports.97–100

The rarity of these tumors makes the true inci-dence of postradiotherapy angiosarcoma of thebreast difficult to determine. The estimated riskof patients treated with whole breast irradiationranges from 0.06 to 0.4 percent.97,100–102 Strobbeand colleagues, reporting on 21 patients withpostradiotherapy angiosarcomas of the breastcollected from the Netherlands cancer registry,cite a potential incidence as high as 1.59 per-

cent when a median latency period of 74months is considered.97

The median interval between breast-preserv-ing therapy and the occurrence of angiosarcomaof the breast has been reported to be between 6and 11 years (range 2 to 44).95,97,98,100 Patientstypically present with skin changes reminiscentof a hematoma. These changes may be presentover a broad area within the radiation field. Anunderlying breast mass may be present. Red-dish-purple skin patches as well as vesicles havealso been reported.97 Findings on mammog-raphy, ultrasound, and MRI are nonspecific.103

Grossly, the tumors may be friable, firm, orspongy. Areas of cystic hemorrhagic necrosisare common in high-grade lesions.3 Histologi-cally, the postirradiated angiosarcoma of thechest wall primarily occurs in the skin, withoccasional extension to underlying subcuta-neous tissue or breast. It shows a wide spectrumof degree of differentiation but is commonlyhigh grade. Well-formed inter-anastomosingvascular channels corresponding to low-gradeangiosarcoma frequently merge with high-gradesolid or spindle-cell-containing regions (Figure6–8). Distinction from postirradiation benignvascular changes including vascular or lym-phatic ectasia and the so called “atypical vascu-lar lesions” in addition to hemangiomas ismandatory.104 Axillary lymph node involvementis uncommon. Gutman and colleagues, report-ing on 17 patients with angiosarcoma, identifiedaxillary nodal metastases only in the context ofdisseminated disease.89

Surgical treatment of postradiotherapy angio-sarcoma should consist of complete resectionwith wide negative margins. Salvage mastec-tomy with en-bloc resection of involved skinand adjacent structures is often required.97,105

Reconstruction of the resulting defect mayrequire musculocutaneus flaps and/or skingrafts. Rarely, patients with small lesions maybe adequately treated with a partial mastec-tomy.105 Axillary dissection need be performedonly if required for complete excision of thetumor.105,106

108 BREAST CANCER

Figure 6–8. Angiosarcoma (original magnification x100).

The prognosis for primary angiosarcoma ofthe breast appears to depend on the grade of thetumor.105 The prognosis for postradiotherapyangiosarcomas, however, remains unclear. Thenumber of reported patients with postradiother-apy angiosarcomas is insufficient to draw accu-rate conclusions regarding prognosis. High-grade lesions are aggressive and tend to recurlocally.107 Low-grade tumors appear to have amore favorable prognosis.3,98,108

REFERENCES

1. Haagensen CD. In: Diseases of the breast. 2nd ed.Philadelphia: W. B. Saunders; 1971.

2. Fisher ER, Palekar AS, Redmond C, et al. Patho-logic findings from the National SurgicalAdjuvant Breast Project (Protocol No. 4). VI.Invasive papillary cancer. Am J Clin Pathol1980;73:313–20.

3. Rosen PP. In: Rosen’s breast pathology. Philadel-phia: Lippincott-Raven Publishers; 1997.

4. World Health Organization. Histological typing ofbreast tumours. 2nd ed. International histolog-ical classification of tumours #2. Genova:World Health Organization; 1981.

5. Foote FW Jr, Stewart FW. A histologic classifica-tion of carcinoma of the breast. Surgery1946;19:74–99.

6. Carter D. Intraductal papillary tumors of thebreast. A study of 76 cases. Cancer 1977;39:1689–92.

7. Lefkowitz M, Lefkowitz W, Wargotz ES. Intra-ductal (intracystic) papillary carcinoma of thebreast and its variants: a clinicopathologicalstudy of 77 cases. Hum Pathol 1994;25:802–9.

8. Kraus FT, Neubecker RD. The differential diagno-sis of papillary tumors of the breast. Cancer1962;15:444–5.

9. Tsuda H, Uei Y, Fukutami T, Hirohashi S. Differ-ent incidence of loss of heterozygosity onchromosome 16q between intraductal papil-loma and intracystic papillary carcinoma of thebreast. Jpn J Cancer Res 1994;85:992–6.

10. Tsuda H, Takarabe T, Susumu N, et al. Detection ofnumerical and structural alterations and fusionsof chromosomes 16 and 1 in low-grade papil-lary breast carcinoma by fluorescence in situhybridization. Am J Pathol 1997;151:1027–34.

11. Tavassoli FA. Classification of metaplastic carci-noma of the breast. Pathol Annu 1992;27:89–119.

12. Rottino A, Wilson K. Osseous, cartilaginous andmixed tumors of the human breast: a review ofthe literature. Arch Surg 1945;50:184–93.

13. Kaufman MW, Marti JR, Gallager HS, Hoehn JL.Carcinoma of the breast with pseudosarcoma-tous metaplasia. Cancer 1984;53:1908–17.

14. Oberman HA. Metaplastic carcinoma of the breast:a clinicopathologic study of 29 patients. Am JSurg Pathol 1987;11:918–29.

15. Rosen P. Invasive mammary carcinoma. In: HarrisJR, Hellman S, Harris JR, et al., editors. Dis-ease of the breast. Philadelphis: Lippincott–Raven Publishers, 1996.

16. Rosen PP, Ernsberger D. Low-grade adenosqua-mous carcinoma a variant of metaplastic mam-mary carcinoma. Am J Surg Pathol 1987;11:351–8.

17. Bauer TW, Rostock RA, Eggleston JC, Baral E.Spindle cell carcinoma of the breast. HumPathol 1984;15:148–52.

18. Gersell DJ, Katzenstein AA. Spindle cell carcinomaof the breast. Hum Pathol 1981;12:550–60.

19. Huvos AG, Lucas JC, Foote FW Jr. Metaplasticbreast carcinoma. NY State J Med 1973;73:1078–81.

20. Mossler JA, Barton TK, Brinkhous AD, et al.Apocrine differentiation in human mammarycarcinoma. Cancer 1980;46:2463–71.

21. Azzopardi JG. Problems in breast pathology. Lon-don; W. B. Saunders; 1979.

22. Eusebi V, Betts C, Haagensen DE, et al. Apocrinedifferentiation in lobular carcinoma of thebreast: a morphologic, immunologic and ultra-structural study. Hum Pathol 1984;15:134–40.

23. Lee BJ, Pack GT, Scharnagel I. Sweat gland can-cer of the breast. Surg Gynecol Obstet 1933;54:975–96.

24. Abati AD, Kimmel M, Rosen PP. Apocrine mam-mary carcinoma a clinicopathologic study of72 cases. AJCP 1990;94:371–7.

25. d’Amore ESG, Terrier-Lacombe MJ, Travagli JP,et al. Invasive apocrine carcinoma of thebreast: a long term follow-up study of 34 cases.Breast Cancer Res Treat 1988;12:37–44.

26. O’Malley FP, Page DL, Nelson EH, Dupont WD.Ductal carcinoma in situ of the breast withapocrine cytology: definition of a borderlinecategory. Hum Pathol 1994;25:164–8.

27. Tavossoli FA, Norris HJ. Intraductal apocrine car-cinoma: a clinicopathologic study of 37 cases.Mod Pathol 1994;7:813–8.

28. Carter D, Rosen PP. Atypical apocrine metaplasiain sclerosing lesions of the breast: a study of 51patients. Mod Pathol 1991;4:1–5.


29. Cavanzo FJ, Taylor HB. Adenoid cystic carcinomaof the breast. Cancer 1969;24:740–5.

30. Geschickter CF. In: Diseases of the breast.Philadelphia: J. B. Lippincott; 1945.

31. Peters GN, Wolff M. Adenoid cystic carcinoma ofthe breast. Report of 11 new cases: review ofthe literature and discussion of biologicalbehavior. Cancer 1982;52:680–6.

32. Qizilbash AH, Patterson MC, Oliveira KF. Ade-noid cystic carcinoma of the breast. ArchPathol Lab Med 1977:101;302–6.

33. Rosen PP. Adenoid cystic carcinoma of the breast:a morphologically heterogeneous neoplasm.Pathol Annu 1989;24:237–54.

34. Ro JY, Silva EG, Callager HS. Adenoid cystic car-cinoma of the breast. Hum Pathol 1987;18:1276–81.

35. Orenstein JM, Dardick I, van Nostrand AW. Ultra-structural similarities of adenoid cystic carci-noma and pleomorphic adenoma. Histopathol-ogy 1985;9:623–38.

36. Herzberg AJ, Bossen EH, Walther PJ. Adenoidcystic carcinoma of the breast metastatic to thekidney. A clinically symptomatic lesion requir-ing surgical management. Cancer 1991;68:1015–20.

37. Leeming R, Jenkins M, Mendelsohn G. Adenoidcystic carcinoma of the breast. Arch Surg1992;127:233–5.

38. Rostock RA, Bauer TW, Eggleston JC. Primarysquamous carcinoma of the breast: a review.Breast 1984;10:27–31.

39. Shousha S, James AH, Ferandez MD, Bull TB.Squamous cell carcinoma of the breast. ArchPathol Lab Med 1984;108:893–6.

40. Tashjian J, Kuni CC, Bohn LE. Primary squamouscell carcinoma of the breast; mammographicfindings. J Can Assoc Radiol 1989;40:228–9.

41. McDivitt, RW, Stewart FW. Breast carcinoma inchildren. JAMA 1966;195:388–90.

42. Tavassoli FA, Norris HJ. Secretory carcinoma ofthe breast. Cancer 1980;45:2404–13.

43. Akhtar M, Robinson C, Ali MA, Godwin JT.Secretory carcinoma of the breast in adults.Light and electron microscopic study of threecases with review of the literature. Cancer1983;51:2245–54.

44. Rosen PP, Cranor ML. Secretory carcinoma of thebreast. Arch Pathol Lab Med 1991;115:141–4.

45. Krausz T, Jenkins D, Grontoft O, et al. Secretorycarcinoma of the breast in adults: emphasis onlate recurrence and metastasis. Histopathology1989;14:25–36.

46. Maluf HM, Koerner FC. Carcinomas of the breast

with endocrine differentiation: a review. Vir-chows Arch 1994;425:449–57.

47. Coombes RC, Easty GC, Detre SI, et al. Secretionof immunoreactive calcitonin by human breastcarcinomas. Br Med J 1975;4:197–9.

48. Woodard BH, Eisenbarth G, Wallace NR, et al.Adrenocorticotropin production by a mam-mary carcinoma. Cancer 1981;47:1823–7.

49. Kaneko H, Hojo H, Ishikowa S, et al. Norepi-nephrine-producing tumors of bilateral breasts.A case report. Cancer 1978;41:2002–7.

50. Muller J. Uber den fein eran Bau and die Formander Krakhaften Geschwilste. Berlin: G. Reiver;1838.

51. Oberman HA. Cystosarcoma phyllodes. A clini-copathologic study of hypercellular periductalstromal neoplasms of the breast. Cancer1965;18:697–710.

52. Rowell MD, Perry RR, Hsiu JG, Barranco SC.Phyllodes tumors. Am J Surg 1993;165:376–9.

53. Reinfuss M, Mitus J, Krzysztof D, et al. The treat-ment and prognosis of patients with phyllodestumor of the breast an analysis of 170 cases.Cancer 1996;77:910–16.

54. Rajan PB, Cranor ML, Rosen PP. Cystosarcomaphyllodes in adolescent girls and youngwomen: a study of 45 patients. Am J SurgPathol 1998;22:64–9.

55. Foster ME, Garrahan N, Williams S. Fibroade-noma of the breast: a clinical and pathologicalstudy. J R Coll Surg Edin 1988;33:16–19.

56. Bernstein L, Deapen D, Ross RK. The descriptiveepidemiology of malignant cystosarcoma phyl-lodes tumors of the breast. Cancer 1993;71:3020–4.

57. Grimes MM. Cystosarcoma phyllodes of the breast:histologic features, flow cytometry analysis andclinical correlations. Mod Pathol 1992;5:232–9.

58. Palmer ML, De Risi DC, Pelikan A, et al. Treat-ment options and recurrence potential for cys-tosarcoma phyllodes. Surg Gynecol Obstet1990;170:193–6.

59. Norris HJ, Taylor HB. Relationship of histologicfeatures to behavior of cystosarcoma phyl-lodes: analysis of ninety-four case. Cancer1967;20:2090–9.

60. Ciatto S, Bonardi R, Cataliotti L, Cardona G andmembers of the Coordinating Center and Writ-ing Committee of FONCAM. Phyllodes tumorof the breast: a multicenter series of 59 cases.Eur J Surg Oncol 1992;18:545–9.

61. Zurrida S, Bartoli C, Galimberti V, et al. Whichtherapy for unexpected phyllode tumour of thebreast? Eur J Cancer 1992;28:654–7.

110 BREAST CANCER

62. Hines JR, Marad TM, Beal JM. Prognostic indica-tors in cystosarcoma phyllodes. Am J Surg1987;153:276–80.

63. Sheen-Chen S, Chou F, Chen W. Cystosarcomaphyllodes of the breast: a review of clinical,pathological and therapeutic options in 18cases. Int Surg 1991;76:101–4.

64. Al-Jurf A, Hawk Wa, Crile G. Cystosarcoma phyl-lodes. Surg Gynecol Obstet 1978;146:358–64

65. Blichert-Toft M, Hansen JPH, Hansen OH,Schiodt T. Clinical course of cystosarcomaphyllodes related to histologic appearance.Surg Gynecol Obstet 1975;140:929–31.

66. Cohn-Cedermark G, Rutqvist L, Rosendahl I, Sil-fversward C. Prognostic factors in cystosar-coma phyllodes—a clinicopathologic study of77 patients. Cancer 1991;68:2017–22.

67. Hopkins ML, McGowan TS, Rawlings G, et al.Phyllodes tumor of the breast: a report of 14cases. J Surg Oncol 1994;56:108–12.

68. Fischer MG, Chideckel NG. “Primary” lym-phoma of the breast. Breast 1984;10:7–9.

69. Giardini R, Piccolo C, Rilke F. Primary non-Hodgkin’s lymphomas of the female breast.Cancer 1992;69:725–35.

70. Anania G, Baccarani U, Risaliti A, et al. Primarynon-Hodgkin’s T-cell lymphoma of the breast.Eur J Surg 1997;163:633–5.

71. Mattia AR, Ferry JA, Harris NL. Breast lymphoma.A B-cell spectrum including the low grade B-cell lymphoma of mucosa associated lymphoidtissue. Am J Surg Pathol 1993;17:574–87.

72. Isaacson PG. Lymphomas of mucosa-associatedlymphoid tissue (MALT). Histopathology1990;16:617–19.

73. Lamovec J, Jancar J. Primary malignant lymphomaof the breast. Lymphoma of the mucosa-associ-ated lymphoid tissue. Cancer 1987;60:3033–41.

74. Arber DA, Simpson JF, Weiss LM, Rappaport H.Non-Hodgkin’s lymphoma involving thebreast. Am J Surg Pathol 1994;18:288–95.

75. El-Ghazaway IMH, Singletary SE. Surgical man-agement of primary lymphoma of the breast.Ann Surg 1991;214:724–6.

76. Misra A, Kapur BML, Rath GK. Primary breastlymphoma. J Surg Oncol 1991;47:265–70.

77. Hugh JC, Jackson FI, Hanson J, Poppema S. Pri-mary breast lymphoma. An immunohistologicstudy of 20 new cases. Cancer 1990;66:2602–17.

78. Jeon HJ, Akagi T, Hoshida Y, et al. Primary non-Hodgkin malignant lymphoma of the breast.An immunohistochemical study of sevenpatients with breast lymphoma in Japan. Can-cer 1992;70:2451–9.

79. Wiseman C, Liao KT. Primary lymphoma of thebreast. Cancer 1972;29:1705–12.

80. Adair FE, Hermann JB. Primary lymphosarcomaof the breast. Surgery 1944;16:836–53.

81. Liberman L, Giess CS, Dershaw DD, et al. Non-Hodgkin lymphoma of the breast: imagingcharacteristics and correlation with histopatho-logic findings. Radiology 1994;192:157–60.

82. DeBlasio D, McCormick B, Straus D, et al. Defin-itive irradiation for localized non-Hodgkin’slymphoma of the breast. Int J Radiat OncolBiol Phys 1989;17:843–6.

83. Smith MR, Brustein S, Straus DJ. Localized non-Hodgkin’s lymphoma of the breast. Cancer1987;59:351–4.

84. Chelius MJ. Teleangiektasie. Heidelberger KlinAnn 1828;499–517.

85. Oberman HA. Sarcomas of the breast. Cancer1965;10:1233–43.

86. Callery CD, Rosen PP, Kinne DW. Sarcomas ofthe breast. A study of 32 patients with reap-praisal of classification and therapy. Ann Surg1985;201:527–32.

87. May DS, Stroup NE. The incidence of sarcomasof the breast among women in the U.S. 1973–1986. Plast Reconstruct Surg 1991;87:193–4.

88. Moore MP, Kinne DW. Breast sarcoma. Surg ClinNorth Am 1996;76:383–92.

89. Gutman H, Pollock RE, Ross MI, et al. Sarcomaof the breast: implications for extent of therapythe M.D. Anderson experience. Surgery 1994;116:505–9.

90. Pollard SG, Marks PV, Temple LN, ThompsonHH. Breast sarcoma: a clinicopathologicreview of 25 cases. Cancer 1990;66:941–4.

91. Terrier MJ, Terrier-Lacombe H, Mourisees S, etal. Primary breast sarcoma: a review of 33cases with immunohistochemistry and prog-nostic factors. Breast Cancer Res Treat 1989;13:39–48.

92. Elson BC, Ikeda DM, Andersson I, Wattsgard C.Fibrosarcoma of the breast: mammographicfindings in five cases. AJR 1992;158:993–5.

93. Mies C. Mammary sarcoma and lymphoma. In:Bland KI, Copeland EM, editors. The breast.Philadelphia: W. B. Saunders; 1998.

94. Laskin WB, Silverman TA, Enzinger FM. Postra-diation soft tissue sarcomas: an analysis of 53cases. Cancer 1988;62:2330–40.

95. Brady MS, Gaynor JJ, Brennan MF. Radiation-associated sarcoma of bone and soft tissue.Arch Surg 1992;127:1379–85.

96. Cahan WG, Woodard HW, Higinbotham NL, et al.Sarcoma arising in irradiated bone: report of11 cases. Cancer 1948;1:3–29.


97. Strobbe LJA, Peterse HL, van Tinteren H, et al.Angiosarcoma of the breast after conservationtherapy for invasive cancer, the incidence andoutcome. An unforeseen sequela. Breast Can-cer Res Treat 1998;47:101–9.

98. Bolin DJ, Lukas G. Low-grade dermal angiosar-coma of the breast following radiotherapy. AmSurg 1996;62:668–71.

99. Del Mastro L, Garrone O, Guenzi M, et al.Angiosarcoma of the residual breast after con-servative surgery and radiotherapy for primarycarcinoma. Ann Oncol 1994;5:163–5.

100. Slotman BJ, van Hattum AH, Meyer S, et al.Angiosarcoma of the breast following conserv-ing treatment for breast cancer. Eur J Cancer1994;30:416–7.

101. Wijnmaalen A, van Ooijen B, van Geel BN, et al.Angiosarcoma of the breast following lumpec-tomy, axillary lymph node dissection, andradiotherapy for primary breast cancer: threecase reports and a review of the literature. Int JRadiat Oncol Biol Phys 1993;26:135–9.

102. Zuccali R, Merson M, Placucci M, et al. Soft tis-sue sarcoma of the breast after conservative

surgery and irradiation for early mammarycancer. Radiother Oncol 1994;30:271–3.

103. Schnarkowski P, Kessler M, Arnholdt H, Helm-berger T. Angiosarcoma of the breast: mammo-graphic, sonographic, and pathological find-ings. Euro J Radiol 1997;24:54–6.

104. Fineberg S, Rosen PP. Cutaneous angiosarcomaand atypical vascular lesions of the skin andbreast after radiation therapy for breast carci-noma. Am J Clin Path 1994;102:757–63.

105. Rosen PP, Kimel M, Ernsberger D. Mammaryangiosarcoma. The prognostic significance oftumor differentiation. Cancer 1988;62:2145–51.

106. Jardines L. Other cancers in the breast. In: HarrisJR, Hellman S, Harris JR, editors. Diseases ofthe breast. Philadelphia: Lippincott-RavenPublishers; 1996.

107. Naka N, Ohsawa M, Tomita Y, et al. Prognosticfactors in angiosarcoma: a multivariate analy-sis of 55 cases. J Surg Oncol 1996;61:170–6.

108. Donnell RM, Rosen PP, Lieberman PH, et al.Angiosarcoma and other vascular tumors of thebreast. Pathologic analysis as a guide to prog-nosis. Am J Surg Pathol 1981;5:629–42.

112 BREAST CANCER

113

OVERVIEW

Breast cancer is a highly prevalent and morbiddisease, afflicting approximately 1 in 9 womenin the United States. The death rate from breastcancer in the United States has recentlydeclined for most age groups, although itremains a major killer with 45,000 deaths annu-ally.1 Despite the overall decline, the incidenceof ductal carcinoma in situ (DCIS) and stage Idisease has risen significantly. In parallel, ther-apeutic opportunities (both traditional andalternative) for breast cancer patients haverapidly expanded. Surgical, radiologic, andmedical oncologic modalities are increasinglydiverse, including both therapeutic and preven-tive strategies. In this era of diversity, individu-alization of treatment strategies to maximizeresponse and minimize morbidity and mortalityhas become the goal.2,3

As documented in the literature of the1930s, breast masses identified at surgery andpresumed to be cancer could be immediatelyremoved by mastectomy on the basis of physi-cal findings.4 The evolution of biopsy followedby histopathology as the diagnostic test preced-ing definitive surgery began in the second quar-ter of the 20th century5–7 and became firmlyestablished within decades. Histologic criteriabecame further defined with experience, result-ing in subclassification of preneoplastic and

malignant processes.8 Needle biopsies (fine-needle aspiration [FNA] and core-needlebiopsy) have contributed to this process,although their role in screening, diagnosis, andsubclassification is variable and still in evolu-tion.9–13 Marker studies represent another phasein this evolution.

Randomized clinical trials (RCTs) have pro-vided important data that allow comparison ofchemotherapeutic strategies for adjuvant, neoad-juvant, palliative, or preventive treatments. Tis-sues acquired from patients enrolled on theserandomized trials have become an importantresource for correlative studies, allowing analy-sis of tumor markers for prognosis, prediction oftherapeutic benefit, or molecular epidemiologicstudies. Knowing which option to choose, inwhat order, and in which combination is thechallenge that drives prognostic and predictivebreast cancer marker studies.

This explosion in breast cancer care optionshas occurred nearly simultaneously with themolecular revolution that occurred in the laterdecades of the 20th century. The identificationand characterization of deoxyribonucleic acid(DNA), the introduction of monoclonal anti-bodies, and molecular biologic methods havegiven scientists and physicians new tools tostudy old problems. As a result, our under-standing of breast carcinogenesis and cancerbiology has been greatly modified. We now

7

Prognostic and Predictive Markers in Breast CancerANN D. THOR, MDDAN H. MOORE II, PHD

appreciate the immense genetic heterogeneityof the disease, both within and between indi-vidual patients. Numerous genetic, transcrip-tional, and protein alterations may someday beused to diagnose and subclassify breast can-cers, augmenting and perhaps replacinghistopathology as the gold standard (Figure7–1). Until the value of these new markers isdetermined by careful study, however, use ofclassic markers in breast cancer remains criticalto the practice of breast cancer care. It is likelythat, over the next decade, emphasis will beplaced on predictive markers and quantitationof molecular targets to guide novel therapeutics(see Figure 7–1).

Clinical Use of Usual Breast CancerMarkers in the 1990s

Breast cancer markers can be broadly subdi-vided into (1) clinical or histologic “markers”or characteristics (such as tumor size, nodalmetastases), which are useful to define or sub-divide the disease; and (2) markers that areidentifiable by specialized testing of the cancer,sera, nipple aspirate, or other biologic sample.Both types of markers have clinical rele-vance14–22 and utility (Table 7–1).

Many physicians currently use breast can-cer markers only for prognosis. Pathologists

regularly perform careful histopathologicanalyses as recommended by the Associationof Directors of Anatomic and Surgical Pathol-ogy.23 The clinical application of markers, suchas proliferative rate and oncogenes, has beenmore controversial. The American Society ofClinical Oncology (ASCO) and other specialtygroups have been reviewing these issues andare expected to publish updated recommenda-tions. Markers for therapeutic prediction,including erbB-2 (HER-2/neu), represent thenew frontier.

Identification of New Markers

Potential new breast cancer markers are gener-ally evaluated using at least one of three types ofstudies: (1) early exploratory studies, which gen-erally seek associations between markers anddisease characteristics; (2) studies to determinewhether factors provide improved means ofidentifying patients at high or low risk for dis-ease progression or death (using various statisti-cal methods, see below); and (3) studies todetermine if markers predict benefit from agiven therapeutic regimen.24 In general, analy-ses seek to determine if a marker is specific forthe disease or tissue type, whether it relates toother disease characteristics of interest, or if ithas prognostic (the ability to portend outcomeindependent of therapy) or predictive (the abilityto portend outcome dependent on therapy) value.

Widely accepted methodologic principleshave been published to guide the design, conduct,and analysis of clinical trials.16,20,21 However, fewguidelines have been published for clinical test-ing of prognostic or predictive markers. Generalguidelines, which are not marker specific, havebeen reported and are summarized in Table 7–2.For specific breast cancer markers, relevance andapplicability may change with time. Recommen-dations published by governing bodies or profes-sional organizations (such as National CancerInstitute [NCI], Food and Drug Administration[FDA], American Cancer Society [ACS], ASCO,College of American Pathologists [CAP],

114 BREAST CANCER

Cancer–associated markers (1960s-)

Breast cancer–associated markers (1970s-)

Prognostic markers associated with breast cancer (1980s-)

Predictive markers associated with breast cancer (1990s-)

? Molecular targets for therapeutic intervention (2000-)

Figure 7–1. Evolution of cancer markers.

American College of Surgeons) are usuallyupdated every few years. Referral to their rec-ommendations on a regular basis is advisable.

To exemplify the evolving nature of thefield, the NCI convened a consensus conferencein 1985. Only nodal status and tumor size wererecognized. In 1990, five factors were recom-mended by a similar group of experts: nodalmetastases, tumor size, histologic grade, histo-logic subtype, and steroid receptor status.3

Additional markers, including proliferation rate,ploidy, and oncogenes (such as p53 mutation/overexpression and erbB-2 amplification/over-expression), were recognized as promisingmarkers. At that time, predictive markers (asso-ciated with a treatment response) were not rec-ognized except for steroid receptors.

Lymph Node Metastases

Nodal metastases are a well-recognized risk fac-tor for poor outcome in breast cancer patients(discussed in detail elsewhere).23,25,26 Nodal pos-itivity and the number of nodes with metastasesare associated with an increased risk of diseaserecurrence or progression.27 Nodal metastaseshave often been divided into subgroups for prog-nostic purposes or randomized trial entry. Whilesubclassification can be afforded using this tech-nique, the biologic value of nodal metastasesshould be considered as a continuum that corre-lates with outcomes data on survival and recur-rence.27 Lymph node metastases are often thestrongest independent variable (marker) of out-comes in breast cancer patients.

There are several important issues related tolymph node metastases that have not yet beenclarified. These include the biologic signifi-cance (and definition) of microscopic nodalmetastases, nodal metastases detected by onlyimmunohistochemical assay (and the need toperform such assays), and issues related to lim-ited axillary dissections (sentinel node proce-dures). There is no consensus regarding the useof frozen sections or cytology techniques onsentinel lymph nodes, the optimal protocol for

sentinel node processing (step sections,immunohistochemistry), or the clinical rele-vance of micro- or single-cell metastases. Res-olution of these issues will better define therole of the pathologist in this procedure. It willalso determine whether microscopic cellularmetastases is itself a marker of prognosis.

Tumor Size

Tumor size is an independent prognosticmarker that is particularly important in node-negative breast cancer patients.17 Most actuar-ial survival data have emphasized 1 cm andlarger tumors, broken into subgroups, forprognostication. Given the increased inci-dence of tumors ² 1 cm, subclassification ofthese smaller tumors by size is important aswell.27–33 Subsetting this group into ² 0.5 cmversus 0.5 to 1 cm has been used by some. Fur-ther subdivisions of this group are likely asmore data are available. For these small, node-negative tumors (T1a), proliferation rateappears to be a very important prognosticmarker as well.34,35

Prognostic and Predictive Markers in Breast Cancer 115

Table 7–1. UTILITY OF BREAST CANCER MARKERS

• Risk assessment• Early detection• Cancer subclassification• Differential diagnosis• Prognosis• Therapeutic prediction• Disease monitoring

Table 7–2. GENERAL GUIDELINES FOR CLINICAL MARKER UTILIZATION

• Markers should exhibit significant and independent predictive value, validated by clinical testing (ie, they shouldnot be implemented solely on the basis of retrospectivedata analysis).3

• Assays used should be feasible, reproducible, widelyavailable, and subject to quality control.3

• Marker analysis should provide data that are readily inter-pretable by the clinician, with therapeutic implications.3

• The measurement of a factor should not consume tumorspecimen needed for other tests, particularly careful cytologic and/or histologic analysis.20

Tumor Grade

Several schemes have been proposed for gradingbreast carcinomas. These generally includearchitectural cellular arrangement, nuclear fea-tures, and other items such as mitotic rate. TheSurveillance, Epidemiology, and End Results(SEER) data from thousands of patients haveshown that breast tumor grading, irrespective ofthe scheme used, has prognostic signifi-cance.36,37 The 1990 Consensus Conference rec-ommendation—that nuclear grade be evalu-ated—has now been largely superseded byincreasing consensus that a single classificationscheme should be adopted. The Elston scheme,which includes nuclear and architectural featuresand mitotic count, is increasingly used.38,39 Con-sensus on a preferred grading scheme is likelyto be forthcoming.40–44 Nuclear grading is alsopossible on cytology preparations from touchimprint or fine-needle aspiration. Athough notequivalent to the combined architectural/cyto-logic systems used in surgical pathology, itmay provide important data on nucleargrade.9–11 Clearly, patients with low-grade(better-differentiated) breast cancers have abetter prognosis than those with high-grade(poorly differentiated) carcinoma. The readeris referred to Chapters 5 and 6 for a detaileddiscussion on breast pathology.

Steroid Receptor Analysis

Steroid receptors have been routinely deter-mined on surgically resected primary breastcancers since the late 1970s.45 Receptor-containing tumors have a better short-termprognosis, although the magnitude of this dif-ference is relatively small (8 to 10 percent dif-ference in recurrence rate for node-negativepatients at 5 years).17,46 Long-term relapse andsurvival rates between receptor-positive andreceptor-negative tumor patients, however, tendto merge.17,44,47 Despite this, steroid receptorassays are often used as a marker of probablesensitivity to tamoxifen or other agents thatbind the estrogen receptor.

Determination of steroid receptors on cyto-logic or surgical preparations of primary breasttumors is standard and almost universally per-formed. Such tests should also be obtained onpresurgical breast cancer samples (cytology orcore biopsy) if neoadjuvant chemo- or radiother-apy will be given. Receptor determination mayalso be performed on tumor metastases if thesteroid receptor status was not determined on theprimary tumor or if there is reason to suspectbiologic cancer progression (the development ofan estrogen receptor–negative phenotype).

Immunochemical assays for estrogen andprogesterone receptors (ER, PgR) are mostoften used.17 Pathologists evaluate the receptorstatus of the invasive component only. Gener-ally, immunopositivity of benign breast epithe-lium adjacent to the cancer is sought as aninternal positive control for the assay. Manypathologists have developed their own defini-tion of positivity, which is used in reporting.This may include evaluation of the percentageof cells staining as well as consideration ofstain intensity. An effort should be made tohave the local institutional scoring systemdefined and the methodology clearly stated inthe assay report. Methods and scoring differ-ences have been estimated to contribute to dis-agreements between laboratories in up to 30percent of specimens. Data on ER from archivalcases often used different methodologies forER or PgR, and cut-off levels for positivitywere generally determined by the local labora-tory. Immunohistochemical methods now inuse for ER and PgR can be applied to archivalfixed-embedded tissue specimens that aredecades old, should concern arise about oldhormone receptor data.

The estrogen receptor is a good example of amarker that is both prognostic and predictive.Patients with tumors that are ER positive aremore likely to have a better outcome indepen-dent of treatment. These same patients are alsomore likely to respond to tamoxifen therapy (apositive predictive factor).17 Adding to the com-plexity of the issue, recent data suggest that

116 BREAST CANCER

optimal ER assay cut points may be different tooptimize either prognostic or predictive estima-tions.48 In summary, ER and PgR have both pos-itive prognostic and predictive values associatedwith a more favorable patient outcome.

Cancer Subtyping

Tumor subtyping has been recognized to haveindependent prognostic significance in breastcancer.22 Ten to 30 percent of invasive ductalcarcinomas are of a special type, many ofwhich can be recognized on cytology prepara-tions. Three of these were recognized by the1990 National Institutes of Health (NIH) Con-sensus Conference as having a favorable prog-nosis—the tubular, colloid, and papillary vari-ants.3 Each of these three patterns havecytologic correlates and are often classifiableon FNA (refer to Chapters 5 and 6). With therapid expansion of molecular technologies,including the promising array-based formats(which may be used simultaneously to measurehundreds to thousands of genes or proteinsfrom a breast cancer), subtyping based on geneexpression will soon be possible.

Proliferation Rate

Cellular proliferation is an important biologiccharacteristic of cancer but has been lesswidely accepted as an independent prognosticmarker. Part of the reticence to adopt it as aroutinely reported marker may be due to thewide variety of tests for quantitation. The pro-liferation rate is based on the principle of cellreplication (cell cycle states G1, S, M, G2).Cells can also be in a resting phase, known asG0. The mitotic rate, which is generally scoredfrom routine hematoxylin and eosin–stainedslides of primary tumor, has been quantitatedusing several systems, including mitotic fig-ures per 10 high-powered fields, mitotic fig-ures per 1,000 cells, or as a percentage ofinvasive cancer cells. The strengths and weak-nesses of these visual counts have been

reviewed elsewhere.18 Mitotic counts are nowa composite of a commonly used grading sys-tem.39,40 Separate reporting of the mitotic score,in addition to the Elston grade, is supported bya recent multivariate analysis of outcomes,which included both statistical models.34 Whenother systems of mitotic quantitation arereported, the mitotic rate should be comparedwith the mean or median for other similarbreast cancers at the same institution, usingthe same scoring methodology. In general,invasive lobular carcinomas have a signifi-cantly lower mitotic rate than infiltrating duc-tal carcinomas.34 Scoring of in situ carcinomashas not yet been associated with prognostic orpredictive value. Although mitotic countingwas first reported nearly a century ago, itremains an important prognostic marker ofbreast cancer biology.34,49

Other techniques that estimate the percent-age of cells in the S phase include flow cytom-etry and thymidine (or thymidine analogue)uptake and immunohistochemical detection ofproliferation associated antigens, such as Ki-67 or proliferating cell nuclear antigen(PCNA)/cyclin.18,49 Of these, flow cytometryor the immunohistochemical detection of pro-liferation associated antigens (such as Ki-67)are the most commonly used. When flowcytometry is performed on a small sample,macrodissection to increase the tumor/benignratio is advisable, as a false diploid readingmay result.

Summary of Commonly ReportedBreast Cancer Markers

Clearly, no single agent or combination treat-ment is appropriate for all patients.44 Improve-ments in outcome, with accurate forecasting ofwho will derive the greatest benefit, are impor-tant. “Therapeutic modeling” using markers hasbeen evaluated in detail by a multidisciplinarypanel sponsored by the ASCO. Three steps weresuggested for the clinical integration of prognos-tic data: (1) analysis of a given patient’s risk of


recurrence and survival based on historic out-come and multiple prognostic factors; (2) identi-fication of various treatment options and theirpotential therapeutic benefit and risk; and (3) anoverall assessment of the expected benefit, risks,cost, and other personal factors that might influ-ence treatment decisions and outcome.19 Deter-mination of prognostic factors in breast cancers,with subsequent use of those data to make ther-apeutic decisions and predict outcome, is com-plicated but feasible. The goal of using markersshould be to “contribute to a decision in practicethat results in a more favorable clinical outcomefor the patient.”19 Similar guidelines should beapplicable to predictive factor analysis.

Evolving Markers Based on Cancer Biology

In the early days of marker development, it wasassumed that we could subset patients for coun-seling and treatment on the basis of marker andclinical data that predicted outcome. Unfortu-nately, clinical and tumor biology heterogeneityamong breast cancer patients made an exactprediction of outcome for an individual patientmore difficult than anticipated. While prognos-tic studies can provide an estimate of risk,translation to a single patient is inherently morecomplex. Recently reported treatment/markerinteractions have made prognostic marker stud-ies even more difficult. The use of archivaltumor banks, comprising heterogeneouslytreated patients is no longer acceptable for ver-ification studies of prognostic markers.Increasingly, such studies are performed onpatient samples derived from cooperativegroup-based randomized trials. This designallows testing for treatment/marker interac-tions.50,51 Marker/therapy interactions may beconfounding. While the relationship may besimilar (a poor prognosis and marker of poorresponse to outcome), it is not always so. Amarker may be associated with a negative prog-nostic value and a positive predictive value, orvice versa. Given the complexity of such inter-

actions, rapid progression of prognostic mark-ers from the research arena to translational(clinical) applications may be ill advised.

While centralized banking has met resis-tance from some local institutions and patholo-gists, nationally applied standards and safe-guards may eventually make these a safer placethan local archives for long-term storage ofslides or blocks. Given the evolving technolo-gies, banks of tumor DNA may someday becommonplace as well. The emergence of pre-dictive factors and a new format for marker val-idation (the RCT) has greatly affected the labo-ratory development and analysis of newmarkers. Marker/chemotherapy interactionsmay also explain, to some extent, discordantmarker data on distinct patient subsets, obtainedusing retrospective archival tumor tissues.50,51

Historic Overview: Identification ofNovel Cancer Markers

The first widely studied cancer-associatedmarker was carcinoembryonic antigen (CEA).It was detectable in tumors as well as body flu-ids and generated great excitement as a markerfor the diagnosis or monitoring of cancer. It isexpressed by adenocarcinomas (includingbreast cancers); however, benign epitheliumand inflammatory states produce CEA as well.The challenges in CEA research were makingreagents and assays that were specific forCEA, quantitating the protein in human tissuesand body fluids, determining the associationswith clinical and disease parameters, and bet-ter definition of CEA biology and structure.We know now that CEA is a member of a largefamily of proteins with homology to othercross-reacting antigens.52–55 Many of the earlystudies used nonspecific reagents and generatedconflicting results. The CEA has shown lim-ited usefulness in breast cancer immunodiag-nostics and as a marker of disease progression.Because it is not part of a critical molecularpathway, targeting of CEA for therapeuticintent has not been useful.

118 BREAST CANCER

In the early 1980s, scientists used breastcancer cells, cell lines, or derived cellular prod-ucts (eg, membrane extracts) to generate mono-clonal antibodies against breast cancer associ-ated antigens. Numerous reagents werediscovered, such as the human milk fat globule(HMFG) membranes.56,57 Expression patternsof these antigens were variable. None havedemonstrated independent value as prognosticor predictive tumor markers, although some ofthese reagents have been used to monitor dis-ease progression.

The development of molecular methods,with subsequent studies of critical cell surfacereceptors, oncogenes, and tumor suppressorgenes, has significantly altered the emphasis ofbreast cancer marker studies away from anti-gens identified by chance. In the early 1980s,gene sequence data were first used to generatemonoclonal antibody probes against peptides,such as the mutant ras gene.58 This allowedvisualization and quantitation of ras gene alter-ations in situ in human colon and breast adeno-carcinomas.57–60 Rapid expansion of these tech-nologies to other important breast cancermarkers, including the clinical acceptance ofimmunohistochemical assays to analyze estro-gen and progesterone receptors, firmly estab-lished this approach.

Some have suggested a “growing backlashof negative sentiment concerning breast cancerprognostic factors in the oncology communitytoday.”61 However, with a shift in emphasisfrom prognosis to prediction and an increaseduse of targeted molecular therapeutics, thefield of markers in breast cancer care hassolidified. The ability of scientists to insert(transfect) genes into cells and the develop-ment of knock-in and knock-out transgenicmice now allows hypothesis testing to deter-mine the specific biologic role(s) of specificgenes or signal transduction pathways. Thesetools have revolutionized our understanding ofcancer biology and, in the process, have identi-fied entirely new targets (markers) for molecu-lar therapeutics.

Promising Prognostic and Predictive Markers

A number of biology-associated markers withreported prognostic or predictive value havebeen reported (Table 7–3). While many ofthese are of biologic interest, relatively fewwill likely have independent prognostic value.With or without prognostic value, genes ortheir encoded products may be useful as thera-peutic targets.

Growth Factors and Receptors

Breast epithelial cells are, by necessity, respon-sive to a wide range of growth factors and theirreceptors including hormones and their cognatereceptors, ER and PgR, prolactin, insulin, and avariety of other factors. Estrogen and ER pro-mote cancer development through an indirectprocess that includes the promotion of cellgrowth and the activation of estrogen-respon-sive genes. Some growth factors/receptors areconsidered oncogenes as well because in theaberrant state, they may cause cancer. Membersof the type I growth factor receptor family (epi-dermal growth factor receptor [EGFR], erbB-2[HER-2/neu], erbB-3, erbB-4) have prognostic,predictive, and therapeutic target value. For thepurpose of this chapter, only the highlights ofthe voluminous literature will be cited.

erbB-2

Over a decade has passed since the HER-2/neu(erbB-2) gene was first identified in chemically


Table 7–3. IMPORTANT MARKER GROUPS IN BREAST CANCER

• Oncogenes• Tumor suppressor genes• Programmed cell death associated• Angiogenesis associated• Growth factors and their receptors• Adhesion molecules • Proteases/Protease inhibitors• Metastasis associated

induced glial tumors in rats.62,63 The humanequivalent of the proto-oncogenic neu, knownas erbB-2 is located on chromosome 17.64 TheerbB-2 gene encodes a transmembrane protein,p185, with structural homology to EGFR. Thisstructural homology is one of the features link-ing these two genes as members of the type 1receptor tyrosine kinase (RTK) gene superfam-ily, which also contains the less well studiedmembers erbB-3 and erbB-4.65,66 Althoughencoded by individual genes, these membersare highly homologous. Each possesses anextracellular ligand-binding domain, and li-gands that bind to all but erbB-2 have beenidentified. These four members can formhomo- and heterodimers, with 10 possibledimers. The biologic differences in a prognosticor predictive sense are not yet known for thesedifferent configurations.

The erbB-2 overexpression/amplification is acomplex process,67–69 which occurs in approxi-mately one-third of invasive and up to two-thirdsof in situ carcinomas.70–77 The erbB-2 alterationshave been associated with a poor prognosis inbreast cancer patients,17,50,78–89 although it is usu-ally less predictive of outcome than lymph nodemetastases. A resurgence of interest in erbB-2 asa breast cancer marker has recently occurredbecause erbB-2 alterations may predict chemore-sponsiveness81,83,85 and the FDA has recentlyapproved the drug Herceptin® (Trastuzumab,Genentech, Inc.), which targets erbB-2.

Cancers without erbB-2 alterations have twocopies of the gene (unless deletions haveoccurred) and encode low levels of protein. Allnormal cells and the majority of breast cancercells bear two copies of the erbB-2 gene and pro-duce low levels of the encoded protein p185.Assays to evaluate erbB-2 generally measureeither gene copy number or protein expression.Abnormal erbB-2 can be defined as proteinexpression at levels above normal cells or genecopy number > 2. Assays, therefore, need to beprecise and have the discriminatory power toseparate abnormal from normal. Variance inassay procedures or reagents may increase or

decrease the sensitivity or specificity of the test(no matter what procedure is used), resulting infalse negatives or false positives. For this reason,calibration of erbB-2 assays, using controls withvarious levels of gene amplification, are neces-sary and can be purchased commercially. Theseshould be fixed embedded pellets of cell lines,with and without gene amplification, rather thanhuman tumors that have been positive before.

“Kits” that support erbB-2 testing fromreagents to recommended scoring systems havejust been released. It is possible that these sys-tems will be superior to the currently used “in-house” technologies, although there is littledata yet to support that conclusion.

While the pathologist may use a specialscoring system (such as the 0, 1+, 2+, or 3+system for the Dako HercepTest®), providingan estimate of the percentage of erbB-2–posi-tive invasive cancer cells will allow greatercomparison with other laboratories. Whilereagent issues are beyond the scope of thischapter, the methodology for scoring deservesbrief mention. In general, (1) membranousreactivity only should be considered positive;(2) the invasive component of a tumor only (notin situ disease) should be scored; (3) erbB-2staining should not be observed in adjacentstroma or inflammatory cells, nor should benignepithelium show strong membranous reactivity;(4) reporting should include an approximateestimate of the percentage of immunopositiveinvasive cancer cells; (5) positive and negativecontrols should be included in each assay; and(6) the method and primary reagent used by thelaboratory should be reported with the assayresult. While some recommend a reporting/scoring of staining intensity, few have com-pared that data with outcome. There is littledata that intensity, by itself, has prognostic orpredictive value.81 Cells with concentric mem-branous staining only are recommended forscoring by some (Dako HercepTest®). This hasnot been proven to be superior to focal mem-brane staining for prognostic or predictive pur-poses. However, in general, intensity and con-

120 BREAST CANCER

centric staining are associated with higher lev-els of gene amplification. Discrete cut pointsused for data analysis in some studies are sub-optimal; the biologic relevance of erbB-2 likelyrepresents a continuum.

Two commercial fluorescence in situhybridization (FISH) assays for erbB-2 haverecently been approved by the FDA for progno-sis (not qualification of patients for Herceptin).The majority of studies have shown compara-bility between immunohistochemical data andFISH,77 although some have reported that FISHmethods are superior.84 The FISH methodologyis generally similar to immunohistochemistry,although reagents are more expensive, requir-ing special microscopic equipment and greaterpathologist time for scoring. Differencesbetween kits include probe labeling (direct viaindirect) and the use of a centromeric probe forchromosome 17 in addition to the erbB-2probe. Intratumor heterogeneity of erbB-2 genecopy number and chromosome 17 centromericcopy number are common.79 Scoring systemsthat reflect this heterogeneity have not beenwidely applied and, therefore, the biologic rele-vance is unknown.

erbB-2 as a Predictive Marker in Breast Cancers.The erbB-2 data from nearly 1,000 stage IIbreast cancers derived from a randomized three-arm trial (Cancer and Leukemia Group B,CALGB trial 8541) of cytoxan, adriamycin, and5-fluorouracil (CAF) have suggested interac-tions between erbB-2 and chemotherapy (doseof CAF81,83,85). This conclusion is supported byboth molecular and immunohistochemicalerbB-2 data. Patients whose tumors had ampli-fied or overexpressed erbB-2, treated with doseintensive CAF, had a significantly better sur-vival than patients without erbB-2 abnormali-ties assigned to the same treatment arm.81,83 Inthis study, stage II breast cancer patients whosetumors had alterations of both erbB-2 and p53treated with dose-intensive CAF had the mostfavorable outcome (90% 10-year survival).81

Interactions between erbB-2 and response to

CAF have now been reported by others aswell.85,86 Data from older cooperative group tri-als suggest a relative resistance of erbB-2–altered breast cancers to methotrexate-basedregimens.50,87,88 Taxol resistance has also beenassociated with erbB-2 overexpression/ amplifi-cation, although this issue remains controver-sial.80 Some reports have also suggested resis-tance of patients with ER+ tumor to tamoxifenif erbB-2 and/or EGFR are overexpressed,89–93

although the interaction has not been demon-strated by all.94–96

erbB-2 as a Therapeutic Target. In 1998 the FDAannounced approval of Herceptin® (Trastuzumab,Genentech, Inc.) for the treatment of metastaticbreast cancer. This approval occurred in fivemonths, a nearly unparalleled “fast-track.” Her-ceptin is thought to offer a less toxic approachfor treating breast cancer, as it directly targetserbB-2 associated growth promotion.97 Her-ceptin is a genetically engineered (humanized)monoclonal antibody which binds erbB-2.Early studies of breast cancer patients withadvanced disease have shown that as a singleagent, or in combination with other chemo-therapy, Herceptin significantly improved out-come for some patients.98,99 It is somewhatunclear how patients should be selected fortreatment with this agent. Most believe thatbreast cancers without erbB-2 alterations willnot be responsive to Herceptin, although therehas not been a clinical trial to test this hypothe-sis. In completed trials, patients with the greaternumber of cells with concentric erbB-2immunostaining had a greater response rate.

Epidermal Growth Factor Receptor

The EGFR gene is amplified with overexpres-sion or is overexpressed in many breast can-cers. This receptor allows breast cancer cells tobind a variety of autocrine or paracrine growthfactors (including epidermal growth factor[EGF], transforming growth factor-alpha[TGF-a]).100–103 The EGFR is upregulated by


estrogens via direct binding to the promotorregion of the gene.104–106 It is also constitutivelyactivated by amplified erbB-2.107 Binding ofligands such as EGF to EGFR triggers rapidtyrosine phosphorylation of the erbB-2 pro-tein108 as well as other downstream sub-strates.109,110

The EGFR is overexpressed by over one-third of infiltrating ductal carcinomas, is uni-versally expressed by medullary carcinomas,and is generally not detected in lobular or col-loid carcinomas.111 Overexpression of EGFRhas been reported in male breast cancers,although infrequently.112 Overexpression ofEGFR has been associated with increasedmetastatic potential and a worse prognosis inboth node-positive and node-negative breastcancer patients.113–118 The EGFR may interactwith erbB-2 to confer relative resistance totamoxifen in ER-positive patients,89–93,115,116

although, as stated above, this issue is some-what controversial. Co-overexpression oferbB-2 and EGFR is seen in approximatelyone-fifth of breast cancers. In summary, EGFRshould be considered prognostic and possiblypredictive on the basis of the data that are cur-rently available.

p53

Nearly one-third of breast cancers have muta-tions of the tumor suppressor gene p53. This hasbeen associated with histologic and clinicalaggressiveness.86,119–123 Mutations often result inoverexpression of the encoded protein as a resultof a prolonged half-life and protein accumula-tion. Fortunately, this effect allows immunohis-tochemical detection of p53 as a surrogate formutational analyses.81,120,124–128 This should beconsidered a screening method, as some muta-tions are clearly not detected. The p53 gene as abreast cancer marker appears more prognostic innode-negative as compared with node-positivebreast cancer patients. In addition to prognosticvalue, p53 data may help identify patients likelyto respond to chemo- or radiotherapy.129–131

The p53 gene is relatively large, and muta-tions have been reported in both introns andexons. In breast cancers, mutations appear tocluster in exons 5 to 9. Given the molecular com-plexity of this large gene, studies of mutationsbased on genetic sequencing have been limited.Newer technologies are being developed forsequencing in high-throughput formats. Giventhe size of the gene and its many functions, thelocation and type of genetic abnormality may beimportant to determine its clinical value.

Patients with germline p53 mutations(LiFraumeni syndrome) have an increased inci-dence of breast cancers.132 Recent evidencesuggests a relationship between BRCA-1 andp53 in hereditary breast cancer such that p53acts as a cancer cofactor in these patients.133

Most p53 abnormalities identified in breastcancers, however, occur as spontaneous,somatic events. The p53 abnormalities havebeen reported in invasive and in situ carcino-mas as well as in rare precursor lesions.

STATISTICAL ISSUES IN CANCER MARKER STUDIES

There are many statistical tools for studying therelationships between patient characteristics(such as age at diagnosis and genetic makeup),tumor parameters (eg, tumor size and grade),adjuvant therapy (primarily radiation andchemotherapy), markers, and length of sur-vival. This subsection will highlight key issuesthat are important in marker analyses.

Randomized Clinical Trials

An RCT is the most rigorous way to evaluatetreatment efficacy, compare different treat-ments, or test the predictive value of a givenmarker. The FDA requires proof of efficacyfrom one or more RCTs for approval of anynew treatment for cancer. The key to an RCT isblinded randomization of patients into treat-ment arms. Randomization minimizes the like-lihood that differences in outcome between two

122 BREAST CANCER

treatment groups are due to factors other thanthe difference in treatment. Formal statisticalmethods for ensuring that patients are assignedat random to a treatment group must be applied.The importance of this step cannot be overem-phasized. Promising results from small studieswithout randomization are often discovered tobe due to baseline differences between thosewho received the new treatment and those whodid not, rather than the differences in treatment.Failures in randomization can result in statisti-cal nightmares in the interpretation of markerdata. The study by Thor34 is a case in point.

An RCT usually involves patients from mul-tiple treatment centers. Once such trials havebeen reviewed and approved, patients areassigned at random to treatment groups duringthe enrollment period. Enrollments cease whentrials are closed. Patients “on protocol” are thenfollowed up for a predetermined length of time,and their outcome is recorded. The usual out-comes of interest are disease recurrence (other-wise known as disease-free survival [DFS]),disease-specific death (DSS), and death fromother causes. Recurrence or death is oftenreferred to as “events” in statistical jargon.Depending on the eligibility criteria, manypatients will not have had a recurrence and/orwill still be alive at the end of the study so thatthe survival time for those patients is notknown. These patients are generally removedfrom analyses of outcomes, a process called“censoring.” The number of patients who arecensored or have events determines the statisti-cal power of a given study.

Kaplan-Meier Survival Curves

Survival experience is usually quantitated asthe length of time patients are followed up andwhether or not outcomes of interest, or events,occurred. When an event has occurred, thelength of time is recorded as length of follow-up to the event time. Time after the event is dis-regarded. The simplest way to summarize dataof this type is to plot them as Kaplan-Meier

(K-M) survival curves. A description of how tocalculate points for drawing this curve can befound in many medical-statistical books.134

Comprehensive statistical software can calcu-late these curves with precision. The K-Mcurve is nonparametric, that is, there are noparameters to be estimated to determine itsshape. It is highly flexible and can be used to fitany set of survival data consisting of time and acensoring indicator. The basic idea underlyingthe K-M curve is that it starts with 100 percentsurvival at time 0 (usually the left side of thecurve along the Y axis). A decrement is made ateach event, and the size of the decrement isequal to the number of patients who experiencethat event at that time, divided by the number ofpatients who were still alive and being followedimmediately prior to the event time. A littleknown but useful feature of the curve is that thenumber of patients at the end of the study canbe estimated by examining the horizontaldecrement at the last event. The numberremaining is equal to the reciprocal of thisdecrement. For example, if the survival curvefalls by 0.1 at the time of the last event, therewere, with high probability, 10 patients fol-lowed up for this length of time.134

Comparisons of Survival

The survival experience of any number ofgroups can be compared visually by plotting K-M survival curves for each group. There aremany statistical tests for comparing survivalcurves; the most widely used is the log-ranktest. It summarizes the survival experience oftwo (or more) groups by forming a 2 ´ 2 con-tingency table each time an event occurs. Thisleads to a series of contingency tables and dif-ferent weighting schemes for combining theresults. The log-rank test weights the chi-squarestatistics from these tables equally. Harringtonand Fleming proposed a weighting schemeunder greater control by the user. This allowscomparisons of survival curves at specific timepoints, such as early or late. These might be


appropriate to determine if the effect of a newtreatment or marker is most pronounced earlyon. Similarly, later events may be analyzed bychanging the Harrington-Fleming parameter.135

Proportional Hazard Statistical Models

The K-M survival curves can be used to studythe effect of a continuous variable (eg, age), butit is necessary to define one or more cut pointsto subdivide patients into strata. A K-M curvecan be calculated for each stratum, and thestrata can be compared using the log-rank orother tests. However, results of such compar-isons are usually highly dependent on the cutpoints; therefore, a researcher who finds con-flicting results may not know how to report hisfindings. Cut points are also commonly used tomaximize the p values, a practice which shouldbe discouraged.

The most widely used statistical tool forstudying the effects of continuous variables onsurvival or recurrence is the proportional haz-ards model. This method was formulated byCox;136 hence, it is also known as the Cox model.It is based on an assumption that the hazard(defined as the instantaneous risk of experienc-ing an event at any given time) for a patient witha risk factor, say age, at level x is proportional tothe hazard for another patient at level x' and thatthe ratio of these hazards is constant over the fol-low-up time. A simple equation describes theratio of these hazards for any values of x and x':

Ratio of hazards = exp[ _ (x – x')]

where _ is a parameter to be estimated from thedata. This model is readily extended to manyvariables.

Comprehensive statistical software pro-grams include routines for estimating the para-meters of the Cox proportional hazards modeland for testing their statistical significance.This method is widely used for studying therelation between multiple factors and survival.It is said to be semiparametric because even

though the shape of the survival function is notspecified, the relative hazards associated withdifferent factors require estimation of parame-ters. In using this model to assess the impor-tance of factors on survival, it is important totest the proportional hazards assumption. Forexample, when the number of positive nodes isused, the model assumes that the relative hazardof having 1 positive node compared with noneis the same as that of having 20 nodes com-pared with 19. Some software programs havethe capability of testing these assumptions sta-tistically. However, it is important for the userto understand the meaning of the assumptionsbehind the proportional hazards model and tomake adjustments when possible to make theassumptions more relevant. For example, thelogarithm of the number of positive nodes (with1 added before taking the logarithm to avoidtaking the log of 0) is often a more realistic wayto model the effects of positive nodes thanusing the actual number. A similar transforma-tion may be appropriate for tumor size. Thiskind of variable transformation is often appliedin marker analyses, although it is usuallydescribed in detail only in figure legends or thestatistical methods section.

Univariate and Multivariate Analyses

When establishing the usefulness of a breastcancer marker, it is important to perform bothunivariate and multivariate analyses. Univariateanalysis determines whether the factor predictsthe end point. It does not consider the influenceof other factors and, therefore, can be mislead-ing. In evaluating novel markers, it is oftenunknown whether the marker is a cause or aresult of the cancerous process. Thus, associa-tions between the presence or absence of amarker with a better or worse survival does littleto advance our understanding of the underlyingbiology or improve predictability of outcome.Probability values (p-values) are poor indicatorsof relative statistical ranking of the importanceof multiple factors. For example, a factor that is

124 BREAST CANCER

“significant” at p = .001 may or may not be abetter predictor of outcome than one with a p =.02. Each factor may interact with others in waysthat are not assessed by the size of the p value.The only robust way to evaluate the performanceof different subsets of factors is to perform ran-domization tests of their performance.137

A much better understanding of the rela-tionship of the marker can be acquired when itis tested in the “presence” of other well-estab-lished factors (ie, multivariate analyses). Forexample, a factor that is highly correlated withthe number of positive nodes may provide noadditional independent prognostic/predictiveinformation when the latter factor is added tothe predictive equation. The only way to findthis out is to perform a multivariate analysis,where all factors are considered together in astatistical model for predicting outcome. TheCox multivariate model described above is awidely used and useful tool for determining thestatistical utility of a new factor in the presenceof established factors. When this analysis isperformed, it is important to include the mostpowerful usual markers in the model. Failure todo so may cause a new marker to appear impor-tant, when in reality it lacks independent value.

SUMMARY

It is important for clinicians and translationalscientists to work closely with statisticians forboth marker study design and analysis. Statisti-cians who regularly participate in breast cancermarker or outcome studies often have the great-est insights; they know which variables shouldbe considered in the analysis and what patientpopulations are best suited for hypothesis test-ing. Survival analysis requires “expert knowl-edge” as many variables can affect study out-comes.138 One of the most common errors that ismade in breast marker studies is the use ofdiverse patients with short-term follow-up. Well-established factors associated with survival, suchas tumor stage and patient age, must be consid-ered in trial design and statistical analyses.

In deciding how many patients should beincluded in a study, it is important to realizethat statistical power is dictated by the numberof events (ie, recurrences or deaths), not thenumber of subjects at the start of follow-up.When multiple factors are under study, a rule ofthumb is that 10 events are required for everyfactor studied. In node-negative disease where5-year survival is often above 90 percent, thismeans that 100 patients per factor under studywill be required. Prognostic marker studiesshould include both univariate and multivariateanalyses of outcomes. Prognostic marker stud-ies generally exclude consideration of treat-ment/marker interactions. If such interactionsexist, they will go unrecognized when the studypopulation is treated heterogeneously. To iden-tify such interactions, study of tumors derivedfrom patients entered in RCTs is necessary.

REFERENCES

1. Parker SL, Tong T, Bolden S, Wingo PA. Cancerstatistics, 1996. CA Cancer J Clin 1996;46(1):5–27.

2. Ernster VL, Barclay J, Kerlikowske K, et al. Inci-dence of and treatment for ductal carcinoma insitu of the breast. JAMA 1996;275(12):913–8.

3. NIH Consensus Conference. Treatment of earlystage breast cancer. JAMA 1991;265:391–5.

4. Tod MC, Dawson EK. The diagnosis and treat-ment of doubtful mammary tumours. Lancet1934;II:1041–5.

5. Muir R. The pathogenesis of Paget’s disease of thenipple and associated lesions. Br J Surg 1934;22:728–37.

6. Foote FW Jr, Stewart F. Lobular carcinoma in situ.A rare form of mammary cancer. Am J Pathol1941;7:491–6.

7. Ewing J. Epithelial and other tumours of thebreast. Neoplastic disease. Philadelphia: W. B.Saunders; 1940.

8. Page DL, Anderson TJ. How should we categorizebreast cancer? Breast 1993;2:217–9.

9. Kreuzer G, Boquoi E. Aspiration biopsy cytology,mammography and clinical exploration: amodern set up in diagnosis of tumors of thebreast. Acta Cytol 1976;20:319–23.

10. Kline TS, Joshi LP, Neal HS. Fine-needle aspirationof the breast: diagnoses and pitfalls. A review of3545 cases. Cancer 1979;44:1458–64.


11. Dabbs DJ. Role of nuclear grading of breast car-cinomas in fine needle aspiration specimens.Acta Cytol 1993;37(3):361–6.

12. Abiati A, Consensus Committee. The uniformapproach to breast fine needle aspiration biopsy:a synopsis. Acta Cytol 1996;40:1120–6.

13. Bassett L, Winchester DP, Caplan RB, et al.Stereotatic core-needle biopsy of the breast: areport of the Joint Task Force of the AmericanCollege of Radiology, American College ofSurgeons, and College of American Patholo-gists. CA Cancer J Clin 1997;47(3):171–90.

14. Aziz K. Tumour markers: current status and futureapplications. Scand J Clin Lab Invest 1995;Suppl 221:153–5.

15. Miller WR, Ellis IO, Sainsbury JR, Dixon JM.ABC of breast diseases. Prognostic factors.BMJ 1994;309:1573–6.

16. Levine MN, Browman GP, Gent M, et al. When isa prognostic factor useful? A guide for the per-plexed. J Clin Oncol 1991;9:348–56.

17. Clark GM. Prognostic and predictive factors. In:Harris JR, Hellman S, Lippman M, Morrow M,editors. Diseases of the breast. Philadelphia: J.B. Lippincott-Raven; 1996.

18. Thor AD, Edgerton SM. Cellular markers of pro-liferation and oncogenes. In: Colvin RB, BhanAD, McCluskey RT, editors. Diagnosticimmunopathology. 2nd ed. New York: RavenPress, Ltd; 1995.

19. Hayes DF, Bast RC, Desch CE, et al. Tumormarker utility grading system: a framework toevaluate clinical utility of tumor markers. JNatl Cancer Inst 1996;88(20):1456–66.

20. Gasparini G, Pozza F, Harris AL. Evaluating thepotential usefulness of new prognostic and pre-dictive indicators in node-negative breast can-cer patients. J Natl Cancer Inst 1993;85(15):1206–19.

21. McGuire WL. Breast cancer prognostic factors:evaluation guidelines. J Natl Cancer Inst 1991;83(3):154–5.

22. Page DL. Prognosis and breast cancer. Recogni-tion of lethal and favorable prognostic sub-types. Am J Surg Pathol 1991;15(4):334–49.

23. Connolly JL, Fechner RE, Kempson RL, et al.Recommendations for the reporting of breastcarcinoma. Association of Directors ofAnatomic and Surgical Pathology. Hum Pathol1996;27(3):220–4.

24. Simon R, Altman DG. Statistical aspects of prog-nostic factor studies in oncology. Br J Cancer1994;69(6):979–85.

25. Carter CL, Allen C, Henson DE. Relation of

tumor size, lymph node status and survival in24,740 breast cancer cases. Cancer 1989;63:181–7.

26. Cazin JL, Gosselin P, Boniface B, et al. Compara-tive values of several tumour markers: exampleof untreated breast carcinoma. Br J Cancer1990;62(6):1031–3.

27. Harris JR, Morrow M, Norton L. Malignanttumors of the breast. In: DeVita VT, HellmanS, Rosenberg SA, editors. Cancer: principlesand practice of oncology. 5th ed. New York:Lippincott-Raven Publishers; 1997.

28. Mansour EG, Ravdin PM, Dressler L. Prognosticfactors in early breast carcinoma. Cancer1994;74:381–400.

29. Hermanek P, Sovin LH, Fleming ID. What do weneed beyond TNM? Cancer 1996;77(5):815–7.

30. Goldhirsch A, Wood WC, Senn HJ, et al. Interna-tional Consensus Panel on the Treatment ofPrimary Breast Cancer. Eur J Cancer 1995;31A:1754–9.

31. Bast RC, Desch CE, Hayes DF, et al. Update ofrecommendations for the use of tumor markersin breast and colorectal cancer. J Clin Oncol1997;16:793–5.

32. Nemoto T, Vana J, Bedwani RN, et al. Managementand survival of female breast cancer: results of anational survey by the American College ofSurgeons. Cancer 1980;45(12):2917–24.

33. Koscielny S, Tubiana M, Le MG, et al. Breast can-cer: relationship between the size of the pri-mary tumour and the probability of metastaticdissemination. Br J Cancer 1984;49(6):709–15.

34. Thor AD, Liu S, Moore DH, Edgerton SM. Com-parison of mitotic index, in vitro bromo-deoxyuridine labeling, and MIB-1 assays toquantitate proliferation in breast cancers. J ClinOncol 1999;17(2):470–7.

35. Rosen PP, Groshen S, Saigo PE, et al. A long-termfollow up study of survival in stage I(T1NoMo) and stage II (T,N, M0) breast carci-noma. J Clin Oncol 1989;7:355–66.

36. Henson DE. The histologic grading of neoplasms.Arch Pathol Lab Med 1988;112:1091–6.

37. Henson DE, Ries L, Freedman LS, Carriaga M.Relationship among outcome, stage of disease,and histologic grade for 22,616 cases of breastcancer. The basis for a prognostic index. Can-cer 1991;68:2142–9.

38. Elston CW, Ellis IO. Method for grading breastcancer. J Clin Pathol 1993;46:189–90.

39. Elston CW, Ellis IO. Pathological prognostic fac-tors in breast cancer: I. The value of histologi-cal grade in breast cancer: experience from a

126 BREAST CANCER

large study with long-term follow-up. Histo-pathology 1991;19:403–10.

40. Robinson IA, McKee G, Kissin MW. Typing andgrading breast carcinoma on fine-needle aspi-ration: is this clinically useful information?Diagn Cytopathol 1995;13(3):260–5.

41. Cajulis RS, Hessel RG, Hwang S, et al. Simpli-fied nuclear grading of fine-needle aspirates ofbreast carcinoma: concordance with corre-sponding histologic nuclear grading and flowcytometric data. Diagn Cytopathol 1994;11(2):124–30.

42. Howell LP, Gandour-Edwards R, O’Sullivan D.Application of the Scarff-Bloom-Richardsontumor grading system to fine-needle aspiratesof the breast. Am J Clin Pathol 1994;101(3):262–5.

43. Dabbs DJ, Silverman JF. Prognostic factors fromthe fine-needle aspirate: breast carcinomanuclear grade. Diagn Cytopathol 1994;10(3):203–8.

44. Hirshaut Y, Pressman P. Breast cancer: the com-plete guide. New York: Bantam; 1996.

45. Jordan VC. Tamoxifen: a guide for clinicians andpatients. Huntington NY: PRR Publishers; 1996.

46. Clark GM, McGuire WL. Steroid receptors andother prognostic factors in primary breast can-cer. Semin Oncol 1988;15:20.

47. Early Breast Cancer Trialists’ CollaborativeGroup. I. Systemic treatment of early breastcancer by hormonal, cytotoxic, or immune ther-apy: 133 randomized trials involving 31,000recurrences and 24,000 deaths among 75,000women. Lancet 1992;339:1–15, 71–85.

48. Elias JM, Masood S. Estrogen receptor assay: arewe all doing it the same way? A survey. J His-totechnol 1995;18:95–6.

49. Thor AD, Yandell DW. Molecular pathology of thebreast. In: Harris, JR, Hellman S, Lippman M,Morrow M, editors. Diseases of the breast.Philadelphia: J. B. Lippincott-Raven; 1996.

50. Ravdin PM, Chamness GC. The c-erbB-2 proto-oncogene as a prognostic and predictivemarker in breast cancer: a paradigm for thedevelopment of other macromolecular mark-ers—a review. Gene 1995;159(1):19–27.

51. Berry DA, Thor A, Cirrincione C, et al. Scientificinference and predictions: multiplicities andconvincing stories: a case study in breast can-cer therapy. In: Bernardo JM, Berger JO,Dawid AP, et al., editors. Bayesian statistics.Vol 5. Oxford University Press; 1996.

52. Nap M, Klaski A, Hoor R, Fleuren G-J. Cross-reactivity with normal antigens in commercial

anti-CEA sera, used for immunohistology. Theneed for tissue controls and absorptions. Am JClin Pathol 1983;79:25–31.

53. Nap M, Keuning H, Burtin P, et al. CEA and NCAin benign and malignant breast tumors. Am JClin Pathol 1984;82:526–34.

54. Thompson J, Zimmermann W. The carcinoembry-onic antigen gene family: structure, expressionand evolution [review]. Tumor Biol 1988;9:63–83.

55. Thor A, Muraro R, Gorstein F, et al. Adjunct to thediagnostic distinction between adenocarcino-mas of the ovary and the colon utilizing amonoclonal antibody (COL 4) with restrictedcarcinoembryonic antigen reactivity. CancerRes 1987;47:505–12.

56. Foster CS, Edwards PAW, Dinsdale EA, NevilleAM. Monoclonal antibodies to the human mam-mary gland. Virchows Arch 1982;394:279–93.

57. Burchell J, Gendler S, Taylor-Papadimitriou J, etal. Development and characterization of breastcancer reactive monoclonal antibodies directedto the core protein of the human milk mucin.Cancer Res 1987;47:5476–82.

58. Thor A, Horan Hand P, Wunderlich D, et al.Monoclonal antibodies define differential rasgene expression in malignant and benigncolonic diseases. Nature 1984;311:562–5.

59. Horan Hand P, Thor A, Wunderlich D, et al.Monoclonal antibodies of predefined speci-ficity detect activated ras gene expression inhuman mammary and colon carcinomas. ProcNatl Acad Sci U S A 1984;81:5227–31.

60. Thor A, Ohuchi N, Horan Hand P, et al. ras genealterations and enhanced levels of ras p21expression in a spectrum of benign and malig-nant human mammary tissues. Lab Invest1986;55(6):603–15.

61. Osborne CK. Prognostic factors for breast cancer:have they met their promise? J Clin Oncol1992;10(5):679–82.

62. Bargmann CI, Hung MC, Weinberg RA. Multipleindependent activations of the neu oncogene bya point mutation altering the transmembranedomain of p185. Cell 1986;45(5):649–57.

63. Schecter AL, Stern DF, Vaidyanathan L, et al. Theneu oncogene: an erbB related gene encoding a185,000 Mr tumour antigen. Nature 1984;312:513–6.

64. Coussens L, Yang-Feng TL, Liao YC, et al. Tyro-sine kinase receptor with extensive homologyto EGF receptor shares chromosomal locationwith neu oncogene. Science 1985;230:1132–9.

65. Kraus MH, Issuing T, Miki N, et al. Isolation andcharacterization of erbB-3, a third member of


the erbB/epidermal growth factor receptorfamily: evidence for overexpression in a set ofhuman mammary tumors. Proc Natl Acad SciU S A 1989;86(23):9193–7.

66. Plowman GD, Culouscou J-M, Whitney GS, et al.Ligand-specific activation of HER4/p180erbB-4, a fourth member of the epidermalgrowth factor receptor family. Proc Natl AcadSci U S A 1993;90:1746–50.

67. Nandi S, Guzman RC, Yang J. Hormones andmammary carcinogenesis in mice, rats, andhumans: a unifying hypothesis. Proc Natl AcadSci U S A 1995;92(9):3650–7.

68. Ethier SP. Growth factor synthesis and humanbreast cancer progression. J Natl Cancer Inst1995;87(13):964–73.

69. Stern DF. Biology of ErbB2/HER2/Neu. [Sub-mitted]

70. King CR, Kraus MH, Aaronson SA. Amplificationof a novel v-erb-related gene in a human mam-mary carcinoma. Science 1985;229:974–6.

71. Kraus MH, Popescu NC, Amsbaugh SC, KingCR. Overexpression of the EGF receptor-related proto-oncogene erbB-2 in human mam-mary tumor cell lines by different molecularmechanisms. EMBO J 1987;6:605–10.

72. Slamon DJ, Clark GM, Wong S, et al. Humanbreast cancer: correlation of relapse and sur-vival with amplification of the HER-2/neuoncogene. Science 1987;235:177–82.

73. Slamon DJ, Godolphin W, Jones LA, et al. Studies ofthe HER-2/neu proto-oncogene in human breastand ovarian cancer. Science 1989;244:707–12.

74. Thor AD, Schwartz LH, Koerner FC, et al. Analy-sis of c-erbB-2 expression in breast carcinomaswith clinical follow-up. Cancer Res 1989;49:7147–52.

75. Paik S, Hazan R, Fisher ER, et al. Pathologic find-ings from the National Surgical Adjuvant Breastand Bowel Project: prognostic significance oferbB-2 protein overexpression in primary breastcancer. J Clin Oncol 1990;8:103–12.

76. Liu E, Thor A, He M, et al. The HER2 (c-erbB-2)oncogene is frequently amplified in in situ carci-nomas of the breast. Oncogene 1992;7:1027–32.

77. Kallioniemi OP, Kallioniemi A, Kurisu W, et al.erbB-2 amplification in breast cancer analyzedby fluorescence in situ hybridization. Proc NatlAcad Sci U S A 1992;89(12):5321–5.

78. Dittadi R, Catozzi L, Gion M, et al. Comparisonbetween Western blotting, immunohistochemi-cal and ELISA assay for p158neu quantitationin breast cancer specimens. Anticancer Res1993;13(5C):1821–4.

79. Szollosi J, Balazs M, Feuerstein BG, et al. erbB-2(HER2/neu) gene copy number, p185HER-2overexpression and intratumor heterogeneity inhuman breast cancer. Cancer Res 1995;55(22):5400–7.

80. Press MF, Bernstein L, Thomas PA, et al. HER-2/neu gene amplification characterized by flu-orescence in situ hybridization: poor prognosisin node-negative breast carcinomas. J ClinOncol 1997;15(8):2894–904.

81. Thor AD, Berry DA, Budman DR, et al. erbB-2,p53 and efficacy of adjuvant therapy in lymphnode-positive breast cancer. J Natl Cancer Inst1998;90:1346–60.

82. Thor AD, Budman DR, Berry DA, et al. Selectingpatients for higher dose adjuvant CAF: c-erbB-2,p53, dose and dose intensity in stage II, nodepositive breast cancer [abstract]. Proc ASCO1997;16:128a.

83. Muss HB, Thor AD, Berry DA, et al. C-erbB-2expression and response to adjuvant therapy inwomen with node-positive early breast cancer.N Engl J Med 1994;330:1260–6.

84. Pauletti G, Godolphin W, Press MF, Slamon DJ.Detection and quantitation of HER-2/neu geneamplification in human breast cancer archivalmaterial using fluorescence in situ hybridiza-tion. Oncogene 1996;13(1):63–72.

85. Paik S, Bryant J, Park C, et al. erbB-2 andresponse to doxorubicin in patients with axil-lary lymph node-positive, hormone receptor-negative breast cancer. J Natl Cancer Inst1998;90(18):1361–70.

86. Clark GM. Should selection of adjuvantchemotherapy for patients with breast cancerbe based on erbB-2 status? J Natl Cancer Inst1998;90(18):1320–1.

87. Gusterson BA, Gelber RD, Goldhirsch A, et al.Prognostic importance of c-erbB-2 expression inbreast cancer. J Clin Oncol 1992;10(7):1049–56.

88. Allred DC, Clark GM, Tandon AK, et al. HER-2/neu in node-negative breast cancer: prognos-tic significance of overexpression influencedby the presence of in situ carcinoma. J ClinOncol 1992;10(4):599–605.

89. Leitzel K, Teramoto Y, Konrad K, et al. Elevatedserum c-erbB-2 antigen levels and decreasedresponse to hormone therapy of breast cancer.J Clin Oncology 1995;13(5):1129–35.

90. Borg A, Baldetorp B, Ferno M, et al. erbB-2amplification is associated with tamoxifenresistance in steroid-receptor positive breastcancer. Cancer Lett 1994;81:137–44.

128 BREAST CANCER

91. Benz CC, Scott GK, Sarup JC, et al. Estrogen-dependent, tamoxifen-resistant tumorigenicgrowth of MCF-7 cells transfected with HER2/neu. Breast Cancer Res Treat 1993;24(2):85–95.

92. Wright C, Nicholson S, Angus B, et al. Relation-ship between c-erbB-2 protein product expres-sion and response to endocrine therapy inadvanced breast cancer. Br J Cancer 1992;65(1):118–21.

93. Carlomagno C, Perrone F, Gallo C, et al. c-erbB-2 overexpression decreases the benefit of adju-vant tamoxifen in early-stage breast cancerwithout axillary lymph node metastases. J ClinOncol 1996;14(10):2702–8.

94. Archer SG, Eliopoulos A, Spandidos D, et al.Expression of ras p21, p53 and c-erbB-2 inadvanced breast cancer and response to firstline hormonal therapy. Br J Cancer 1995;72(5):1259–66.

95. Loaiciga K, Beckmann MW, Niederacher D, etal. p185/HER2 and pS2 protein overexpressionin breast cancer specimens: improvement forprediction of response to endocrine therapy?Oncol Rep 1994;1:625–9.

96. Elledge RM, Green S, Ciocca D, et al. HER-2expression and response to tamoxifen in estro-gen receptor-positive breast cancer: a South-west Oncology Group study. Clin Cancer Res1998;4:7–12.

97. Diamond A, The Advisory Board Company. Her-ceptin® offers new treatment option foradvanced breast cancer patients. The OncologyRoundtable, Oncology Watch 1998;Issue #7.

98. Park JW, Hong K, Carter P, et al. Development ofanti-p185HER2 immunoliposomes for cancertherapy. Proc Natl Acad Sci U S A 1995;92:1327–31.

99. Pegram M, Lipton A, Pietras R, et al. Phase IIstudy of intravenous recombinant humanizedanti-p185 HER-2 monoclonal antibody (rhuMAbHER-2) plus cisplatin in patients with HER-2/neu overexpressing metastatic breast cancer.Proc Am Soc Clin Oncol 1995;14:A124.

100. Davidson NE, Gelmann EP, Lippman ME, Dick-son RB. Epidermal growth factor receptor geneexpression in estrogen receptor-positive andnegative human breast cancer cell lines. MolEndocrinol 1987;1(3):216–23.

101. Ennis BW, Lippman ME, Dickson RB. The EGFreceptor system as a target for antitumor ther-apy. Cancer Invest 1991;9(5):553–62.

102. Falette N, Lefebvre MF, Meggouh F, et al. Mea-surement of occupied and non-occupied epi-

dermal growth factor receptor sites in 216human breast cancer biopsies. Breast CancerRes Treat 1992;20(3);177–83.

103. Ro J, North SM, Gallick GE, et al. Amplified andoverexpressed epidermal growth factor recep-tor gene in uncultured primary human breastcarcinoma. Cancer Res 1988;48:161–4.

104. Chrysogelos SA, Yarden RI, Lauber AH, MurphyJM. Mechanisms of EGF receptor regulation inbreast cancer cells. Breast Cancer Res Treat1994;31:227–36.

105. Ishii S, Xu Y-H, Stratton RH, et al. Characteriza-tion and sequence of the promoter region of thehuman epidermal growth factor receptor gene.Proc Natl Acad Sci U S A 1985;82:4920–4.

106. Koenders PG, Beex LV, Geurts-Moespot A, et al.Epidermal growth factor receptor-negativetumors are predominantly confined to the sub-group of estradiol receptor-positive human pri-mary breast cancers. Cancer Res 1991;51(17):4544–8.

107. Worthylake R, Opresko LK, Wiley HS. ErbB-2amplification inhibits down-regulation andinduces constitutive activation of both erbB-2and epidermal growth factor receptors. J BiolChem 1999;274:8865–74.

108. King CR, Borrello I, Bellot F, et al. EGF bindingto its receptor triggers a rapid tyrosine phos-phorylation of the erbB-2 protein in the mam-mary tumor cell line SK-BR-3. EMBO J 1988;7(6):1647–51.

109. Buday L, Downward J. Epidermal growth factorregulates p21ras through the formation of acomplex of receptor, Grb2 adapter protein, andSos nucleotide exchange factor. Cell 1993;73(3):611–20.

110. Sadowski HB, Shuai K, Darnell JE Jr, GilmanMZ. A common nuclear signal transductionpathway activated by growth factor andcytokine receptors. Science 1993;261:1739–44.

111. Skoog I, Macias A, Azavedo E, et al. Receptorsfor EGF and oestradiol and thymidine kinaseactivity in different histological subgroups ofhuman mammary carcinomas. Br J Cancer1986;54(2):271–6.

112. Winchester DJ, Goldschmidt RA, Kahn SH, et al.Flow cytometric and molecular prognosticmarkers in 91 male breast carcinoma patients[meeting abstract]. 46th Annual Cancer Sympo-sium in Conjunction with Society of Head andNeck Surgeons; 1993 March 18–21; Los Ange-les, CA: Society of Surgical Oncology; 1993.

113. Toi M, Nakamura T, Mukaida H, et al. Relation-


ship between epidermal growth factor receptorstatus and various prognostic factors in humanbreast cancer. Cancer 1990;65:1980–4.

114. Sainsbury JRC, Farndon JR, Harris AL, SherbetGV. Epidermal growth factor receptors onhuman breast cancers. Br J Surg 1985;72:186–8.

115. Nicholson S, Wright C, Sainsbury JR, et al. Epi-dermal growth factor receptor (EGFr) as amarker for poor prognosis in node-negativebreast cancer patients: neu and tamoxifen failure.J Steroid Biochem Mol Biol 1990;37(6):811–4.

116. Nicholson S, Richard J. Sainsbury C, et al. Epi-dermal growth factor receptor (EGFr); resultsof a 6 year follow-up study in operable breastcancer with emphasis on the node negativesubgroup. Br J Cancer 1991;63:146–50.

117. Sainsbury JR, Farndon JR, Sherbet GV, HarrisAL. Epidermal growth factor receptors andoestrogen receptors in human breast cancer.Lancet 1985;1(8425):364–6.

118. Klijn JG, Berns PM, Schmitz PI, Foekens JA. Theclinical significance of epidermal growth factorreceptor (EGF-R) in human breast cancer: areview on 5232 patients. Endocr Rev 1992;13(1):3–17.

119. Stark A, Hulka BS, Conway JS, et al. Her-2/neuamplification and the risk of subsequent inva-sive breast cancer. [In press]

120. Thor AD, Yandell DW. Prognostic significance ofp53 overexpression in node-negative breastcarcinoma: preliminary studies support cau-tious optimism [editorial]. J Natl Cancer Inst1993;85(3):176–7.

121. Thor AD, Moore DH II, Edgerton SM, et al.Accumulation of p53 tumor suppressor geneprotein: an independent marker of prognosis inbreast cancers. J Natl Cancer Inst 1992;84:845–55.

122. Yandell DW, Thor AD. p53 analysis in diagnosticpathology: biologic implications and possibleclinical applications. Diagn Mol Pathol 1993;2(1):1–3.

123. Thor AD. Prognostic and predictive markers inbreast cancer: issues related to moleculardeterminants of outcome. Breast J 1998;4(5):379–82.

124. Horak EK, Smith K, Bromley L, et al. Mutantp53, EGF receptor and c-erbB-2 expression in

human breast cancer. Oncogene 1991;6(12):2277–84.

125. Nigro JM, Baker SJ, Preisinger AC, et al. Muta-tions in the p53 gene occur in diverse humantumour types. Nature 1989;342:705–8.

126. Davidoff AM, Humphrey PA, Iglehart JD, MarksJR. Genetic basis for p53 overexpression inhuman breast cancer. Proc Natl Acad SciU S A 1991;88(11):5006–10.

127. Prosser J, Thompson AM, Cranson G, Evans HJ.Evidence that p53 behaves as a tumour sup-pressor gene in sporadic breast tumours.Oncogene 1990;5(10):1573–9.

128. Runnebaum IB, Nagarajan M, Bowman M, et al.Mutations in p53 as potential molecular mark-ers for human breast cancer. Proc Natl AcadSci U S A 1991;88(23):10657–61.

129. Bergh J, Norberg T, Sjogren S, et al. Completesequencing of the p53 gene provides prognosticinformation in breast cancer patients, particu-larly in relation to adjuvant systemic therapy andradiotherapy. Nature Med 1995;1(10):1029–34.

130. Hawkins DS, Demers GW, Galloway DA. Inacti-vation of p53 enhances sensitivity to multiplechemotherapeutic agents. Cancer Res 1996;56(4):892–8.

131. Levine AJ. p53, the cellular gatekeeper forgrowth and division. Cell 1997;88(3):323–31.

132. Kleihues P, Schauble B, zur Hausen A, et al.Tumors associated with p53 germline muta-tions: a synopsis of 91 families. Am J Pathol1997;150(1):1–13.

133. Sobol H, Stoppa-Lyonnet D, Bressac-De PailleretsB, et al. BRCA1-p53 relationship in hereditarybreast cancer. Int J Oncol 1997;10:349–53.

134. Marubini E, Valsecchi MG. Analysing survivaldata from clinical trials and observational stud-ies. New York: Wiley & Sons; 1995.

135. Harrington DP, Fleming TR. A class of rank testprocedures for censored survival data. Bio-metrika 1982;69:53–56.

136. Cox DR. Regression models and life tables (withdiscussion). J R Statist Soc 1972;34(B):187–220.

137. Edgington ES. Randomization tests. New York:Marcel Dekker; 1995.

138. Berry DA. When is a confirmatory randomizedclinical trial needed? [editorial]. J Natl CancerInst 1996;88:1606–7.

130 BREAST CANCER

131

Although Broders first defined the pathologicentity of ductal carcinoma in situ (DCIS) in1932, in situ disease remained a clinical curios-ity until the mid-1970s because of the unusualassociation of a palpable mass with noninfil-trating cancer.1 With the widespread acceptanceof screening mammography for breast cancerdetection came a significant increase in thenumber of patients with nonpalpable DCIS.Reports over the last two decades have demon-strated equivalent survival results for the treat-ment of DCIS with mastectomy versus breast-conservation therapy. Yet, the limitations ofearly studies led to ambivalence about the effi-cacy of breast conservation. Illustrating thissentiment was the fact that mastectomy wasmore commonly used than lumpectomy forpatients with DCIS from 1985 through 1991, asreported by the American College of Surgeonsusing the National Cancer Data Base (NCDB).2

This chapter focuses on clinical researchthat has attempted to predict risk factors foripsilateral recurrence after treatment of DCISwith breast conservation therapy.

INCIDENCE

The increased incidence of DCIS over the lasttwo decades has resulted from a significantincrease in the number of screening mammo-grams per year, heightened awareness by radi-

ologists of the natural evolution of calcifica-tions related to DCIS, and technologicadvances in mammography equipment.

Since 1976, the Cancer Incidence and EndResults (CIER) Committee of the AmericanCancer Society, Illinois Division, has publishedincidence data on approximately 85 percent ofpatients treated for cancer at Illinois hospitals. Aretrospective report from the CIER Committeewas based on 10,974 breast cancer patientsdiagnosed from 1970 to 1975.3 Only 2 percentof patients during this time period had in situdisease, reflecting the stage distribution com-monly seen in the era before the availability ofmammography.4,5 The number of patients withDCIS, as a percentage of the total number offemale breast cancer patients, has steadilyincreased to 12.9 percent in 1995, corroboratingdata from the NCDB.2,6 The combined statewideregistry and mammography survey data from1985 to 1994 revealed that 2.2 patients withDCIS were identified per 1,000 mammograms.Data from Evanston Northwestern Healthcarecompiled from 1994 to 1996 demonstrated thatthe number of patients with DCIS graduallyincreased until age 70 years and then remainedconstant thereafter (Figure 8–1).7

Despite the fact that most investigatorsregard DCIS as a preinvasive phase of malig-nant transformation, in some women, it is pre-sent but never becomes clinically relevant. For

8

Surgical Management of Ductal Carcinoma In SituSTEPHEN F. SENER, MDLAURIE H. LEE, PA-C

example, the incidence of occult multicentricDCIS in mastectomy specimens is higher thanthe ipsilateral breast tumor recurrence rate afterlumpectomy, with or without radiation.8 As fur-ther evidence for the heterogeneous natural his-tory of DCIS, seven autopsy series of womenwithout a history of breast cancer were collec-tively evaluated, and it was demonstrated thatthe median prevalence of DCIS was 8.9 percent(range 0 to 14.7 percent), depending on the levelof scrutiny of the pathologic examination.9

TREATMENT OF PRIMARY DUCTAL CARCINOMA IN SITU

Total Mastectomy

Reports on the treatment of DCIS by mastec-tomy serve as historical benchmarks for com-parison with breast conservation therapy, con-sisting primarily of patients presenting prior tothe wide acceptance of mammographic screen-ing. Patients in these series frequently had nip-ple discharge, Paget’s disease of the nipple, orpalpable DCIS, and pathologic review fre-quently demonstrated evidence of invasion.Locoregional recurrence and disease-relatedmortality rates were about 1 percent.10–13 Chestwall recurrences were invasive and defined the

cancers as biologically aggressive.14 The risk ofdisease-related mortality, although low, wasfinite, most likely due to difficulty in identify-ing areas of invasion within the excised breast.Clearly distinct from these subsets are the cur-rent, more common patients with screeningmammograms who present with localized, non-palpable areas of calcification representingDCIS, in whom the incidence of multicentricityor invasion is low.

The difficulty in identifying areas of inva-sion continues to exist today and leads to a lowbut continuing risk of mortality from DCIS,especially for patients with multicentric or geo-graphically large, comedo DCIS.15 Treatmentrecommendations must take into account therisk of noninvasive and invasive recurrence aswell as that of disease-related mortality. Cur-rently, mastectomy is generally reserved forpatients in whom lumpectomy results in eitherpositive pathologic margins or unacceptablecosmesis. A small percentage of patients whomight otherwise be candidates for breast con-servation therapy will elect mastectomy due toa lack of interest in cosmesis, inaccessibility ofradiation treatment facilities, or a history ofconnective tissue disease.

Lumpectomy with Radiation Therapy

The National Surgical Adjuvant Breast Project(NSABP) Protocol B-17 was initiated in 1985to test whether radiation after lumpectomy forlocalized DCIS prevented recurrence of cancerin the surgically treated breast.16–18 The updatedreport (1997) presented findings on 814 eligi-ble patients through 8 years of follow-up. Of403 patients treated by lumpectomy alone, 104(26.8%) had an ipsilateral breast recurrence.Fifty-one (13.4%) recurrences were noninva-sive and 53 (13.4%) were invasive. Of the 411patients treated by lumpectomy and radiation,47 (12.1%) had an ipsilateral breast recurrence.Thirty (8.2%) recurrences were noninvasive,and 17 (3.9%) were invasive. Thus, the additionof radiation to lumpectomy in the treatment of

132 BREAST CANCER

Figure 8–1. Numbers of noninvasive and invasive breastcancers by age among 609 patients treated at EvanstonNorthwestern Healthcare, 1994 to 1996.

localized DCIS significantly reduced the inci-dence of noninvasive and invasive ipsilateralbreast recurrence (p < .000005). Despite a bet-ter disease-free survival in those treated withlumpectomy and radiation (75% versus 62%,p = .00003), overall survival was equivalent(95% versus 94%).

Numerous retrospective studies have beendone to determine variables associated with anincreased risk of local recurrence after breastconservation therapy. The recurrence and sur-vival results of lumpectomy and radiation formammographically detected DCIS are shownin Table 8–1.12,16–24 Ipsilateral breast recurrencerates ranged from 0 to 10 percent at 5 years and8 to 23 percent at 10 years. The most compre-hensive evaluations had the lowest recurrencerates and included mammographic and patho-logic correlation, with microscopic marginanalysis, classification of architectural pattern,determination of tumor size, and use ofpostlumpectomy mammography to assess thecompleteness of excision.

The patterns of recurrence after lumpectomyand radiation for mammographically detectedDCIS are shown in Table 8–2. Most ipsilateralbreast recurrences occurred in the same vicinityas the primary tumor, and approximately 50 per-cent were invasive cancers when detected. Solinand colleagues reported that the median timeinterval from diagnosis of DCIS to an invasive

recurrence was 5 years and to a noninvasiverecurrence was 4 years.24 A longer time intervalto an invasive than a noninvasive recurrence hasalso been reported by two other groups.20,22 InNSABP B-17, 58 percent of all recurrenceswere within 2 years of treatment.17 Fowble andcolleagues concluded from their data thatincreased attention to efforts that assure com-plete excision of DCIS prior to radiation (exci-sion to negative margins > 2 mm, and negativepreradiotherapy mammography) reduced therisk of noninvasive local recurrence. And, asthat type of recurrence was eliminated, lateinvasive recurrence became the predominanttype of ipsilateral breast failure.20

Factors Associated with IpsilateralBreast Recurrence after Lumpectomy

with Radiation Therapy

Treatment-Related Factors

Recent studies have indicated that microscopicmargin status is a predictor for ipsilateral breastrecurrence.12,17,20,21,23,24 For example, in theseries reported by Fowble and colleagues, the5-year actuarial breast recurrence rate was 0percent for patients with negative or unknownmargins and 8 percent for those with positive orclose margins.20 Solin and colleagues reporteda recurrence rate of 29 percent for patients with

Surgical Management of Ductal Carcinoma In Situ 133

Table 8–1. RECURRENCE AND SURVIVAL RESULTS OF LUMPECTOMY AND RADIATION FOR MAMMOGRAPHICALLY DETECTED DCIS

Ipsilateral Cause-SpecificRecurrence (%) Survival (%)

Authors (ref. #) Number of Patients 5 yr 10 yr 5 yr 10 yr Median Follow-Up (y)

NSABP B-1716–18 399* 10 12.1 — 75† 8.0 (mean)Kuske, et al19 44 7 — — — 4.0Fowble, et al20 110 1 15 100 100 5.3Vicini, et al21 105 8.8 10.2 — 99 6.5Hiramatsu, et al22 54 2 23 — 96 6.2Sneige, et al23 31 0 8 — — 7.2Silverstein, et al12 133‡ 7 19 — 97 7.8Solin, et al24 110 7 14 100 96 9.3

*81 percent detected by mammography†8-year disease-free survival‡89 percent detected by mammography

positive or close margins and 7 percent forthose with negative margins.24 The median timeinterval to recurrence was 3.6 years for patientswith positive margins and 4.3 years for thosewith negative margins. From the NSABP B-17trial, the breast recurrence rates, with a meanfollow-up of 43 months, were 10 percent forpatients with positive or unknown margins and4 percent for those with negative margins.17

The presence of malignant-appearing calcifi-cations on preradiation mammography has beenhighly predictive of recurrence. Two series havereported five such patients, all of whom had dis-ease recurrence.23,25 A third series reported thatthere were no recurrences in 37 patients withnegative preradiation mammography.20

Pathologic Factors

Confounding the pathologist’s ability to assessthe adequacy of surgical margins is the growthpattern of DCIS within the duct system. Silver-stein and colleagues demonstrated that even withnegative margins and preradiation mammogra-phy to confirm excision of all calcifications,more than 50 percent of patients had residualDCIS in re-excision or mastectomy specimens.15

Holland demonstrated that mammography mayunderestimate the extent of DCIS by 2 cm in 15to 20 percent of patients.26 An additional studyby Faverly and colleagues demonstrated a differ-ence in the growth patterns of well- and poorlydifferentiated DCIS.27 Poorly differentiated

DCIS grew in a continuous pattern, implyingthat margin assessment should be accurate.However, well-differentiated DCIS grew in adiscontinuous (multifocal) pattern in 70 percentof patients, making margin analysis problematic.Of specimens with multifocal DCIS, about 65percent had gaps < 5 mm, 20 percent had gaps 5 to 10 mm, and 10 percent had gaps >10 mm.

The influence of pathologic factors on ipsi-lateral breast recurrence remains an area foractive investigation. It was initially suggested bySolin and colleagues that high-grade DCIS orcomedo necrosis was associated with a higherrate of breast recurrence.28 However, this serieswith a shorter follow-up underestimated thenumber of recurrences in low-grade and non-comedo DCIS, and recurrences with high-gradeor comedo DCIS were predominant. Longer fol-low-up revealed late recurrences with low-gradeand noncomedo DCIS, a finding also reportedby Silverstein and colleagues.15,29

Combination of Treatment-Related and Pathologic Factors

(Van Nuys Prognostic Index)

In 1995, Silverstein and colleagues devised ascoring system, based on retrospective data,that combined three independent predictors oflocal recurrence after breast conservation treat-ment in patients with DCIS: tumor size, marginwidth, and pathologic classification.30 Scores,ranging from 1 to 3, were assigned to each of

134 BREAST CANCER

Table 8–2. PATTERNS OF IPSILATERAL BREAST RECURRENCE AFTER LUMPECTOMYAND RADIATION FOR MAMMOGRAPHICALLY DETECTED DCIS

Number of Local Invasive Median Time (y) Authors Recurrences Recurrence*(%) Recurrence(%) to Recurrence

NSABP B-1718 47 — 36 —Kuske, et al19 3 33 100 2.6Fowble, et al20 3 33 100 8.8Vicini, et al21 10 70 70 2.4Hiramatsu, et al22 4 75 25 6.1Sneige, et al23 1 — 0 8Silverstein, et al12 16 100 50 4.9Solin, et al24 15 73 40 5

* recurrence within lumpectomy site

the variables defined by multivariate analyses(Table 8–3). The Van Nuys Prognostic Index(VNPI) scoring system was validated usingdata from 394 patients treated for DCIS withbreast conservation: 209 by lumpectomy aloneand 185 by lumpectomy and radiation. Patientswere divided into three groups with differentprobabilities for ipsilateral breast recurrence,on the basis of VNPI scores (3 to 4, 5 to 7, 8 to9). The 12-year local recurrence-free survivalrates were 98 percent for those with VNPI = 3and 4; 70 percent for those with VNPI = 5 to 7;and 28 percent for those with VNPI = 8 and 9.The 12-year breast cancer-specific survivalrates were 100 percent for those with VNPI = 3and 4; 99 percent for those with VNPI = 5 to 7;and 95 percent for those with VNPI = 8 and 9.

In patients with VNPI scores of 3 and 4,there was no difference in disease-free survivalin those treated with lumpectomy and radiationversus those treated by lumpectomy alone. Inpatients with intermediate VNPI scores (5 to 7),there was a 13 percent lower local recurrencerate in those treated with lumpectomy and radi-ation versus those treated with lumpectomyalone (p = .027). Even though there was a sig-nificantly lower local recurrence rate inpatients with VNPI = 8 and 9 treated bylumpectomy and radiation versus lumpectomyalone, close to 60 percent of those treated bylumpectomy and radiation had an ipsilateralbreast recurrence with an 81-month median fol-low-up. Although this prognostic scheme has arational formula, it was defined using a retro-spectively identified cohort. Further validationin a prospective analysis will be important toconfirm its conclusions.

Clinical Factors

Solin and colleagues reported that the ipsilat-eral breast recurrence rate for women < 50years of age was 25 percent compared with 2percent for those ³ 50 years, and the time inter-val to recurrence was 4.9 years for the youngerwomen compared with 8.7 years for the singlerecurrence in women ³ 50 years.24 Van Zee andcolleagues also reported an increased recur-rence risk for women < 40 years of age versusthose ³ 40 years.31 However, other investigatorswere unable to confirm the correlation ofyoung age with increased risk of breast recur-rence after breast conservation treatment.20,22,23

An increased risk of breast recurrence wasassociated with a family history of breast cancerby two groups of investigators.22,25 However,this finding remains unconfirmed by others, sothe impact of family history on ipsilateral breastrecurrence remains uncertain at this time.20

Lumpectomy Alone

Lagios and colleagues proposed that lumpec-tomy alone was appropriate treatment forselected patients with mammographicallydetected DCIS.14,32 Selection criteria includedtumor size of 25 mm or less, histologicallynegative margins of excision, and postopera-tive mammography to confirm the absence ofcalcifications remaining in the breast. Follow-up of the original 79 patients reported in 1989revealed a 15-year actuarial local recurrencerate of 19 percent.33 The local recurrence rateswere 33 percent for patients with high-gradeDCIS, 10 percent for those with intermediate-


Table 8–3. THE VAN NUYS PROGNOSTIC INDEX (VNPI) SCORING SYSTEM

Score 1 2 3

Size (mm) 15 or less 16 to 40 41 or moreMargin width (mm) 10 or more 1 to 9 Less than 1Pathologic class Non–high grade Non–high grade High grade

No necrosis NecrosisNuclear grade 1 to 2 Nuclear grade 1 to 2 Nuclear grade 3

VNPI = size score + margin score + pathologic class score.

grade DCIS, and 6 percent for those with low-grade DCIS. In addition, local recurrence rateswere 68 percent for patients with margins < 1mm, 20 percent for those with margins 1 to 9mm, and 7 percent for those with margins of10 mm or more. Other authors have reportedsimilar results (14 to 27% breast recurrencerates) with follow-up times of 45 to 90months.18,34-37 These data have led to the con-clusion that there may be subsets of patientswith DCIS for whom lumpectomy alone is ade-quate treatment.

The NSABP B-17 trial represents the onlyrandomized comparison of lumpectomy with orwithout radiation therapy. Although the datafrom this study strongly supported the use ofradiation therapy to decrease the risk of ipsilat-eral breast recurrence, DCIS consists of a broadspectrum of disease defined by grade andextent. Failure to scrutinize these pathologicvariations has been a criticism of the study.38

Although there may be subsets of patients thatare adequately treated by lumpectomy alone, asuccessful outcome with this approach isdependent on careful selection of good-riskpatients and demonstration of clear surgicalmargins after lumpectomy.

RESULTS OF TREATMENT FOR IPSILATERAL BREAST TUMOR

RECURRENCE

Survival results after treatment of an ipsilateralbreast recurrence following conservativesurgery for DCIS are shown in Table 8–4.Patients with noninvasive recurrence weretreated by complete total mastectomy and nonedeveloped distant metastases thereafter. In 707patients treated by Silverstein and colleagues(which included 259 patients initially treated bymastectomy), the 8-year local recurrence ratewas 12.5 percent.39 There were 74 recurrences:39 were noninvasive, and 35 were invasive. Atthe time of local invasive recurrence, 18 of 35(51%) were stage I, 4 (11%) were stage IIA, 8(23%) were stage IIB, 4 (11%) were stage IIIB,

and 1 (3%) was stage IV. However, node infor-mation was available for only 5 of the 35patients with invasive recurrence because 30patients initially underwent axillary dissectionwith the treatment of their primary tumor. Thus,some of the remaining 30 patients were probablyunderstaged. At 8-year follow-up after treatmentof the 35 patients for local invasive recurrence,the probability of developing distant metastaseswas 27 percent and the breast cancer–specificmortality rate was 14.4 percent. The 8-yearbreast cancer mortality rate for 448 patientsthat had breast conservation treatment forDCIS was 2.1 percent. The results indicatedthat, regardless of treatment choice, the overallmortality rates were low.

Re-excision of recurrent disease may be aconsideration for patients treated initially withlumpectomy alone. After re-excision, radiationtherapy should be included in the treatment pro-gram. For most, a recurrence occurs in the con-text of an irradiated breast, and a completionmastectomy is usually the treatment of choice.

RECOMMENDATIONS FOR FOLLOW-UPCARE OF PATIENTS WITH DCIS

In 1992, a collaborative effort of the AmericanColleges of Radiology and Surgeons, the Col-lege of American Pathologists, and the Societyof Surgical Oncology led to a publication regard-

136 BREAST CANCER

Table 8–4. RESULTS OF TREATMENT FOR IPSILATERAL BREAST RECURRENCE FOLLOWING

LUMPECTOMY AND RADIATION FOR DCIS

Number Developing Metastases /Total Number

Noninvasive InvasiveAuthors Recurrence Recurrence

NSABP B-1716 0/20 1/8Kuske, et al19 — 0/3Fowble, et al20 — 1/3Vicini, et al21 0/3 1/7Hiramatsu, et al22 0/3 1/1Sneige, et al23 0/1 —Silverstein, et al39 0/39 5/35Solin, et al24 0/9 1/6Total 0/75 10/61 (16%)

ing standards of care for invasive and noninva-sive breast cancer treated by breast conservationtherapy.40 A task force of the same four nationalorganizations published a subsequent separatestandard of care for patients with DCIS.41

The goals of routine follow-up include theidentification of treatment sequelae and earlydetection of recurrent or new breast cancers.Regular clinical examinations and breast imag-ing are the cornerstones of effective follow-upcare. Routine tests for metastatic disease are notindicated for asymptomatic patients after treat-ment for DCIS. Clinical examinations should bedone every 6 months for at least 5 years and per-haps through 8 years, when the risk of ipsilateralbreast recurrence after breast-conservation ther-apy approaches that of contralateral breast can-cer. A preradiation therapy ipsilateral mammo-gram (with magnification views, as necessary)should be done to ensure that there are no resid-ual suspicious calcifications after lumpectomy.A baseline mammogram of the treated breastshould be done during the first year after breast-conservation therapy, and thereafter annually ormore frequently, if warranted by clinical or radi-ographic findings. Mammography of the con-tralateral breast should be done at least annually,depending on clinical or radiographic findings.

REFERENCES

1. Broders AC. Carcinoma in situ contrasted withbenign penetrating epithelium. JAMA 1932;99:1670–4.

2. Winchester DJ, Menck HR, Winchester DP.National treatment trends for ductal carcinomain situ of the breast. Arch Surg 1997;132:660–5.

3. Cancer Incidence and End Results Committee.Cunningham MP, Chairman. Breast cancer: areport by 53 Illinois hospitals on cases diag-nosed 1970–1975. Chicago: American CancerSociety, Illinois Division, Inc.; 1982.

4. Hughes KS, Lee AK, Rolfs A. Controversies inthe treatment of ductal carcinoma in situ. SurgClin North Am 1996;76(2):243–65.

5. Rosen PP, Braun DW, Kinne D. The clinical sig-nificance of pre-invasive breast carcinoma.Cancer 1980;46:919–25.

6. Cancer Incidence and End Results Committee.

Cancer in Illinois: incidence reports. Chicago:American Cancer Society, Illinois Division,Inc.; 1983–1995.

7. Sener SF, Winchester DJ, Winchester DP, et al.Spectrum of mammographically detected breastcancers. Am Surg 1999;65:731–6

8. Schwartz GF, Patchefsky AS, Feig SA, et al. Mul-ticentricity of non-palpable breast cancer. Can-cer 1980;45:2913–6.

9. Welch HG, Black WC. Using autopsy series toestimate the disease “reservoir” for ductal car-cinoma in situ of the breast: how much morebreast cancer can we find? Ann Intern Med1997;127:1023–8.

10. Ashikari R, Hajdu SI, Robbins GF. Intraductalcarcinoma of the breast (1960–1969). Cancer1971;28:1182–7.

11. Kinne D, Petrek JA, Osborne MP, et al. Breast car-cinoma in situ. Arch Surg 1989;124:33–6.

12. Silverstein MJ, Barth A, Poller DN, et al. Ten-yearresults comparing mastectomy to excision andradiation therapy for ductal carcinoma of thebreast. Eur J Cancer 1995;37:1425–7.

13. Rosner D, Bedwani RN, Vana J, et al. Noninvasivebreast carcinoma: results of a national surveyby the American College of Surgeons. AnnSurg 1980;192:139–47.

14. Lagios MD, Westdahl PR, Margolin FR, RoseMR. Duct carcinoma in situ: relationship ofextent of noninvasive disease to the frequencyof occult invasion, multicentricity, lymph nodemetastases, and short-term failures. Cancer1982;50:1309–14.

15. Silverstein MJ, Waisman JR, Gamagami P, et al.Intraductal carcinoma of the breast (208cases): clinical factors influencing treatmentchoice. Cancer 1990;66:102–7.

16. Fisher B, Constantino J, Redmond C, et al.Lumpectomy compared with lumpectomy andradiation therapy for the treatment of intraductalbreast cancer. N Engl J Med 1993;328:1581–6.

17. Fisher ER, Constantino J, Fisher B, et al. Patho-logic findings from the National SurgicalAdjuvant Breast Project (NSABP) Protocol B-17. Intraductal carcinoma (ductal carcinomain situ). Cancer 1995;75:1310–9.

18. Fisher B, Dignam J, Wolmark N, et al. Lumpec-tomy and radiation therapy for the treatment ofintraductal breast cancer: findings from NSABPB-17. J Clin Oncol 1998;16(2):441–52.

19. Kuske RR, Bean JM, Garcia DM, et al. Breastconservation therapy for intraductal carcinomaof the breast. Int J Radiat Oncol Biol Phys1993;26:391–6.


20. Fowble B, Hanlon AL, Fein DA, et al. Results ofconservative surgery and radiation for mam-mographically detected ductal carcinoma insitu (DCIS). Int J Radiat Oncol Biol Phys1997;38(5):949–57.

21. Vicini FA, Lacerna MD, Goldstein NS, et al. Duc-tal carcinoma in situ detected in the mammo-graphic era: an analysis of clinical, pathologic,and treatment-related factors affecting out-come with breast-conserving therapy. Int JRadiat Oncol Biol Phys 1997;39(3):627–35.

22. Hiramatsu H, Bornstein BA, Recht A, et al. Localrecurrence after conservative surgery and radi-ation therapy for ductal carcinoma in situ. Pos-sible importance of family history. Cancer JSci Am 1995;1:55–61.

23. Sneige N, McNeese MD, Atkinson EN, et al. Duc-tal carcinoma in situ treated with lumpectomyand irradiation: histopathologic analysis of 49specimens with emphasis on risk factors andlong term results. Hum Pathol 1995;26:642–9.

24. Solin LJ, McCormick B, Recht A, et al. Mammo-graphically detected, clinically occult ductalcarcinoma in situ (intraductal carcinoma)treated with breast conserving surgery anddefinitive breast irradiation. Cancer J Sci Am1996;2:158–65.

25. McCormick B, Rosen PP, Kinne D, et al. Ductalcarcinoma in situ of the breast: an analysis oflocal control after conservative surgery andradiotherapy. Int J Radiat Oncol Biol Phys1991;21:289–92.

26. Holland R, Hendriks JHCL, Verbeek ALM, et al.Extent, distribution, and mammographic/histo-logical correlations of breast ductal carcinomain situ. Lancet 1990;335:519–22.

27. Faverly DRG, Burgers L, Bult P, Holland R. Threedimensional imaging of mammary ductal car-cinoma in situ: clinical implication. SeminDiagn Pathol 1994;11:193–8.

28. Solin LJ, Yeh IT, Kurtz J, et al. Ductal carcinoma insitu (intraductal carcinoma) of the breast treatedwith breast-conserving surgery and definitiveirradiation. Cancer 1993;71:2532–42.

29. Solin LJ, Kurtz J, Fourquet A, et al. Fifteen-yearresults of breast conserving surgery and defin-itive breast irradiation for the treatment of duc-tal carcinoma in situ of the breast. J Clin Oncol1996;14:754–63.

30. Silverstein MJ. The Van Nuys Prognostic Index.In: Silverstein MJ, editor. Ductal carcinoma insitu of the breast. Baltimore: Williams andWilkins; 1997. p. 491–501.

31. Van Zee P, Liberman L, McCormick B, et al. Longterm follow-up of DCIS treated with breastconservation: effect of age and radiation[Abstr]. Soc Surg Oncol 1996;82:26.

32. Lagios MD, Margolin FR, Westdahl PR, RoseMR. Mammographically detected ductal carci-noma in situ. Frequency of local recurrencefollowing tylectomy and prognostic effect ofnuclear grade and local recurrence. Cancer1989;63:618–24.

33. Lagios M. Lagios experience. In: Silverstein MJ,editor. Ductal carcinoma in situ of the breast.Baltimore: William and Wilkins; 1997. p. 361–5.

34. Shreer I. Conservation therapy of DCIS withoutradiation. Breast Dis 1996;9:27–36.

35. Arnesson LG, Olsen K. Linkoping experience. In:Silverstein MJ, editor. Ductal carcinoma in situof the breast. Baltimore: Williams and Wilkins;1997. p. 373–8.

36. Schwartz GF. Treatment of subclinical ductal car-cinoma in situ by excision and local surveil-lance. In: Silverstein MJ, editor. Ductal carci-noma in situ of the breast. Baltimore: Williamsand Wilkins; 1997. p. 353–60.

37. Silverstein MJ. Van Nuys experience by treatment.In: Silverstein MJ, editor. Ductal carcinoma insitu of the breast. Baltimore: Williams andWilkins; 1997. p.443–8.

38. Page D, Lagios M. Pathologic analysis of theNational Surgical Adjuvant Breast Project(NSABP) B-17 Trial. Unanswered questionsremaining unanswered considering currentconcepts of ductal carcinoma in-situ [Editor-ial]. Cancer 1995;75:1219–22.

39. Silverstein MJ, Lagios MD, Martino S, et al. Out-come after invasive local recurrence in patientswith ductal carcinoma in situ of the breast. JClin Oncol 1998;16:1367–73.

40. Winchester DP, Cox JD. Standards for breast-con-servation treatment. CA Cancer J Clin 1992;42:134-62.

41. Winchester DP, Strom EA. Standards for diagno-sis and management of ductal carcinoma insitu of the breast. CA Cancer J Clin 1998;48(2):108–28.

138 BREAST CANCER

139

Implementation of screening mammographyand increased breast cancer awareness accountfor the vast majority of breast cancers present-ing at an earlier stage. This combined withextensive data supporting the choice of breastpreservation has lead to a dramatic change inthe treatment for early stage breast cancer. Hal-sted firmly established radical mastectomy asthe sole surgical procedure for breast cancer.Although this provided improved locoregionalcontrol of disease, the results were disfiguring(Figure 9–1, right breast). Clinical trials haveled to the evolution of therapy that has lessenedphysical deformity (Figure 9–1, Figure 9–2)and improved survival.

DIAGNOSTIC EVALUATION

Although the surgeon is sometimes the firstphysician that encounters the patient with breastcancer, this diagnosis may be initially suspectedby primary care physicians or other specialists.Although few of these physicians will bedirectly involved in the diagnostic procedures,all should be familiar with the key issues rele-vant to the initial evaluation of women with sus-pected breast cancer.

A suspicious finding or an interval changeon a mammogram may require additional imag-ing studies. Prior to obtaining a histologic or

9

Evaluation and Surgical Management of Stage I and II Breast CancerDAVID J. WINCHESTER, MD, FACS

cytologic diagnosis, a thorough breast exami-nation and an explanation of options should beconducted by the surgeon. If a stereotactic corebiopsy is performed prior to surgical evalua-tion, it is possible that a subtle physical findingmay be overlooked, leading to a more compli-cated image-directed biopsy instead of an in-office needle biopsy. If complicated by a post-procedural hematoma, the therapeutic surgicalprocedures may also require localization.Although large postprocedural hematomas areunusual, when they do occur, the physicalexamination may have very little role in formu-lating a treatment plan.

Figure 9–1. Metachronous bilateral breast cancers treatedwith radical mastectomy (left) and modified radical mastec-tomy (right).

The same concerns also apply to the radio-graphic evaluation. Prior to embarking on a tis-sue diagnosis, it is important to obtain a mam-mogram and additional views or a sonogram, ifindicated. The usefulness of these studies isadversely affected by a breast hematoma. It isalso imperative that the contralateral breast bethoroughly evaluated by physical examination,mammography, and, if indicated, sonography.

After the physical examination and radio-graphic studies, a cytologic or histologic diagno-sis is required prior to any therapeutic proce-dures. If the lesion can be appreciated on physicalexamination, an office-based aspiration or corebiopsy is simple, cost effective, and expeditiousand simplifies the subsequent treatment in termsof localization of the malignancy. The diagnosiscan be reliably obtained with either a cytologicaspiration or a histologic core biopsy.

Establishing a diagnosis with core biopsyprovides histologic confirmation of malignancyand has the ability to distinguish between inva-sive and in situ carcinoma. This approachrequires a local anesthetic and has a greaterpotential for formation of a hematoma. Com-pared with core biopsy, fine-needle aspiration(FNA) cytology is a simpler, less invasive tech-nique but requires expertise in the preparationand interpretation of cytologic preparations. Ithas very limited capabilities in distinguishing

invasive from in situ tumors, although there maybe reliable signs in certain tumor types.1 Hor-mone receptor assays and immunohistochemi-cal stains for prognostic markers such as HER-2/neu and p53 can be determined from bothcore samples and cytology preparations. Thefalse-negative and false-positive rates of corebiopsy and FNA cytology are comparable.2,3

THERAPY

The treatment of breast cancer continues to berefined to an individualized approach thatstrives to preserve the breast, chest wall mus-cles, and lymphatics, when possible. To achieveimproved survival, multimodality therapy hasalso been increasingly used, but according toneed and efficacy. To define the optimal local,regional, and systemic therapies of breast can-cer, the patient needs to be staged according tothe TNM (tumor, nodes, metastases) stagingsystem defined by the American Joint Commit-tee on Cancer4 (Tables 9–1 and 9–2). For mostpatients with early stage breast cancer, surgicalintervention serves as the first phase of treat-ment. This step also moves beyond clinical stag-ing to pathologic staging to provide importantprognostic information to direct adjuvant ther-apy decisions. In addition to the stage of thetumor, the presence of ductal carcinoma in situ

140 BREAST CANCER

Figure 9–2. A and B, Excellent cosmetic outcome from breast preserving therapy (left breast).

A B

(DCIS) also has implications for the local man-agement of breast cancer. Small invasive breastcancers accompanied by extensive DCIS mayrequire total mastectomy.

Breast Preservation Therapy

Adherence to screening mammography guide-lines has made most patients candidates forbreast preservation therapy (BPT). Acceptanceof this treatment approach has been gradual anddependent upon regional preferences and avail-ability of radiation therapy facilities.5 Whenconsidering treatment options for early breastcancer, good cosmesis is an important goal.Improvements in surgery and radiotherapy haveminimized the incidence of poor results seeninitially (Figures 9–4 and 9–5). For mostpatients, the best cosmetic result can beachieved with breast preservation therapy and,

along with that, the shortest recovery as com-pared with a mastectomy with reconstruction.Multiple prospective randomized studies haveconfirmed the efficacy of BPT.6–11 No studyhas identified a survival disadvantage of thisapproach. A nonrandomized comparison12 aswell as a meta-analysis of randomized trials13

have shown equivalent survival rates betweenthese two approaches.

Patient Selection

Despite the extensive data to support the use ofBPT, there are still patients who are not candi-dates for this approach. As surgery is usually thefirst treatment modality, the choice of therapy isguided by the surgeon’s evaluation. A moreeffective evaluation process ideally includes thepreoperative evaluation from other treatingphysicians to provide a cohesive and compre-hensive treatment plan before therapy begins.

Most randomized clinical trials includedpatients with T1 and T2 tumors.6–11 Thus,tumors up to 5 cm can be safely managed withthis approach. A good cosmetic result may bedifficult to achieve with large T2 tumors, andneoadjuvant chemotherapy may improve theability to preserve breast tissue without com-promising survival.14,15 Saving the breast isimportant for psychological and cosmetic rea-sons. This requires sound surgical judgment andmeticulous techniques. A large breast can more

Evaluation and Surgical Management of Stage I and II Breast Cancer 141

Table 9–1. TUMOR-NODE-METASTASIS (TNM) CLASSIFICATION SYSTEM FOR BREAST CANCER

Primary Tumor (T)TX Primary tumor cannot be assessedT0 No evidence of primary tumorTis Carcinoma in situ or Paget’s disease of the

nipple, with no associated tumorT1 Tumor 2 cm in greatest dimension

T1a 0.5 cmT1b > 0.5 cm and 1.0 cmT1c > 1.0 cm and 2.0 cm

T2 Tumor > 2 cm, and 5 cm in greatest dimensionT3 Tumor > 5 cm in greatest dimensionT4 Tumor of any size with direct extension to chest

wall or skinT4a Extension to chest wallT4b Edema, ulceration, or satellite nodulesT4c Both T4a and T4bT4d Inflammatory carcinoma

Regional Lymph Nodes (N)NX Regional lymph nodes cannot be assessedN0 No regional lymph node metastasisN1 Metastasis to ipsilateral axillary lymph node(s)N2 Metastasis to ipsilateral axillary node(s) fixed to

one anotherN3 Metastasis to ipsilateral internal mammary lymph

node(s)Distant Metastasis (M)

MX Presence of distant metastasis cannot be assessed

M0 No distant metastasisM1 Distant metastasis (includes metastases to supra

clavicular lymph node[s])

Table 9–2. BREAST CARCINOMA STAGES

STAGE T N M

0 Tis N0 M0I T1 N0 M0IIA T0 N1 M0

T1 N1 M0T2 N0 M0

IIB T2 N1 M0T3 N0 M0

IIIA T0 N2 M0T1 N2 M0T2 N2 M0T3 N1 or N2 M0

IIIB T4 Any N M0Any T N3 M0

IV Any T Any N M1

Figure 9–6. Paget’s disease.

readily accommodate a larger lumpectomy thana smaller breast. Resection of more than a quad-rant of the breast begins to have a significantimpact on the cosmetic result and leads to con-sideration of alternatives such as mastectomywith reconstruction.16 With the development ofmicrovascular techniques, reconstructive optionsand results have made total mastectomy a goodoption for many patients.

In the preoperative evaluation of the patient,the size of the primary tumor is an importantdiscriminator in the selection of the surgicaltherapy. Given an acceptable ratio between thesize of the tumor and the size of the breast,

patients who are interested in cosmesis andcommitted to radiation therapy are good candi-dates for BPT. Centrally located lesions includ-ing Paget’s disease (Figure 9–6) were at onetime considered a relative contraindication toBPT. Adequate treatment of tumors in this loca-tion may necessitate resection of a portion or allof the nipple-areolar complex (Figure 9–7).This treatment has the distinct advantage ofmaintaining the breast mound and a sensatebreast. Nipple reconstruction may be per-formed, if desired.

Physical findings and mammographicdimensions may not correlate with the histo-logic findings after a lumpectomy is per-formed.17 Thus, the choice of lumpectomy isultimately contingent upon the ability toachieve histologically clear surgical margins.Close or involved surgical margins are impor-tant predictors of local failure and shouldprompt consideration of either re-excision orcompletion mastectomy, depending on theextent of the margin involvement.18–20

Placement of the incision is important tocreate a good cosmetic result and to allow foradditional surgery in the case of microscopi-cally involved margins. Optimal cosmesis usu-ally places incisions within skin folds or in acurvilinear fashion around the nipple (Figure9–8). Incisions should be placed directly over

142 BREAST CANCER

Figure 9–5. Poor cosmetic outcome resulting from radiationinjury.

Figure 9–4. Poor cosmetic outcome resulting from poorincision placement, incision size, and hematoma formation.

the primary to avoid tunneling and limit thedeformity and extent of dissection in the breast.With the exception of superficial lesions, resec-tion of skin or subcutaneous tissue is notrequired. A small ellipse of skin may be helpfulfor specimen orientation.

Aside from the size of the tumor, multicen-tric tumors or extensive intraductal cancer arealso important in identifying poor candidatesfor BPT. Clearing the surgical margin for in situdisease is equally important for local control.Mammographically occult in situ disease maytherefore have the potential to alter the surgicaltherapy for even small invasive tumors. Syn-chronous tumors located in different quadrantsmust be approached with the same margin cri-teria as those for solitary lesions, leading to acompromised cosmetic result. In addition, mul-ticentric disease also suggests that other areasin the breast may contain unrecognized foci ofcancer. Two or more ipsilateral tumors shouldlead to strong consideration of a mastectomywith reconstruction.

The histologic subtype of the breast primarymay have an impact on the local management.In addition to DCIS, which has the propensityto extend great distances in the breast withoutany mammographic or physical findings, inva-sive lobular carcinoma may also have a perva-sive presentation. In the case of in situ ductalcarcinoma, preoperative magnification viewsmay help identify extensive pleomorphic calci-fications around a stellate mass. The surgicalprocedure should attempt to remove all suspi-cious calcifications.

Invasive lobular carcinomas have a moreindolent presentation with a less defined masswith indistinct borders. Mammographic find-ings are subtle21,22 and more likely to underesti-mate tumor dimensions, compared with otherinvasive cancers.17 These characteristics accountfor the greater likelihood of requiring re-excision after lumpectomy.17 The histologicevaluation of the lumpectomy and regionallymph nodes is more difficult because of thefrequency of single malignant lobular cells that

can extend into the breast parenchyma. Cyto-keratin stains may facilitate the identificationof lobular cells but have uncertain prognosticinformation in axillary staging.23

Randomized clinical trials addressing breastpreservation therapy have not separated orexcluded lobular carcinomas.6–11 Several non-randomized studies have found no difference inthe local disease-free survival rates betweenbreast preservation patients with lobular carci-noma and those with ductal carcinoma24–26

whereas others have noted a difference.17,27–29

Analysis of the National Cancer Data Base didnot identify any significant differences in size,stage, or survival according to histology.30 Itwould appear that there are not any specific his-tologic categories that should exclude consider-ation of BPT. The same principles of a carefulpreoperative assessment and microscopic eval-uation of lumpectomy margins should lead tosuccessful BPT for all histologic variants.


Figure 9–8. Incision placement for lumpectomy.

Figure 9–7. Central lumpectomy.

Risk Factors for Local Recurrence

Several important variables exist for localrecurrence after lumpectomy. The only variableto predict local recurrence from analysis ofNational Surgical Adjuvant Breast and BowelProject (NSABP) B-06 was age under 45years.31 Other variables analyzed included vas-cular and lymphatic invasion, tumor grade, andsize.31 Treatment-related variables include theextent of resection, margin involvement, andthe implementation of radiotherapy. Thus, therisk of local recurrence can be dramaticallyaffected by treatment decisions.

Defining the optimal margin distance isbased on a subjective and individualized assess-ment of each patient, balancing the cosmeticoutcome and pathologic characteristics. Whencomparing a more extensive quadrantectomywith lumpectomy, the risk of a local recurrence isreduced significantly in the former.32 Despite asignificant reduction in local recurrence with amore extensive resection,32 radiation therapy hasbeen shown to add to local control in thesepatients.33 Margin involvement is a strong predic-tor of local recurrence, and identification of aninvolved margin should prompt consideration ofre-excision or completion mastectomy.19,20

Despite the usefulness of a microscopic marginassessment, clear surgical margins under themost stringent conditions still do not ensure localcontrol rates that are equivalent to those achievedwith the addition of radiotherapy. In the UppsalaSwedish trial, only patients with tumors < 20mm were included. Each patient underwent asector resection consisting of removal of a por-tion of the skin and pectoralis fascia. Each mar-gin was assessed twice. Any microscopic margininvolvement or lymph node involvement was anexclusion criterion. Patients were randomized toobservation or radiotherapy after sector resec-tion. Despite these favorable conditions andcareful analysis of margins, local recurrence wassignificantly more common in the observationarm of the study.34 Serial sectioning studies ofmastectomy specimens of patients that would be

lumpectomy candidates have also shown thatmicroscopic foci of cancer are identified beyond2 cm of the primary in 41 percent of patients.35

Although an adequate margin is important, amore extensive resection needs to be balancedwith the cosmetic result of the operation. In mostinstances, resection of the pectoralis fascia withthe lumpectomy specimen will avoid concernsabout posterior extension. Without muscleinvolvement, inclusion of the pectoralis fasciawith the lumpectomy specimen should assist ingood local control, even with a close margin.

The handling of the surgical specimenbecomes critical when close or involved micro-scopic margins are identified on the lumpec-tomy specimen. Inability to accurately definespecimen orientation risks having to remove toolarge a specimen or removing the wrong area ofpersistent involvement. The surgeon and thepathologist should work closely together at thetime of surgery. Specimens should either beinked on six sides by the surgeon or have appro-priate markers to allow the pathologist to do so.Specimens should be submitted fresh for patho-logic examination so that any questions aboutthe orientation can be addressed immediately.

Achieving good hemostatis is important forobvious reasons. Large lumpectomy cavities canbe defined for the radiotherapist by placingradiopaque surgical clips at the borders of thespecimen. This may help facilitate the delivery ofa radiation therapy boost to the lumpectomy site.

Total Mastectomy

Total mastectomy remains an excellent choice formany patients with breast cancer. A clear advan-tage of mastectomy is the avoidance of radiationtherapy for patients without large tumors or mul-tiple involved lymph nodes. This has more appealfor patients that are not motivated to achievegood cosmesis. Older, less mobile patients mayfind this preferable to the alternative of lumpec-tomy and radiation therapy.

Total mastectomy is indicated for multicen-tric disease or tumors with extensive coexistent

144 BREAST CANCER

DCIS, where achieving a clear surgical marginbecomes difficult with a segmental mastec-tomy. It is also indicated for individuals whoare not radiation therapy candidates, includingthose with active scleroderma, history of priorradiotherapy, ataxia telangectasia, and earlypregnancy and for those who opt for it. Excel-lent cosmetic results can be achieved with avariety of reconstructive options, which canoccur either simultaneously or as a delayed pro-cedure. If a patient is contemplating recon-struction, a skin-sparing mastectomy should beperformed. This operation involves the removalof the nipple-areolar complex and breast tissuebut differs from a standard incision in preserv-ing as much of the skin over the breast as pos-sible (Figure 9–9).

Most patients with early-stage breast cancercan undergo immediate reconstruction. This hasthe advantages of limiting the surgical interven-tions to a single-stage procedure and providingthe patient with the psychological benefit of animmediately reconstructed breast. Immediatereconstruction also best preserves the elasticityof the elevated flaps and helps maintain the nat-ural contour of the breast, including the infra-mammary fold, which may be affected with adelayed reconstruction. Considerations fordelayed reconstruction include urgency toaddress adjuvant systemic treatment, a patientwho remains undecided regarding reconstruc-tion options, or a patient who is likely to receivechest wall radiation therapy. Although radiationtherapy can be successfully delivered after auto-genous reconstruction with good cosmeticresults, the incidence of capsular contractionafter radiation therapy is prohibitive in thosepatients undergoing implant reconstruction.36

For patients with a strong familial history ofbreast cancer, a decision may be made to com-bine a treatment operation with a prophylacticprocedure. The identification of breast cancersusceptibility genes has fostered this concept,but in practical terms, it is very difficult toassess risk and screen for a genetic mutation ina timely fashion before embarking on a thera-

peutic operation for a diagnosis that led to thegenetic evaluation. Counseling these patientscan be very difficult as they have to cope withboth a diagnosis of cancer and an emotionaldecision as to whether or not they wish toundergo a bilateral mastectomy. For thosepatients who might have greater difficulty inreaching a comfortable decision regarding abilateral operation, a safe approach is to pro-ceed with a lumpectomy and axillary stagingprocedure in conjunction with genetic counsel-ing, with or without genetic testing. With thisapproach, the more important delivery of sys-temic therapy is not delayed. Duringchemotherapy, the time-consuming process ofgenetic testing can be performed, if indicated.If the patient is found not to carry a geneticmutation, radiation therapy serves as the laststep of the treatment plan. For those patientswith an identified mutation, a completion mas-tectomy and contralateral prophylactic mastec-tomy with reconstruction can be performed.Although a prophylactic mastectomy does notguarantee prevention of future breast cancerevents, recent data suggest that it is an effectivemeans of reducing risk.37 As an alternative,tamoxifen can be used for both adjuvant treat-ment and chemoprevention. Although data arerelatively early, a clear reduction in high-riskpatients was identified in a randomized studyof tamoxifen users.38 If elected, compartmen-talizing treatment and prophylactic issues helpsto ease the sudden burden of complex decisionsthat a patient will face at the time of diagnosis.


Standard Skin-sparing

Figure 9–9. Incision placement for mastectomy.

Radical Mastectomy

In the context of early breast cancer, there is vir-tually no need to resect the pectoralis musclesand axillary tissue. Occasionally, tumors locatedposteriorly along the chest wall may focallyinvade the pectoralis muscle. It should be pointedout that invasion of the pectoralis muscle doesnot constitute chest wall invasion and is stagedaccording to the size of the primary tumor.4

Small breast cancers that present with muscleinvolvement are usually located peripherally orposteriorly, and extension, in part, reflectsproximity to the muscle. This scenario can besafely managed with resection of a portion ofthe muscle as part of either a lumpectomy or atotal mastectomy. With either surgical approach,radiation therapy should be considered.

Management of the Axilla

Just as treatment of the primary tumor of thebreast has evolved from a single, radical opera-tion for all scenarios to a more directedapproach consisting of lumpectomy, the stan-dard axillary dissection is quickly beingreplaced by sentinel lymphadenectomy. Intro-duced by Morton and colleagues in 1992 for thetreatment of melanoma,39 this technique wasquickly applied to breast cancer.40,41 Like lym-phatic mapping for other disease sites, the sen-tinel lymph node is identified through the con-stant anatomic relationship between a tumorand draining lymphatics. Conceptually, eachspecific area in the breast drains to a sentinellymph node which may be located anywherewithin the axilla or internal mammary chain(Figure 9–10). Larger tumors may have morethan one draining lymphatic (Figure 9–11). Thesentinel lymph node biopsy continues to berefined and defined for patients with earlybreast cancer; in several studies, it has beendemonstrated to yield reliable correlation to anaxillary dissection.40–46

The axillary dissection has always been rec-ognized as an excellent procedure for two

important reasons: staging of the breast cancerand providing regional control. Lymph nodeinvolvement represents the most importantvariable, aside from metastatic disease, to pre-dict outcome.47 This information is importantin defining the prognosis and in tailoring theadjuvant therapy of patients with breast cancer.Staging of the axilla represents a critical vari-able in defining the prognosis of patients pre-senting with early breast cancer. In the contextof early detection and screening mammog-raphy, nodal involvement is at times the onlyprognosticator that leads to the clear recom-mendation of chemotherapy.

Aside from providing important prognosticinformation, axillary dissection represents themost effective means of controlling regionaldisease.48 What constitutes an adequate axillarydissection? This has been well established withmultiple studies analyzing the inclusion ofmetastatic disease based on the arbitrary divi-sion of level I, II, and III axillary lymphnodes.49–51 On average, there is a one percentchance of metastatic disease in level III lymphnodes that would not be detected in levels I andII.49–51 Mathiesen demonstrated that the poten-tial for identifying micrometastases increasedtill 10 lymph nodes were removed from theaxilla.52 Unless extensive axillary involvementis recognized at the time of surgery, a level I/IInode dissection should encompass axillary dis-ease in 99 percent of patients.

In the presence of axillary disease, an axil-lary dissection is an excellent operation forregional control and prognostic information.However, for the great majority of patients withearly breast cancer, an axillary dissection doesnot confer any therapeutic benefit. The greatestconcern, particularly for younger, activepatients is the risk of developing lymphedema.This risk is directly related to the extent of theaxillary dissection and is further increased withthe addition of radiation therapy.53 This riskremains indefinitely for the life of the patient.Other potential side effects include paresthe-sias, loss of mobility, and cosmetic deformity.

146 BREAST CANCER

Avoiding the side effects of an axillary dis-section in a substantial number of patients whodo not achieve any therapeutic benefit has beena major impetus in identifying an alternativemeans of staging the axilla. Other methods thathave been evaluated to replace axillary dissec-tion have been imaging studies including ultra-sonography, computed tomography, and scintig-raphy. All these techniques have had the samemajor limitation of an unacceptably high false-negative rate.54–56 Positron emission tomog-raphic (PET) scanning has emerged as a moresensitive test that relies on the metabolic differ-ences of tumors rather than on anatomic changes.To date, this test shows promise but is still not

sensitive enough to exclude the presence of axil-lary disease.57 It is unlikely that any imagingtechnology will compare with the sensitivity of amicroscopic examination, which has the poten-tial to identify single metastatic cells. The short-coming of a standard axillary dissection is thatpathologists are incapable of reviewing every cellof every lymph node. The ability to detect micro-metastases is directly related to the intensity withwhich a lymph node is analyzed. Serial section-ing studies have identified a higher incidence oftrue nodal positivity and mortality in those withunrecognized micrometastases.23,58 Outside ofinvestigational studies, serial sectioning of axil-lary dissection specimens is impractical.

Sentinel lymphadenectomy has several con-ceptual advantages over standard axillary dissec-tion. Most significant to the patient is that therisk of long term complications is virtually elim-inated by avoiding an extensive axillary dissec-tion. Recovery is much shorter, and for most,BPT, under these circumstances, can be accom-plished as an outpatient procedure. Compared toa standard axillary incision, the sentinel lymphnode can be removed through a smaller incisionwith transcutaneous localization of the nodewith a hand held gamma probe (Figure 9–12). Inaddition to the reduction in morbidity, sentinellymphadenectomy provides the pathologist withthe opportunity to perform a much more com-prehensive analysis of the specimen, given themore limited material to analyze. Although diffi-


Figure 9–10. Breast lymphatic drainage patterns.

Figure 9–11. Lymphatic drainage pattern of an upper breasttumor.

Figure 9–12. Transcutaneous localization of axillary sentinellymph node.

Figure 9–13. Lymphscintigram of breast pri-mary and axillary sentinel lymph nodes.

cult to prove, this may, in fact, provide a moresensitive means of staging the axilla, providedthe sentinel node is correctly identified.

The immediate question addressed by pre-liminary studies has been the ability of sentinellymphadenectomy to detect micrometastases asefficiently as an axillary dissection can. As aresult, comparative studies have included bothoperations in the same patient. Although the endpoints for these studies have all been the same,the technique has varied widely. Conceptually, avisual tracer, a radiolabeled protein, or a combi-nation of both are injected around the breasttumor. Initially, this process was performed withisosulfan blue, a vital blue dye used for lym-phatic mapping for melanoma. This approachcreated difficulties in defining the breast lym-phatics because of the three-dimensional natureof breast cancers and the potential for missingmultiple sentinel nodes located within the three-dimensional axillary nodal basin. Because of thebrevity of time in which isosulfan blue migratesthrough the sentinel node, the timely identifica-tion of multiple sentinel nodes is difficult.

The introduction of radiocolloid for this tech-nique greatly enhanced this procedure by provid-ing a second means of localizing a sentinel nodeusing a hand-held gamma probe. This has simpli-fied the procedure by obviating the lymphaticmapping required to identify the blue lymphnode. Additionally, technitium-labeled sulfur col-loid binds to lymphatic tissue and provides amuch greater window of opportunity to localizesentinel lymph nodes. Lymphoscintigraphy doneprior to the surgical procedure can assist in con-firming the migration and location of radiocol-loid (Figure 9–13). Another conceptual benefit ofthe use of radiocolloid is that it makes it possibleto localize internal mammary nodes. For tumorslocated medially in the breast, these may be theonly sentinel nodes for the tumor.

Although data are quickly emerging to sup-port sentinel lymphadenectomy in the staging ofbreast cancer, it has not yet become uniformlyaccepted as the standard of care. A question yetto be answered is the optimization of the tech-

nique to most accurately identify the sentinelnode. Variables to be defined include the loca-tion of the injection, the radiopharmaceuticalcompound and optimal size, the time intervalbetween the injection and the surgical proce-dure, the combination of radiocolloid with iso-sulfan blue, size limitations of the tumor, effectof excisional biopsy on the accuracy of sentinelnode localization, and the microscopic and sub-microscopic evaluation of nodal tissue. Thelong-term regional recurrence risk after a sen-tinel lymph node biopsy has yet to be addressed.Until these questions have been answered, sen-tinel lymphadenectomy should be performed asa protocol.

Adjuvant Therapy

The surgical treatment for early breast cancerserves as a very important therapeutic step butalso provides important prognostic informationto define subsequent adjuvant therapy deci-sions. Adjuvant chemotherapy has evolved froma narrowly defined node-positive indication tomore encompassing indications based on identi-fied benefits. Nonetheless, the efficacy of com-bination chemotherapy is well correlated withnodal involvement, and surgical staging, partic-ularly for early breast cancer, is important indefining optimal adjuvant therapy decisions.

148 BREAST CANCER

RECURRENCE

An ipsilateral breast recurrence after BPT may bedifficult to distinguish from a second primarytumor. Information important to help make thisdistinction includes the location of the recurrencerelative to the initial primary tumor, the histo-logic features, and the disease-free interval. Prox-imity to a previous lumpectomy site increases thelikelihood of the tumor being a recurrence. In theabsence of any previous history of in situ carci-noma, an in situ component is suggestive of anew primary lesion. Local recurrences are mostlikely to occur within several years of initialtreatment. A long disease-free interval is moresuggestive of a second primary tumor. The dis-tinction between a new primary tumor and recur-rent breast cancer is more important in under-standing the biology of the disease and theefficacy of the treatment selected. In practicalterms, an ipsilateral event is managed in a simi-lar fashion with either scenario. Without theability to deliver additional radiotherapy in mostpatients, re-excision is not a safe option, andmost patients are treated with a completion mas-tectomy. An ipsilateral event should also lead toa metastatic evaluation; recurrences are com-monly the harbinger of metastatic disease.59

In patients previously treated with mastec-tomy, a chest wall recurrence can usually bemanaged with a local excision. The goal of exci-sion should be a margin-free resection. In somesituations, this may necessitate resection ofmuscle or a portion of the chest wall. In patientswho have undergone previous radiation treat-ments, a more extensive resection may representthe only therapeutic option to achieve local con-trol. Closure of the chest wall defect may befacilitated by a myocutaneous flap closure. Hor-monal and cytotoxic chemotherapy must also beconsidered at the time of a local recurrence.

CONCLUSION

Local and regional surgical treatment of earlybreast cancer continues to evolve toward a more

limited and tailored approach. Diagnostic eval-uation also continues to improve in defining theextent of the disease and in providing accuratestaging information to guide resection andadjuvant therapy. With these strategies, thetreatment of breast cancer has become moreprecise and effective.

REFERENCES

1. Shin HJ, Sneige N. Is a diagnosis of infiltratingversus in situ ductal carcinoma of the breastpossible in fine-needle aspiration specimens?Cancer 1998;84:186–91.

2. Shabot M, Goldberg I, Schick P, et al. Aspirationcytology is superior to Tru-Cut needle biopsyin establishing the diagnosis of clinically sus-picious breast masses. Ann Surg 1982;196:122–26.

3. Ballo SM, Sneige N. Can core needle biopsyreplace fine-needle aspiration cytology in thediagnosis of palpable breast carcinoma? Acomparative study of 124 women. Cancer 1996;78:773–7.

4. American Joint Committee on Cancer. Breast. In:Fleming ID, Cooper JS, Henson DE, et al. edi-tor. AJCC Cancer Staging Manual. 5th Edition.Philadelphia: J. B. Lippincott; 1997. pp. 171–80.

5. Winchester DJ, Menck HR, Winchester DP.National treatment trends for ductal carcinomain situ of the breast. Arch Surg 1997;132(6):660–5.

6. Sarrazin D, Le M, Rouesse J, et al. Conservativetreatment versus mastectomy in breast cancertumors with macroscopic diameter of 20 mil-limeters or less. Cancer 1984;53(5):1209–13.

7. Fisher B, Anderson S, Redmond CK, et al.Reanalysis and results after 12 years of follow-up in a randomized clinical trial comparingtotal mastectomy with lumpectomy with orwithout irradiation in the treatment of breastcancer. N Engl J Med 1995;333(22):1456–61.

8. Veronesi U, Banfi A, Del Vecchio M, et al. Com-parison of Halsted mastectomy with quadran-tectomy, axillary dissection, and radiotherapyin early breast cancer: long term results. Eur JCan Clin Oncol 1986;22(9):1085–9.

9. Van Dongen J, Bartelink H, Fentimen I, et al. Ran-domized clinical trial to assess the value ofbreast-conserving therapy in stage I and IIbreast cancer, EORTC 10801 trial. J Natl Can-cer Inst Monogr 1992;11:15–8.


10. Blichert-Toft M, Rose C, Anderson JA, et al. Dan-ish randomized trial comparing breast conser-vation therapy whit mastectomy: six years oflife table analysis. J Natl Cancer Inst Monogr1992;11:19–25.

11. Lichter A, Lippman M, Danforth D, et al. Mastec-tomy versus breast conserving therapy in thetreatment of Stage I and II carcinoma of thebreast: a randomized trial at the National Can-cer Institute. J Clin Oncol 1992;10:976–83.

12. Winchester DJ, Menck HR, Winchester DP. TheNational Cancer Data Base report on theresults of a large nonrandomized comparisonof breast preservation and modified radicalmastectomy. Cancer 1997; 80:162–7.

13. Early Breast Cancer Trialist’s Collaborative Group.Effects of radiotherapy and surgery in early breastcancer. N Engl J Med 1995;333:1444–55.

14. Fisher B, Bryant J, Wolmark N, et al. Effect ofpreoperative chemotherapy on the outcome ofwomen with operable breast cancer. J ClinOncol 1998;16:2672–85.

15. Fisher B, Brown A, Mamounas E, et al. Effect ofpreoperative chemotherapy on local-regionaldisease in women with operable breast cancer:findings from NSABP B-18. J Clin Oncol1997;15:2483–93.

16. Olivotto IA, Rose MA, Osteen RT, et al. Late cos-metic outcome after conservative surgery andradiotherapy: analysis of causes of cosmeticfailure. Int J Radiat Oncol Biol Phys 1989;1:747–53.

17. Yeatman TJ, Cantor AB, Smith TJ, et al. Tumorbiology of infiltrating lobular carcinoma.Implications for management. Annals of Surg1995;222(4):549–59.

18. Gage I, Schnitt SJ, Nixon AJ, et al. Pathologicmargin involvement and the risk of recurrencein patients treated with breast conservingsurgery. Cancer 1996;78:1921–8.

19. Smitt MC, Nowels KW, Zdeblick MJ, et al. Theimportance of the lumpectomy surgical marginstatus in long term results of breast conserva-tion. Cancer 1995;76:259–67.

20. Anscher M, Jones P, Prosnitz L, et al. Local fail-ure and margin status in early-stage breast car-cinoma treated with conservative surgery andradiation therapy. Ann Surg 1993;218:22–8.

21. Krecke KN, Gisvold JJ. Invasive lobular carcinomaof the breast: mammographic findings andextent of disease at diagnosis in 184 patients.AJR Am J Roentgenol 1993;161(5):957–60.

22. Le Gal M, Ollivier L, Asselain B, et al. Mammo-

graphic features of 455 invasive lobular carci-nomas. Radiology 1992;185(3):705–8.

23. Clare SE, Sener SF, Wilkens W, et al. Prognosticsignificance of occult lymph node metastasesin node-negative breast cancer. Ann SurgOncol 1997;4(6):447–51.

24. Weiss MC, Fowble BL, Solin LJ, et al. Outcomeof conservative therapy for invasive breast can-cer by histologic subtype. Int J Radiat Oncol,Biol, Phys 1992;23(5):941–7.

25. Poen JC, Tran L, Juillard G, et al. Conservationtherapy for invasive lobular carcinoma of thebreast. Cancer 1992;69(11):2789–95.

26. White JR, Gustafson GS, Wimbish K, et al. Con-servative surgery and radiation therapy forinfiltrating lobular carcinoma of the breast.The role of preoperative mammograms inguiding treatment. Cancer 1994;74(2):640–7.

27. du Toit RS, Locker AP, Ellis IO, et al. An evalua-tion of differences in prognosis, recurrencepatterns and receptor status between invasivelobular and invasive carcinomas of the breast.Eur J Surg Oncol 1991;17(3):251–7.

28. Silverstein MJ, Lewinsky BS, Waisman JR, et al.Infiltrating lobular carcinoma. Is it differentfrom infiltrating duct carcinoma? Cancer1994;73(6):1673–77.

29. Mate TP, Carter D, Fischer DB, et al. A clinicaland histopathologic analysis of the results ofconservation surgery and radiation therapy instage I and II breast carcinoma. Cancer 1986;58(9):1995–2002.

30. Winchester DJ, Chang HR, Graves TA, et al. Acomparative analysis of lobular carcinoma ofthe breast: presentation, treatment, and out-come. J Am Coll Surg 1998;186(4):416–22.

31. Fisher ER, Sass R, Fisher B, et al. Pathologic find-ings from the National Surgical AdjuvantBreast Project (protocol 6). II. Relation of localbreast recurrence to multicentricity. Cancer1986;57(9):1717–24.

32. Veronesi U, Volterrani F, Luini A, et al. Quadran-tectomy versus lumpectomy for small sizebreast cancer. Eur J Cancer 1990;26:671–3.

33. Veronesi U, Luini A, Del Vecchio M, et al. Radio-therapy after breast preserving surgery inwomen with localized cancer of the breast. NEngl J Med 1993;328:1587–91.

34. Liljegren G, Holmberg L, Adami HO, et al. Sectorresection with or without postoperative radio-therapy for stage 1 breast cancer. Five yearresults of a clinical trial. J Natl Cancer Inst1994;86:717–22.

150 BREAST CANCER

35. Holland R, Veling S, Mravunac M, et al. Histologicmultifocality of Tis, T1–2 breast carcinomas:implication for clinical trials of breast conserv-ing treatment. Cancer 1985;56:979–90.

36. Schuster RH, Kuske Rb, Young VL, Fineberg B.Breast reconstruction in women treated withradiation therapy for breast cancer: cosmesis,complications, and tumor control. Plast Recon-str Surg 1992;90:445–52.

37. Hartmann LC, Schaid DJ, Woods JE, et al. Effi-cacy of bilateral prophylactic mastectomy inwomen with a family history of breast cancer.N Engl J Med 1999;340:77–84.

38. Fisher B, Joseph P, Constantino D, et al. Tamox-ifen for prevention of breast cancer: report ofthe National Surgical Adjuvant Breast andBowel Project P-1 study. J Natl Cancer Inst1998;90:1371–88.

39. Morton DL, Wen DR, Wong JH, et al. Technicaldetails of intraoperative lymphatic mapping forearly stage melanoma. Arch Surg 1992;127:392–9.

40. Krag DN, Weaver DL, Alex JC, Fairbank JT. Sur-gical resection and radiolocalization of the sen-tinel lymph node in breast cancer using agamma probe. Surg Oncol 1993;2:335–9.

41. Giuliano AE, Kirgan DM, Guenther JM, MortonDL. Lymphatic mapping and sentinel lym-phadenectomy for breast cancer. Ann Surg1994;3:391–8.

42. Albertini JJ, Lyman GH, Cox C, et al. Lymphaticmapping and sentinel node biopsy in thepatient with breast cancer. JAMA, 1996;276:1818–22.

43. Veronesi U, Paganelli, Galimberti V, et al. Sen-tinel-node biopsy to avoid axillary dissectionin breast cancer with clinically negative lymph-nodes. Lancet 1997;349:1864–7.

44. Krag D, Weaver D, Ashikaga T, et al. The sentinelnode in breast cancer: a multicenter validationstudy. N Engl J Med 1998;339:941–6.

45. Guiliano AE, Jones RC, Brennan M, Statman R.Sentinel lymphadenctomy in breast cancer. JClin Oncol 1997;15:(6):2345–50.

46. Winchester DJ, Sener SF, Winchester DP, et al.Sentinel lymphadenectomy for breast cancer:Experience with 180 consecutive patients: effi-cacy of filtered technetium 99m sulphur col-loid with overnight migration time. J Am CollSurg 1999;188:597–603.

47. Ciatto S, Cecchini S, Iossa A, Grazzini G. “T” cat-

egory and operable breast cancer prognosis.Tumori 1989;75:18–22.

48. Fisher B, Redmond C, Fisher ER, et al. Ten-yearresults of a randomized clinical trial comparingradical mastectomy and total mastectomy withor without radiation. N Engl J Med 1985;312:674–81.

49. Rosen P, Martin M, Kinne D, et al. Discontinuousor “skip”: metastases in breast carcinoma:analysis of 1228 axillary dissections. Ann Surg1983;197:276.

50. Veronesi U, Rilke F, Luini A, et al. Distribution ofaxillary node metastases by level of invasion.Cancer 1987;59:682–7.

51. Boova R, Bonanni R, Rosato F. Patterns of axil-lary nodal involvement in breast cancer. AnnSurg 1982;196:642–4.

52. Mathiesen O, Carl J, Bonderup O, Panduro J. Axil-lary sampling and the risk of erroneous stagingof breast cancer. An analysis of 960 consecutivepatients. Acta Oncol 1990;29:721–5.

53. Larson D, Weinstein M, Goldberg I, et al. Edema ofthe arm as a function of the extent of axillarysurgery in patients with Stage 1 and 2 carcinomaof the breast treated with primary radiotherapy.Int J Radiat Oncol Biol Phys 1986;12:1575–82.

54. Tate JJ, Lewis V, Archer T, et al. Ultrasound detec-tion of axillary lymph node metastases in breastcancer. Eur J Surg Oncol 1989;15:139–41.

55. March DE, Wechsler RJ, Kurtz AB, et al. Ct-pathologic correlation of axillary lymph nodesin breast cancer. J Comput Assist Tomogr1991;15:440–4.

56. Kao Ch, Wang SJ, Yeh SH. Technetium-99m MIBIuptake in breast carcinoma and axillary lymphnode metastases. Clin Nucl Med 1994;19:898–900.

57. Avril N, Dose J, Janicke F, et al. Assessment ofaxillary lymph node involvement in breast can-cer patients with positron emission tomog-raphy using radiolabeled 2–(fluorine-18)-flu-oro-2–deoxy-D-glucose. J Natl Cancer Inst1996;88:1204–9.

58. International (Ludwig) Breast Cancer StudyGroup. Prognostic importance of occult axil-lary lymph node metastases from breast can-cers. Lancet 1990;335:1565–8.

59. Bedwinek J, Fineberg B, Lee J, et al. Analysis offailures following local treatment of isolatedlocal-regional recurrence of breast cancer. Int JRadiat Oncol Biol Phys 1981;7:581–5.


153

Integration of systemic chemotherapy and/orhormonal therapy with surgery and irradiationis considered the standard of care in the treat-ment of locally advanced breast cancer(LABC). Because the greatest risk for patientswith LABC is the development of distantmetastases and subsequent death, the goals ofsurgery are to provide maximal locoregionalcontrol with minimal disfigurement and to per-mit accurate staging to determine prognosis.Breast conservation surgery is sometimes pos-sible after tumor downstaging with inductionchemotherapy, but close cooperation betweenthe medical and surgical oncologists and theradiation therapist is required to determine thefeasibility of this option. Similarly, the surgeonmust be familiar with the natural history ofLABC to assess the advisability of major resec-tions of either persistent advanced primary disease or locoregional recurrences. If lifeexpectancy is very short, as is the case withpatients who have bulky visceral disease ormetastases nonrespondent to multiple chemo-therapy regimens, the true benefit of a complexbut technically feasible operation should beevaluated carefully. However, in selectedpatients with advanced disease, surgery mayachieve quality palliation of local symptoms ofpain, hemorrhage, and malodorous ulceration.

This chapter defines LABC and addressesthe role of surgery after tumor downstagingwith induction chemotherapy, the use of mas-tectomy for inflammatory breast cancer, thefeasibility of immediate reconstruction in

10

Locally Advanced Breast CancerS. EVA SINGLETARY, MD, FACS

selected patients with LABC, and recent inno-vations in systemic therapy.

DEFINITION OF LOCALLY ADVANCEDBREAST CANCER

Locally advanced breast cancer generally refersto large primary tumors (> 5 cm) associatedwith skin or chest-wall involvement or withfixed (matted) axillary lymph nodes (T3/T4;N2/N3).1 In the most recent TNM staging sys-tem,1 tumors associated with disease in the ipsi-lateral supraclavicular nodal basin have beeneliminated from the LABC category becausethe supraclavicular basin lies outside the pri-mary lymphatic drainage pathways of the axillaand internal mammary nodes; tumors associ-ated with supraclavicular disease have beenreclassified as stage IV disease. However, aspatients with distant metastases confined tosupraclavicular nodes have a better prognosisthan patients with metastases at other distantsites and can be rendered disease free withlocoregional therapy,2 metastases limited to theipsilateral sub- or supraclavicular fossa will beincluded in the definition of LABC offeredhere. Large primary tumors (> 5 cm) with noevidence of nodal involvement (T3;N0) have amore favorable prognosis than LABC, with a 5-year survival rate of 70 to 80 percent; thus, inthe most recent TNM staging system, T3N0lesions have been reclassified as stage IIB dis-ease. However, as most series have classifiedT3N0 lesions as LABC for the purposes of

treatment, these tumors will also be included inthe present definition of LABC.

ROLE OF SURGERY AFTER INDUCTION CHEMOTHERAPY

Since the mid-1970s, patients with LABCtreated at The University of Texas M. D. Ander-son Cancer Center have received three to fourcycles of doxorubicin-based combinationchemotherapy prior to local therapy; local ther-apy is followed by the completion of systemictherapy and irradiation. Between 1974 and1996, patients with LABC were treated in fourtrials addressing four major concerns about theuse of induction chemotherapy: (1) whethertumor progression will occur during inductionchemotherapy, rendering the tumor unre-sectable even with radical surgery; (2) whetheroperative morbidity is increased after inductionchemotherapy; (3) whether the histologic stag-ing information obtained from the surgicalspecimen after induction chemotherapy main-tains its prognostic correlations with survival;(4) and whether breast conservation therapywith or without an axillary node dissection isfeasible and safe in patients with LABC.

In the first clinical trial at M. D. AndersonCancer Center (1974 to 1985), induction combi-nation chemotherapy was administered to 174evaluable patients (191 registered) with nonin-flammatory stage III breast cancer.3 After threecycles of 5-fluorouracil, doxorubicin, andcyclophosphamide (FAC), patients with anexcellent tumor response underwent irradiationof the chest wall and regional lymph nodes.Patients with a substantial volume of residualtumor underwent mastectomy and irradiation.After completion of locoregional therapy, FACwas reinitiated and continued until a dose of 450mg/m2 of doxorubicin was reached. Then treat-ment with cyclophosphamide, methotrexate,and 5-fluorouracil (CMF) was instituted andcontinued for a total treatment period of 2 years.

After the three cycles of induction chemo-therapy with FAC, 17 percent of patients had a

complete response (no evidence of tumor byphysical or radiographic examination). Seventy-one percent had a partial response (³ 50 percenttumor shrinkage). Only 10 percent had a minoror no significant response to induction chemo-therapy. Tumor progression occurred in 2 per-cent of patients. This trial demonstrated that themajority of patients will have significant tumorshrinkage with induction chemotherapy andthat the likelihood of tumor progression is low.As the virulence of a tumor is associated withchemoresistance, tumors that progress duringaggressive chemotherapy are unlikely to becontrolled with surgery, and a crossoverchemotherapy regimen should be considered.This study also confirmed that inductionchemotherapy is well tolerated and that surgicalprocedures after induction chemotherapy canbe completed without an increased rate ofinfection or delayed wound healing.4

The above trial refuted the concept that histo-logic staging information obtained after induc-tion chemotherapy would not have predictivepower. The histologically confirmed response inthe mastectomy specimen after inductionchemotherapy was an excellent prognostic fac-tor for survival and was more accurate thanclinical assessment of response.5,6 The numberof positive axillary nodes after inductionchemotherapy also remained prognostic forsurvival: actuarial 5-year survival rates were 70percent for patients with negative lymph nodes,62 percent for patients with one to three posi-tive lymph nodes, 47 percent for patients withfour to ten positive lymph nodes, and 21 per-cent for patients with more than ten positivelymph nodes.6 The 5-year disease-free survivalrates were 72, 46, 35, and 6 percent, respec-tively. When the subsets of patients with four ormore positive lymph nodes were combined, theoverall survival rate at 5 years was 38 percent,and the disease-free survival rate dropped toonly 20 percent. As patients with four or morepositive lymph nodes after induction chemo-therapy have a survival rate similar to thatobtained in historical trials of mastectomy and

154 BREAST CANCER

postoperative irradiation without systemic ther-apy,2,7 these patients should be considered forinnovative clinical trials.

The second M. D. Anderson clinical trial(1985 to 1989) was designed to determinewhether the extent of residual disease in themastectomy specimen after induction chemo-therapy can be used as a guide in planning post-operative adjuvant therapy. Three cycles of vin-cristine, doxorubicin, cyclophosphamide, andprednisone (VACP) were administered at 21-day intervals, then a modified radical mastec-tomy was performed. Patients with histologi-cally confirmed complete remission and thosewith < 1 cm3 of residual tumor received fiveadditional cycles of VACP; those with noresponse to induction chemotherapy werecrossed over to receive five cycles of methotrex-ate, 5-fluorouracil, and vinblastine (MFVb).Patients with partial response and ³ 1 cm3 ormore of residual tumor were randomly assignedto receive five additional cycles of either VACPor MFVb. All patients received radiation to thechest wall and regional lymph nodes. Eightpatients whose tumors remained inoperableafter induction chemotherapy underwent irradi-ation before mastectomy and MFVb. Irradiationhad a minimal effect on wound healing providedwound tension and thin skin flaps were avoided.If mastectomy resulted in a large defect, flapcoverage consisting of healthy autogenous tis-sue was preferred to the use of skin grafts.

Of 193 evaluable patients in this second trial(200 registered), 161 had a partial or greaterclinical response to the three cycles of induc-tion chemotherapy. Among the patients with apartial response, no statistically significant dif-ference (p = .64) was detected in the 4-year sur-vival rates for the MFVb and VACP groups (75and 58%, respectively).8 Of the 32 patients inthis study whose tumors showed a minor or noresponse to the induction chemotherapy, only16 remain alive and only 8 are disease-free atthe time of writing. The lack of impact on sur-vival of the crossover regimen in this study wasprobably due to the absence of an effective sec-

ond-line therapy. Significantly, there was exten-sive downstaging in a large proportion ofpatients in the study: 17 mastectomy specimenshad no evidence of residual tumor, and 54 mas-tectomy specimens had < 1 cm3 of residualtumor. This finding led us to consider the pos-sibility of performing breast conservationsurgery for locally advanced disease.

In a retrospective review of the mastectomyspecimens in which the tumor shrank by ³ 50percent with induction chemotherapy, the fac-tors most commonly associated with multiple-quadrant involvement that would excludebreast conservation surgery were demonstratedto be persistent skin edema, residual tumor size> 4 cm, extensive intramammary lymphaticinvasion, and known mammographic evidenceof multicentric disease.8

The objective in the third M. D. Andersonclinical trial (1989 to 1992) was to determineprospectively what fraction of patients withLABC may be candidates for breast conserva-tion surgery after induction chemotherapy.9 Of203 evaluable patients with LABC who com-pleted four cycles of induction chemotherapywith (FAC), 51 (25 %) elected and underwentbreast conservation surgery (Figure 10–1). Thebreast preservation rate for patients with ulcer-ative lesions or dermal lymphatic involvement(stage IIIB) was only 6 percent. With a medianfollow-up of > 60 months, only 5 (ten %) of the51 patients who underwent breast conservationsurgery had relapses in the breast.

In the fourth M. D. Anderson clinical trial(1992 to 1996), the objective was to determineif high-dose chemotherapy would increase theextent of tumor downstaging with inductionchemotherapy and allow more patients theoption of breast conservation surgery. One hun-dred and seventy patients with LABC were ran-domly assigned to receive either four cycles ofstandard FAC (1000 mg/m2 5-fluorouracil, 50mg/m2 doxorubicin, and 500 mg/m2 cyclophos-phamide) at 21-day intervals or dose-intensiveFAC (1200, 60, and 1000 mg/m2 of the threedrugs, respectively) at 18-day intervals with

Locally Advanced Breast Cancer 155

156 BREAST CANCER

prophylactic subcutaneous administration ofrecombinant human granulocyte colony-stimu-lating factor (G-CSF). After surgery, patientswith < 1 cm3 of residual tumor received fouradditional cycles of FAC or dose-intensiveFAC. Patients with a clinical partial responsebut with > 1 cm3 of residual tumor and thosewith four or more positive lymph nodes in thesurgical specimen were treated postoperativelywith four more cycles of FAC or dose-intensiveFAC followed by four cycles of methotrexateand vinblastine. Patients with no change or pro-gression of disease received six cycles ofmethotrexate and vinblastine. In all patients,locoregional radiotherapy was instituted within6 weeks of completion of chemotherapy.

One hundred and sixty-six patients wereevaluable for response. Patients who receivedFAC plus G-CSF were more likely to have acomplete or partial clinical response comparedwith patients who received standard FAC (84 v66%). However, the two regimens producedsimilar results in terms of histologic downstag-ing of the primary tumor. There was a completehistologic response (no tumor present) seen in25 percent of patients treated with FAC plus G-CSF and in 16 percent of patients treated withFAC alone (p = .155). There was a near-com-plete histologic response (< 1 cm3 of tumor pre-sent) seen in 25 percent of patients treated withFAC plus G-CSF and in 24 percent of patients

treated with FAC alone (p = .963). Although ahigher percentage of patients underwent breastconservation therapy in the group that receivedFAC plus G-CSF (42 v 29 %), this differencedid not achieve statistical significance.

To determine if there may be an alternativeto axillary node dissection after tumor down-staging, we analyzed 147 consecutive patientsin the FAC versus high-dose FAC study whohad both physical and ultrasound examinationsof the axilla at diagnosis and prior to surgery10

(Figure 10–2). Of the 133 patients with pal-pable axillary disease on initial examination, 43patients (32%) were downstaged to a negativeaxilla as assessed by physical and ultrasoundexamination following induction chemother-apy. There was a pathologic complete axillarylymph node response found in 30 patients(23%). Of the 72 patients with axillary metas-tases that were cytologically proven by fine-needle aspiration on initial evaluation, 15(21%) were confirmed to have histologicallynegative axillary lymph nodes following induc-tion chemotherapy. Of the 28 patients in whomthe axilla became clinically negative but thefindings on axillary ultrasound remained posi-tive after induction chemotherapy, 21 (75%)were found to have macroscopic axillary nodaldisease upon dissection. When both the physi-cal and ultrasound examination were negativefollowing induction chemotherapy, 53 percent

Figure 10–1. A, Patient with locally advanced breast cancer who desired breast conservation therapy. B, After four cycles ofinduction chemotherapy, a segmental mastectomy and axillary node dissection were performed. The patient then completedchemotherapy followed by irradiation of the breast and regional nodal basins.

BA


of patients (29 of 55) were found to still haveaxillary nodal metastases. However, 96 percent(25 of 26) had only 2 to 5 mm foci of disease.

On the basis of our analysis of the FAC ver-sus high-dose FAC study, we are currently con-ducting a clinical trial to assess whether patientswith a negative axilla by physical and ultra-sound examination can be safely treated withoutaxillary node dissection. In this clinical trial,patients with T2-3, N0-1 breast cancer are ini-tially randomized to receive four cycles of eitherpaclitaxel or standard FAC preoperatively.11

After the completion of induction chemother-apy, patients who have become candidates forbreast conservation surgery and who have clini-cally negative axilla are further randomlyassigned to either irradiation of the axilla or astandard level I and II axillary lymph node dis-section. After completion of four cycles of post-operative FAC, irradiation is delivered to thebreast and, in patients with a nondissectedaxilla, the lower axilla and the supraclavicular

fossa. Preliminary analysis based on 78 evalu-able patients (104 patients registered at the timeof the analysis) who had completed the induc-tion chemotherapy and surgery showed thatpaclitaxel and FAC have a similar ability todownstage both the primary tumor and the axil-lary nodal disease.11 No or minimal residual dis-ease was found in the breast in 41 percent of the41 patients on the FAC treatment arm, com-pared to 32 percent of the 37 patients whoreceived induction chemotherapy with pacli-taxel (p = .44). Sixty-nine patients underwent anaxillary node dissection. Negative or < 4 posi-tive lymph nodes were found in 68 percent ofthe 41 patients who received FAC chemother-apy, compared to 77 percent of the 37 patientstreated with paclitaxel (p = .75). In the eightpatients who did not undergo an axillary nodedissection, no axillary recurrences had beendetected in a 24-month follow-up period.

How safe is conservative surgery for LABCin terms of long-term local control? In review

Figure 10–2. A, Normal fat-replaced axillary lymph node. The sonogram shows a large node completely replaced by echogenicfat with a thin hypoechoic rim outlining the periphery of the node. B, Hypoechoic metastatic foci in an axillary lymph node confirmed by ultrasound-guided fine needle aspiration.

BA

of our database of all patients treated at M. D.Anderson with breast conservation therapy(patients with early-stage breast cancer andthose with LABC), 949 patients were found tohave been treated with breast conservationsurgery at our institution between 1982 and1994.12 Of this group, 93 patients receivedinduction chemotherapy prior to surgery on oroff protocol. The initial stage distribution ofthese 93 patients was as follows: stage IIA, 22.6percent; stage IIB, 24.7 percent; stage IIIA,32.3 percent; stage IIIB, 16.1 percent; and stageIV (supraclavicular lymph node metastasesonly), 4.3 percent. In most patients (88%),induction chemotherapy consisted of FAC orhigh-dose FAC for three to five cycles. Aftersegmental mastectomy and axillary node dis-section, patients underwent four to eight cyclesof chemotherapy followed by radiotherapy.Breast irradiation consisted of 50 Gy of exter-nal-beam radiation to the intact breast and a 10to 15 Gy boost to the segmental mastectomysite, which had been marked intraoperativelywith clips. Of the 93 patients, 86 completedpostoperative therapy. Two patients refusedradiotherapy after chemotherapy, four patientsrefused chemotherapy but did receive radio-therapy, and one patient refused all postopera-tive therapy.

Overall, nine patients had a local recurrence,for a local failure rate of 9.7 percent. In sixpatients, local recurrence was the first site ofrelapse; recurrence was in the breast paren-chyma in three patients, in the skin of the breastin two patients, and in the breast parenchymaand an axillary node in one patient. The mediantime to local recurrence in these six patients was55 months. Three patients had local recurrenceafter the development of distant metastases. Inthe nine patients with local recurrence, the sur-gical margin of the segmental mastectomy spec-imens was negative in all patients but close inthree patients. The six patients with local recur-rence only or local recurrence prior to distantmetastases had an overall survival rate of 83percent at a median follow-up of 88 months.

This survival rate is similar to the overall sur-vival rate of 89 percent for the entire group of93 patients (median follow-up of 73 months).The local recurrence rate in our selected seriesof breast conservation therapy for LABC wassimilar to the local failure rate observed forbreast conservation therapy in our patients withearly-stage breast cancer and was also consis-tent with the experience of other investigators(Table 10–1).13,15–20 The results of our study alsoindicate that most patients with local recurrencecan be treated without an adverse effect on over-all survival.

The role of axillary node dissection afterinduction chemotherapy in patients with LABChas become controversial. There are four mainarguments against the routine use of axillarynode dissection for LABC. First, inductionchemotherapy in patients with LABC and oper-able breast cancer has been shown to down-stage positive axillary lymph nodes to negativenodes in 23 to 44 percent of patients.6,10,14,20,21

Second, in most treatment protocols, patientswith LABC routinely receive additional post-operative chemotherapy and radiotherapyregardless of the findings at axillary node dis-section. Third, some LABC series havereported that axillary node dissection alone,axillary irradiation alone, or a combination ofsurgery and irradiation produce equivalent axil-lary control rates after induction chemother-apy.22–24 The fourth argument is somewhat morecomplex. There has been a survival advantagesuggested for high-dose chemotherapy overstandard anthracycline-based chemotherapy inpatients with multiple positive axillary nodesafter induction chemotherapy.25,26 However,high-dose chemotherapy off-protocol cannot berecommended in the absence of prospectiverandomized data demonstrating such a survivalbenefit. One of the lessons learned from thehigh-dose chemotherapy protocols for metasta-tic breast cancer was that patients with previ-ously demonstrated resistance to chemotherapyusually do not benefit from this procedure.27 Inaddition, proponents of axillary node dissection

158 BREAST CANCER

in patients receiving induction chemotherapyassert that the number of positive nodesdetected after tumor downstaging may affectwhether patients should be crossed over to a dif-ferent chemotherapeutic agent or be given high-dose chemotherapy. For example, phase II trialshave demonstrated high activity of taxane-basedchemotherapy (paclitaxel and docetaxel) inanthracycline-resistant breast cancer.28,29 How-ever, if the trend in therapy is toward a sequen-tial or combined approach using anthracyclineand taxane-based regimens prior to local ther-apy, with no further systemic interventionplanned, then the histologic assessment of theaxilla becomes only a prognostic tool.

Whether chemotherapy can substitute forsurgery for local control of occult axillarymetastases is still unknown. Data concerninglocoregional recurrences of the chest wall fol-lowing mastectomy show that optimal local con-trol is provided by using both systemic therapyand irradiation rather than chemotherapyalone.30 The use of axillary irradiation in patientswith clinically node-negative stage I or stage IIbreast cancer reduced the rate of axillary recur-rence by 1 and 3 percent, respectively.31–33 TheEarly Breast Cancer Trialists’ CollaborativeGroup overview analysis of randomized trialscomparing axillary surgical clearance versusradiotherapy found no difference in mortalitybetween groups regardless of the type of axillarytreatment.34 Sentinel lymph node biopsy inpatients with LABC has not been sufficientlystudied yet and will prove accurate only ifmetastatic deposits within each axillary lymphnode respond identically to chemotherapy.35

Based upon the clinical trials conductedthus far, we have learned that inductionchemotherapy can be given safely withoutincreasing the morbidity of local treatment. Thehistologic findings after induction chemother-apy remain important in defining prognosis.Very few patients will have progression of theirdisease and some will become candidates forbreast conservation therapy. Most importantly,neoadjuvant chemotherapy may identify


patients who will benefit from a cross-overchemotherapy regimen if tumor response isinadequate to the initial drugs.

INFLAMMATORY BREAST CANCER

Because today’s combination chemotherapyregimens can often render inflammatory breastcancer (IBC) resectable, mastectomy now has arole in the treatment of this disease. In ourreview of 178 women treated for IBC in dox-orubicin-based multimodality therapy proto-cols between 1974 and 1993, the addition ofmastectomy led to significant improvement inlocoregional disease control.36 Locoregionalrelapse rates were 16.3 percent (16 of 98patients) for patients who underwent chemo-therapy, mastectomy, and radiotherapy, and35.7 percent (15 of 42 patients) for patientswho underwent only chemotherapy and radio-therapy (p = .016). However, when patientswere stratified on the basis of tumor responseto induction chemotherapy, only patients with apartial response to chemotherapy demonstratedsignificant improvement in local control withthe addition of mastectomy. As only 12 percentof patients (21 of 178) had a complete clinicalresponse, demonstration of a statistically sig-nificant improvement in local control with the

Table 10–1. RATES OF BREAST CONSERVATION THERAPY AND SUBSEQUENT LOCAL

RECURRENCE AFTER INDUCTION CHEMOTHERAPY

Percent Percent LocalAuthor Patients Stage BCT Recurrence

Bonadonna14 157 II/III 81 1Calais15 158 II/III 49 8Veronesi16 226 II/III 90 6Schwartz17 160 II/III 34 2Touboul*18 97 II/IV 62 16Merajver*19 89 III 28 14Fisher20 747 I/II 68 8Peoples12 93 II/III/IV N/A 10

*Local therapy consisted of primary radiation therapy.BCT = breast conservation therapyAdapted from Hunt KK, Buzdar AU. Breast conservation after tumordownstaging with induction chemotherapy. In: Singletary SE, editor.Breast cancer—M. D. Anderson Solid Tumor Oncology Series.New York: Springer-Verlag; 1999. p. 196–207.

160 BREAST CANCER

use of mastectomy was not feasible in thisreview. The amount of residual disease foundon histologic examination of the mastectomyspecimen was highly predictive of long-termlocal control: no patient with < 1 cm3 residualdisease (n = 38) had a locoregional recurrence.

The effect of the addition of mastectomy ondisease-specific and disease-free survival wasalso dependent on the tumor response to induc-tion chemotherapy (Figure 10–3). Patients whohad a complete or partial clinical response toinduction chemotherapy and were treated withmastectomy in addition to chemotherapy andirradiation had significantly improved 5-yeardisease-specific survival compared with patientswho had a similar response to inductionchemotherapy but did not undergo mastectomy(62.0 v 43.0 %; p = .018). No improvement insurvival (disease-specific or disease-free) withthe addition of mastectomy was detected in

patients who had no significant response toinduction chemotherapy.

These retrospective data suggest that optimallocal control for most patients with IBC isobtained with the addition of mastectomy tochemotherapy and radiotherapy. Other benefitsof mastectomy are that it allows accurate assess-ment of the amount of residual disease afterinduction chemotherapy and that lower doses ofradiation can be used for subclinical disease.37

However, mastectomy should be used only selec-tively in patients who have no significantresponse to induction chemotherapy, as thesepatients are at high risk for both local and distantfailure regardless of surgical intervention.Although accelerated fractionation radiotherapyhas been proposed to exploit the biologic char-acteristics of IBC, an improvement in local con-trol rates with this technique has not been con-firmed.38,39 However, this accelerated schedule

Figure 10–3. Kaplan-Meier actuarial survival curves for disease-specific (A) and disease-free (B) survival in patients with acomplete or partial response to induction chemotherapy and for disease-specific (C) and disease-free (D) survival in patientswith no significant response to induction chemotherapy, stratified by type of treatment received. C = chemotherapy; M = mas-tectomy; RT = radiotherapy. Reprinted with permission from Fleming RYD, Armar L, Buzdar AU, et al. Effectiveness of mastec-tomy by response to induction chemotherapy for control in inflammatory breast carcinoma. Ann Surg Oncol 1997;4:452–61.

A B

C D

expedites delivery of the radiotherapy, whichmay be especially useful if the goal of therapy ispalliation. Interstitial irradiation has been stud-ied as a possible substitute for surgery in patientswho experience significant tumor reduction withchemotherapy, but early results in terms of localcontrol do not appear promising.40

Attempts to improve tumor downstaging inIBC with high-dose chemotherapy have shownpromise in terms of clinical response rates, butthe effect on long-term survival is still unclear.In a review of five trials of either single-agentor combination chemotherapy followed byautologous bone marrow transplantation(ABMT) for IBC and other stage III breast can-cers, Antman and colleagues41 reported that 44(79%) of 56 patients had a clinical completeresponse after induction chemotherapy butbefore ABMT, and that 89 percent had a com-plete response after ABMT. Disease-free statuswas maintained in 54 percent of patients, withfollow-up ranging from 1 to 37 months.

The use of a different, crossover chemother-apy regimen prior to mastectomy in patientswith IBC with less than a partial response toinduction chemotherapy may also be effective,as nonanthracyline-resistant drugs such aspaclitaxel and docetaxel are now available. Thecurrent trial at M. D. Anderson involves the useof paclitaxel if less than a partial response isobtained with four cycles of induction FAC. If acomplete response or partial response isachieved with paclitaxel, the patient undergoesmastectomy followed by four cycles of pacli-taxel and irradiation. In patients who have acomplete or partial response with the initial fourcycles of FAC, mastectomy is performed and anadditional four cycles of FAC are given, fol-lowed by four cycles of paclitaxel and irradia-tion. In patients with no significant response toeither FAC or paclitaxel induction chemo-therapy, the radiation oncologist and surgeonplan whether to treat the breast with preopera-tive irradiation and then perform mastectomy orto proceed with definitive irradiation as the onlylocal modality, with the intent of palliation.

Forty-three patients were entered in this IBCprotocol between 1994 and 1998. There was aclinical complete response observed in sevenpercent and a clinical partial response observedin 65 percent. This overall rate of response toinduction chemotherapy of 72 percent is iden-tical to the rate of response in the 178 patientswith IBC treated on FAC induction chemother-apy protocols between 1974 and 1993. How-ever, 2-year disease-free survival and overallsurvival were 61 and 78 percent, respectively,in the current protocol (median follow-up,20 months), compared to 52 and 71 percent,respectively, in our previous studies (medianfollow-up, 86 months). In our next protocol forIBC, patients will receive four cycles of induc-tion chemotherapy with FAC followed by twosix-week courses of paclitaxel (175 mg/m2 as a3-hour infusion weekly), with a two-week breakbetween the two courses. Patients with minimaltumor response after the FAC-paclitaxel induc-tion chemotherapy will be considered as candi-dates for high-dose chemotherapy with autolo-gous peripheral blood progenitor cell support.Surgical therapy will be planned at the comple-tion of all systemic therapy. In patients whohave a complete clinical response as docu-mented by physical examination, radiologicalimaging, and core needle biopsies, locoregionalirradiation concomitant with weekly paclitaxelwill be offered as an alternative to surgery.

RECONSTRUCTIVE SURGERY

The goal of reconstructive surgery for patientswith LABC can be to (1) repair defects, or (2)repair defects and re-create a breast mound. Inpatients with LABC who need or elect to havestandard mastectomy and who desire breastreconstruction to improve the cosmetic out-come, reconstruction is often delayed untilcompletion of both adjuvant chemotherapy andirradiation. As most locoregional recurrencesare in the skin or subcutaneous tissue of thechest wall,42 a flat postmastectomy chest walloften makes irradiation technically easier than


does a reconstructed breast mound, especially ifinclusion of the internal mammary nodal basinis necessary. However, in selected patients withexcellent response to induction chemotherapyor when palliative debulking surgical proce-dures are needed, the use of an autogenous flapto create a breast mound or provide skin cover-age of the operative defect before radiotherapyis instituted if feasible.

Implant-based reconstruction in an irradi-ated field has been associated with a high com-plication rate as well as patient discomfort anddissatisfaction because of loss of skin elasticityand fibrosis of underlying tissues after irradia-tion.43,44 The M. D. Anderson series of 298patients who received submuscular implantsrevealed that the rates of capsular contracture(Baker III or greater), pain, implant exposure,and implant removal were significantly higher(p = .028) in 13 patients with implants withinan irradiated field than in 230 patients withimplants who received no radiotherapy.45 Theeffects of irradiation were slightly less detri-mental in patients with implants placed beneathautogenous-tissue flaps: the complication ratewas 40 percent in 19 patients with implantsplaced in an irradiated area and 8 percent in 36patients with implants who had not undergoneradiotherapy.

The use of a myocutaneous flap for breastreconstruction, either before or after irradiation,does not interfere with the resumption ofchemotherapy or the ability to detect locore-gional recurrence.46 Irradiation of the recon-structed breast-mound flap does not impair theflap’s blood supply. In the M. D. Anderson seriesof 61 patients who required complex chest wallresections,47 prior irradiation that included theinternal mammary artery, which provides bloodto the rectus abdominis flap, or the thoracodor-sal artery, which provides blood to the latis-simus dorsi flap, did not compromise the viabil-ity of these flaps for wound coverage. Providedthat the flap has an adequate vascularizationwithout evidence of significant fat necrosis, theirradiation itself does not alter the cosmetic

result, except for the anticipated skin tanningand slight fibrosis of the reconstructed breastmound. In a series from M. D. Anderson48 of 19patients who received radiotherapy after recon-struction with an autogenous tissue flap, eitherfor known local recurrence (n = 4) or as adju-vant therapy for high risk of recurrence (n = 15),the cosmetic result was dependent on the initialoutcome of the reconstruction.

The two tissue flaps used most frequentlyfor reconstruction after breast surgery are thelatissimus dorsi and rectus abdominis myocuta-neous flaps. The advantages of the latissimusdorsi flap include its reliable blood supply andthe relative rarity of donor site morbidity. Thisflap is also relatively thin, so it matches thethickness of the native chest wall skin fairlyclosely and is excellent for providing coverageof soft-tissue defects (Figure 10–4). The chiefdisadvantage of the latissimus dorsi flap is itslimited size; an implant is usually required ifthe patient desires a reconstructed breastmound. The amount of available surplus skinvaries from patient to patient, but in general thelatissimus dorsi flap is never > 10 cm wide or20 cm long.

Rectus abdominis myocutaneous flaps canbe quite large and are most useful for defectstoo large to repair with a latissimus dorsi flap.The chief disadvantage is that they tend to bebulky and thus do not closely match the thick-ness of the native chest wall skin. The thicknessof this flap can be an advantage, however, if thedefect is located directly over the central area ofthe chest wall; in this case, the excess flap bulkcan be used to reconstruct a breast mound.

The two main types of rectus abdominis myo-cutaneous flaps are the transverse rectus abdo-minis myocutaneous (TRAM) flap and the verti-cal rectus abdominis myocutaneous (VRAM) flap.The TRAM flap has a greater arc of rotation anda more symmetrical and easily concealed donorsite than does the VRAM flap (Figure 10–5). TheVRAM flap leaves a more noticeable donor scarbut is technically easier to construct and has amore reliable blood supply (Figure 10–6). The

162 BREAST CANCER


Figure 10–4. A, B, Patient with locally advanced cancer of theleft breast who refused chemotherapy and radiation therapy.C, Operative defect after modified radical mastectomy.D, Closure of operative defect with a latissimus dorsi myo-cutaneous flap. E, Patient with excellent range of motion 10days after surgery. (Reprinted with permission from Single-tary SE. Breast surgery. In: Roh MS, Ames FC, editors. Atlasof advanced oncologic surgery. New York: Gower Medical Publishing, 1993. p. 14.1–14.9.)

A B

C

D

E

164 BREAST CANCER

TRAM flap is used most often when cosmeticconsiderations are important.

For major chest wall resections, the rectusabdominis flap is capable of covering a widearea from the clavicle to the costal margin andfrom the sternum to the midaxillary line.Because this flap is bulky, it provides sufficientchest wall stability even when up to five ribs orthe entire sternum is resected, without the needfor prosthetic mesh. However, if three or more

ribs have been removed, the use of mesh doesimprove chest wall mechanics and reduces theduration of ventilator dependence. Marlex, anonabsorbable durable mesh, can be used forflat surfaces of the chest wall. If the defect islarge, a “sandwich” of Marlex mesh and methylmethacrylate can be formed to restore a morenormal contour.49 If the mesh is covered bywell-vascularized tissue, the risk of infectionand extrusion is usually low.

Figure 10–5. A, Operative defect following full-thicknesschest-wall excision. B, Myocutaneous TRAM flap harvestedfrom the lower abdominal wall and transferred to the chest.C, Postoperative result with the chest wall reconstructionshaped into a facsimile of a breast. (Reprinted with permis-sion from Singletary SE, Hortobagyi GN, Kroll SS. Surgicaland medical management of local-regional treatment failuresin advanced primary breast cancer. Surg Oncol Clin N Am1995;4(4):671–84.)

A

B

C


Figure 10–6. A, Operative defect of the chest wall after radi-cal excision. B, Myocutaneous VRAM flap harvested from thelateral abdominal wall. C, Postoperative result. (Reprintedwith permission from Singletary SE, Hortobagyi GN, KrollSS. Surgical and medical management of local-regionaltreatment failures in advanced primary breast cancer. SurgOncol Clin N Am 1995;4(4):671–84.)

A

B

C

STRATEGIES FOR IMPROVING SYSTEMIC THERAPY

Advances in the treatment of LABC are largelydependent on improvements in systemic chemo-therapy. Two major strategies to improve sys-temic therapy include improved selection andindividualization of chemotherapy regimens andthe development of novel targeted therapies. Ifthe chemosensitivity of a specific breast cancercould be predicted before or soon after the initia-tion of chemotherapy, an optimal treatment regi-men could be designed for that tumor. One pos-sible way to predict chemosensitivity is tomeasure levels of cellular proteins associatedwith drug resistance, including MDR1 (mul-tidrug-resistance protein, or P-glycoprotein),MRP (multidrug resistance-associated protein),glutathione S-transferase, and dihydrofolatereductase.50 Although MDR1 is often notexpressed in early-stage breast cancer, it isdetectable at a high frequency in patients withLABC 51 and appears to correlate with a poorresponse to chemotherapy. However, the otherproteins associated with drug resistance areeither not detectable at a sufficient frequency orhave not yet been shown to be reliable enough topredict in vivo drug resistance. The earlier invitro chemosensitivity assays were problematicbecause the patient’s tumor cells had to be grownin culture for a prolonged period and the assayscould reliably predict only chemoresistance, notchemosensitivity. Newer chemosensitivity assaysthat preserve the cellular spatial relationshipsof the tumor and do not require a prolongedculture period are currently being assessed.52

Another approach to predicting chemosensi-tivity is to measure the effects of the chemother-apy on the intact tumor in vivo. This can be doneby studying sequential needle aspirates to deter-mine changes in flow cytometric DNA profilesand nuclear morphometric features that measurealterations in DNA content and cell cycle char-acteristics during chemotherapy; changes inthese features have been shown to correlate withsubsequent tumor regression.53,54 Alternatively,

results on positron emission tomography, whichreflect the metabolic alterations in the breastcancer following chemotherapy, may holdpromise as a predictor of response.55,56 Studiesare also under way to evaluate the role of mag-netic resonance imaging in accurately measur-ing true tumor response to chemotherapy.57

However, the optimal timing of these modalitiesin relation to the cycle of chemotherapy has notyet been determined.

An exciting new area is the identification ofseveral specific targets for novel therapeuticapproaches based on an understanding of themolecular genetic and biochemical features ofthe tumor. These therapeutic approaches mayinclude monoclonal antibodies either alone orconjugated to a cytotoxic substance; vaccines; orgene therapy to either suppress an oncogene orreplace the product of an inactivated tumor sup-pressor gene.58 These strategies may be com-bined with current chemotherapy regimens toproduce a synergistic effect. Antibodies againstthe HER-2/neu and epidermal growth factorreceptor oncogene products have been demon-strated to have a synergistic effect when com-bined with cisplatin, doxorubicin, and cyclo-phosphamide.59,60 Novel therapies may also beused to protect normal cells against the effects ofchemotherapy drugs and thus lessen side effects.For example, gene therapy may allow the humanmultidrug-resistance gene to be transfected intohuman marrow progenitors to instill a preferen-tial resistance to a chemotherapy drug such aspaclitaxel.61 Although the BRCA1 gene is rarelymutated in sporadic breast cancer, levels ofBRCA1 mRNA and its protein are decreased inboth hereditary and sporadic disease. Results ofa pilot trial with an ovarian cancer nude mousemodel indicate that delivery of a nonmutatedBRCA1 gene into the tumor via a retroviral vec-tor can suppress tumor growth.62 Encouragingobservations have also been reported for an E1Bgene-attenuated adenovirus, ONYX-015, thattargets the p53 gene of tumors but not of normalcells.63 The tumor-specific cytolysis produced bythis adenovirus appears to augment the efficacy

166 BREAST CANCER

of concomitant chemotherapy. Other therapeuticpossibilities include inhibitors of angiogene-sis64–66 and matrix metalloproteinases67; retinoidsto induce differentiation68; and vaccines directedagainst tumor antigens such as muc-1.69

Translational research that brings new treat-ment concepts from the laboratory to the clini-cal arena is essential for continued progress inthe management of LABC. Clinicians must beprepared to consider the feasibility of molecu-lar control of the underlying process of mam-mary carcinogenesis as part of their treatmentarmamentarium for both early-stage breast can-cer and LABC.

REFERENCES

1. Anonymous. Breast. In: Fleming ID, Cooper JS,Henson DE, et al, editors. AJCC cancer stagingmanual. 5th ed. Philadelphia: Lippincott-Raven; 1997. p.171–80.

2. Strom EA, McNeese MD, Fletcher GH, et al.Results of mastectomy and postoperative irra-diation in the management of locoregionallyadvanced carcinoma of the breast. Int J RadiatOncol Biol Phys 1991;21:319–23.

3. Hortobagyi GN, Ames FC, Buzdar AU, et al. Man-agement of stage III primary breast cancer withprimary chemotherapy, surgery, and radiationtherapy. Cancer 1988;62:2507–16.

4. Broadwater JR, Edwards MJ, Kuglen C. Mastec-tomy following preoperative chemotherapy.Ann Surg 1991;213:126–9.

5. Feldman LD, Hortobagyi GN, Buzdar AU, et al.Pathological assessment of response to induc-tion chemotherapy in breast cancer. CancerRes 1986;46:2578–81.

6. McCready DR, Hortobagyi GN, Kau SW, et al.The prognostic significance of lymph nodemetastases after preoperative chemotherapy forlocally advanced breast cancer. Arch Surg1989;124:21–5.

7. Hortobagyi GN, Singletary SE, Buzdar AU, et al.Primary chemotherapy for breast cancer: M. D.Anderson experience. In: Banzet P, editor. Pro-ceedings of the 3rd International Congress onNeoadjuvant Chemotherapy. New York:Springer-Verlag; 1991. p. 145–8.

8. Singletary SE, McNeese MD, Hortobagyi GN.Feasibility of breast conservation surgery afterinduction chemotherapy for locally advancedcarcinoma. Cancer 1992;69:2849–52.


9. Booser D, Frye D, Singletary S, et al. Response toinduction chemotherapy for breast cancer: aprospective multimodality treatment program.[abstract] Proc Am Soc Clin Oncol 1992;11:82.

10. Kuerer HM, Newman LA, Fornage BD, et al. Roleof axillary lymph node dissection after tumordownstaging with induction chemotherapy forlocally advanced breast cancer. Ann SurgOncol 1998;5(8):673–80.

11. Buzdar AU, Hortobagyi GN, Asmar L, et al.Prospective randomized trial of paclitaxelalone versus 5-fluorouracil/doxorubicin/cyclophosphamide as induction therapy inpatients with operable breast cancer. SeminOncol 1997;24:1–4.

12. Peoples GE, Regan Q, Heaton KM, et al. Breastconservation therapy for large primary andlocally advanced breast cancers after inductionchemotherapy. Ann Surg Oncol. In press.

13. Hunt KK, Buzdar AU. Breast conservation aftertumor downstaging with induction chemother-apy. In: Singletary SE, editor. Breast Cancer-M. D. Anderson Solid Tumor Oncology Series.New York: Springer-Verlag; 1999. p.196–207.

14. Bonadonna G, Veronesi U, Brambilla C, et al. Pri-mary chemotherapy to avoid mastectomy intumors with diameters of three centimeters ormore. J Natl Cancer Inst 1990;82:1539–45.

15. Calais G, Berger C, Descamps P, et al. Conserva-tive treatment feasibility with induction chemo-therapy, surgery, and radiotherapy for patientswith breast carcinoma larger than 3 cm. Cancer1994;74:1283–8.

16. Veronesi U, Bonadonna G, Zurrida S, et al. Con-servation surgery after primary chemotherapyin large carcinomas of the breast. Ann Surg1995;222:612–8.

17. Schwartz GF. Breast conservation followinginduction chemotherapy for locally advancedbreast cancer: a personal experience. TheBreast Journal 1996;2:78–82.

18. Touboul E, Buffat L, Lefranc J, et al. Possibility ofconservative local treatment after combinedchemotherapy and preoperative irradiation forlocally advanced noninflammatory breast can-cer. Int J Radiat Oncol Biol Phys 1996;34:1019–28.

19. Merajver SD, Weber BL, Cody R, et al. Breastconservation and prolonged chemotherapy forlocally advanced breast cancer: the Universityof Michigan experience. J Clin Oncol 1997;15:2873–81.

20. Fisher B, Bryant J, Wolmark N, et al. Effect ofpreoperative chemotherapy on the outcome of

women with operable breast cancer. J ClinOncol 1998;16:2672–85.

21. Schwartz GF, Birchansky CA, Komarnicky LT, etal. Induction chemotherapy followed by breastconservation for locally advanced carcinomaof the breast. Cancer 1994;73:362–9.

22. Scholl SM, Fourquet A, Asselain B, et al. Neoad-juvant versus adjuvant chemotherapy in pre-menopausal patients with tumors consideredtoo large for breast conserving surgery: pre-liminary results of a randomised trial: S6. EurJ Cancer 1994;5:645–52.

23. Delena M, Varini M, Zucali R, et al. Multimodaltreatment for locally advanced breast cancer.Cancer Clinical Trials 1981;4:229–36.

24. Perloff M, Lesnick GJ, Korzun A, et al. Combina-tion chemotherapy with mastectomy or radio-therapy for stage III breast carcinoma: Cancerand Leukemia Group B study. J Clin Oncol1988;6:261–9.

25. Rahman ZU, Frye DK, Buzdar AU, et al. Impact ofselection process on response rate and long-term survival of potential high-dose chemo-therapy candidates treated with standard-dosedoxorubicin-containing chemotherapy inpatients with metastatic breast cancer. J ClinOncol 1997;15:3171–7.

26. Hortobagyi GN, Buzdar AU, Champlin R, et al.Lack of efficacy of adjuvant high-dose (HD)tandem combination chemotherapy (CT) forhigh-risk primary breast cancer (HRPBC)—arandomized trial. [abstract] Proc Am Soc ClinOncol 1998;17: 123a.

27. Dunphy FR, Spitzer G, Buzdar AU, et al. Treat-ment of estrogen receptor negative or hormon-ally refractory breast cancer with double high-dose chemotherapy intensification and bonemarrow support. J Clin Oncol 1990;8:1207–16.

28. Holmes FA, Walters RS, Theriault RL, et al. PhaseII trial of Taxol, an active drug in the treatmentof metastatic breast cancer. J Natl Cancer Inst1991;83:1797–1805.

29. Valero V, Holmes FA, Walters RS, et al. Phase IItrial of docetaxel, a new highly effective anti-neoplastic agent in the management of patientswith anthracycline-resistant breast cancer. JClin Oncol 1995;13:2886–94.

30. Buzdar AU, McNeese MD, Hortobagyi GN, et al.Is chemotherapy effective in reducing the localfailure rate in patients with operable breastcancer? Cancer 1990;65:394–9.

31. Osborne MP, Ormiston N, Harmer OL, et al.Breast conservation in the treatment of earlybreast cancer: a 20-year follow-up. Cancer1984;53:349–55.

32. Delouche G, Bachelot F, Premont M, Kurts JM.Conservation treatment of early breast cancer:long-term results and complications. Int JRadiat Oncol Biol Phys 1987;13:29–34.

33. Wazer DE, Erban JK, Robert NJ, et al. Breast con-servation in elderly women for clinically nega-tive axillary lymph nodes without axillary dis-section. Cancer 1994;74:878–83.

34. Early Breast Cancer Trialists’ CollaborativeGroup. Effects of radiotherapy and surgery inearly breast cancer. N Engl J Med 1995;333:1444–55.

35. Singletary SE. Management of the axilla in earlystage breast cancer. In: Perry MC, editor.American Society of Clinical Oncology Edu-cational Book. Alexandria (VA): AmericanSociety of Clinical Oncology; 1998. p.132–41.

36. Fleming RYD, Asmar L, Buzdar AU, et al. Effec-tiveness of mastectomy by response to inductionchemotherapy for control in inflammatory breastcarcinoma. Ann Surg Oncol 1997;4:452–61.

37. Singletary SE, Ames FC, Buzdar AU. Manage-ment of inflammatory breast cancer. World JSurg 1994;18:87–92.

38. Barker JL, Montague ED, Peters LJ. Clinicalexperience with irradiation of inflammatorycarcinoma of the breast with and without elec-tive chemotherapy. Cancer 1981;45:625–29.

39. Thoms WW, McNeese MD, Fletcher GH, et al.Multimodal treatment for inflammatory breastcancer. Int J Radiat Oncol Biol Phys 1989;17:739–45.

40. Brun B, Ottmezguine Y, Feuilhade F, et al. Treat-ment of inflammatory breast cancer with com-bination chemotherapy and mastectomy versusbreast conservation. Cancer 1988;61:1096–103.

41. Antman K, Bearman SI, Davidson N, et al. Doseintensive therapy in breast cancer: current sta-tus. In: Gale RP, Champlin RE, editors. Newstrategies in bone marrow transplantation. NewYork: Alan R Liss; 1990. p. 423–6.

42. Newman LA, Kuerer HM, Hunt KK, et al. Pre-sentation, treatment and outcome of localrecurrence after skin-sparing mastectomy andimmediate breast reconstruction. Ann SurgOncol 1998;5(7):620–6.

43. Halpern J, McNeese MD, Kroll SS, Ellerbrock N.Irradiation of prosthetically augmented breasts:a retrospective study on toxicity and cosmeticresults. Int J Radiat Oncol Biol Phys 1990;18:189–91.

44. Forman DL, Chiu J, Restifo RJ, et al. Breastreconstruction in previously irradiated patientsusing tissue expanders and implants: a poten-

168 BREAST CANCER

apy. Ann NY Acad Sci 1993;698:423–8.57. Orel SG, Schnall MD, Powell CM, et al. Staging

of suspected breast cancer: effect of MR imag-ing and MR-guided biopsy. Radiology 1995;196:115–22.

58. Dhingra K, Hittelman WN, Hortobagyi GN.Genetic changes in breast cancer—conse-quences for therapy? Gene 1995;159:59–63.

59. Baselga J, Norton L, Masui H, et al. Antitumoreffects of doxorubicin in combination withanti-epidermal growth factor receptor mono-clonal antibody. J Natl Cancer Inst 1993;85:1327–33.

60. Hancock MC, Langton BC, Chan T, et al. Amonoclonal antibody against the c-erbB-2protein enhances the cytotoxicity of cis-diammine dichloroplatinum against humanbreast and ovarian tumor cell lines. Cancer Res1991;51:4575–80.

61. Ward M, Richardson C, Pioli P, et al. Transfer andexpression of the human multiple drug resis-tance gene in human CD34+ cells. Blood 1994;84:1408–14.

62. Holt JT, Thompson ME, Szabo C, et al. Growthretardation and tumor inhibition by BRCA1.Nat Genet 1996;12:298–302.

63. Heise C, Sampson-Johannes A, Williams A, et al.ONYX-015, an E1B gene-attenuated adeno-virus, causes tumor-specific cytolysis and anti-tumoral efficacy that can be augmented bystandard chemotherapeutic agents. Nature1997;3:639–45.

64. Thorpe PE, Burrows FJ. Antibody-directed target-ing of the vasculature of solid tumors. BreastCancer Res Treat 1995;36:237–51

65. O’Reilly MS, Holmgren L, Chen C, Folkman J.Angiostatin induces and sustains dormancy ofhuman primary tumors in mice. Nat Med 1996;2:689–92.

66. Folkman J. Clinical implications of angiogenesisresearch. N Engl J Med 1995;333:1757–63.

67. Sledge GW Jr, Qulali M, Goulet R, et al. Effect ofmatrix metalloproteinase inhibitor batimastat onbreast cancer regrowth and metastasis in athymicmice. J Natl Cancer Inst 1995;87:1546–50.

68. Lotan R. Retinoids in cancer chemoprevention.FASEB J 1996;10:1031–9.

69. Gilewski T, Adluri R, Zhang S, et al. Preliminaryresults: vaccination of breast cancer patientslacking identifiable disease with muc-1-key-hole limpet hemocyanin (klh) conjugate andqs21. [abstract] Proc Am Soc Clin Oncol 1996;15:555.


tially unfavorable result. Ann Plast Surg 1998;40:360–4.

45. Evans GRD, Schusterman MA, Kroll SS, et al.Reconstruction and the radiated breast: is therea role for implants? Plast Reconstr Surg 1995;96:1111–8.

46. Schusterman MA, Kroll SS, Miller MJ, et al. Thefree transverse rectus abdominis musculocuta-neous flap for breast reconstruction: one cen-ter’s experience with 211 consecutive cases.Ann Plast Surg 1994;32:234–42.

47. McKenna RJ, Moutain CF, McMurtrey MJ, et al.Current techniques for chest wall reconstruc-tion: expanded possibilities for treatment. AnnThorac Surg 1988;46:508–12.

48. Hunt KK, Baldwin BJ, Strom EA, et al. Feasibil-ity of postmastectomy radiation therapy afterTRAM flap breast reconstruction. Ann SurgOncol 1997;4:377–84.

49. Kroll SS, Walsh G, Ryan B, King RC. Risks andbenefits of using Marlex mesh in chest wallreconstruction. Ann Plast Surg 1993;31:303–6.

50. Verrele P, Meissonnier F, Fonck Y, et al. Clinicalrelevance of immunohistochemical detection ofmultidrug resistance. P-glycoprotein in breastcarcinoma. J Natl Cancer Inst 1991;83:111–6.

51. Ro J, Sahin A, Ro JY, et al. Immunohistochemicalanalysis of P-glycoprotein expression corre-lated with chemotherapy resistance in locallyadvanced breast cancer. Hum Pathol 1990;21:787–91.

52. Blackman KE, Fingert HJ, Fuller AF, Meitner PA.The fluorescent cytoprint assay in gynecologi-cal malignancies and breast cancer. Methodol-ogy and results. Contrib Gynecol Obstet 1994;19:53–63.

53. Brifford M, Spyratos F, Hacene K, et al. Evaluationof breast carcinoma chemosensitivity by flowcytometric DNA analysis and computer assistedimage analysis. Cytometry 1991;13:250–8.

54. O’Reilly SM, Camplejohn RS, Rubens RD. DNAflow cytometry and response to preoperativechemotherapy for primary breast cancer. Eur JCancer 1992;28:681–3.

55. Jansson T, Westlin JE, Ahlstrom H, et al. Positronemission tomography studies in patients withlocally advanced and/or metastatic breast can-cer: a method for early therapy evaluation. JClin Oncol 1995;13:1470–7.

56. Nieweg OE, Wong W-H, Singletary SE, et al.Positron emission tomography of glucosemetabolism in breast cancer: potential for tumordetection, staging, and evaluation of chemother-

171

In contrast to the 1960s when silicone implantswere the mainstay of breast reconstruction,patients in the 1990s may choose from animpressive spectrum of reconstructive options.Techniques, instruments, and materials haveevolved that provide all patients, regardless ofage, stage, previous treatment, or laterality,choices that may optimally represent theirdesires and expectations. Breast cancer aware-ness has increased the sophistication ofpatients, however, it remains the plastic sur-geon’s responsibility to educate patients andcoordinate expectations and outcome.

Increased detection of breast cancer has par-alleled improved techniques and availability ofscreening mammography, an increased femalepopulation, and the impact of changes in theage of childbearing, menarche, and menopause.Today, ductal carcinoma in situ (DCIS) repre-sents 15 to 20 percent of all breast cancercases;1 it is treated by either localized resectionor total mastectomy. Genetic testing and betterelucidation of risk factors has identified addi-tional patients as potential candidates for pro-phylactic mastectomy. As many as 15 percentof patients undergoing breast conservation, andwho require a proportionately large lumpec-tomy, attain poor esthetic outcome and may bebetter served in the longterm, by preoperativeconsideration of completion mastectomy andautologous reconstruction.2 Included in thisgroup are patients with small breasts, propor-tionately large lesions, and centrally locatedlesions. These women, often having been diag-

11

Breast ReconstructionGEOFFREY C. FENNER, MDTHOMAS A. MUSTOE, MD

nosed at an earlier age and stage, have excellentprognoses, and represent an increasing percent-age of patients seeking consultation and alter-natives for breast restoration.

The female breast is intimately associatedwith a woman’s selfesteem, sexuality, and inter-personal relations. The response to the impactand presumed implications of breast cancervaries widely among women. Breast cancerrepresents a therapeutic myriad with emotionaland physical implications, both for the presentand future. Although breast reconstruction maybe viewed as a positive alternative to breastloss, it represents only one facet newly diag-nosed cancer patients must face. Each patientupholds an individual, often rigid, esthetic stan-dard, emotional drive, and physiology whichguides them towards a specific reconstructivetechnique. It remains the plastic surgeon’sresponsibility to inform, educate, and performwith this in mind.

The first breast reconstruction was per-formed by Czerny, in 1895, when he success-fully transplanted a lipoma from a patient’s flankto a submammary position.3 Multiple develop-ments over the past 100 years have improvedreconstructive options as well as ultimate out-comes for women faced with mastectomy.

IMMEDIATE RECONSTRUCTION

Immediate reconstruction provides significantadvantages for the newly diagnosed breast can-cer patient (Table 11–1). Greater understanding

of tumor biology and technical advances inreconstructive surgery have led to greateracceptance of immediate postmastectomyreconstruction. The first large series reported in1982, conveyed excellent outcome, less expensethan delayed reconstruction, and no apparenteffect on the natural course of the malignancy.4

Initial options for immediate reconstructionthrough the mid-1990s included variousexpanders and implants, yet now include vari-ous flaps and even free-flap reconstruction. Sat-isfactory outcome is dependent upon patientselection as well as communication between theablative and reconstructive surgeons.

The advantages of immediate reconstruc-tion include diminished psychosocial trauma,superior esthetic results, decreased surgicalmorbidity, and lower cost than delayed recon-struction. Historically, delayed reconstructionswere more often performed due to heightenedfear of recurrence, concerns that immediatereconstruction would mask subsequent detec-tion of a recurrence, and the possibility thatimmediate reconstruction would be compro-mised by and hinder the initiation of adjuvanttherapy. It was also felt that patients would bemore appreciative of reconstruction if requiredto live for a time with the postmastectomydefect. These ideas have since been renderedobsolete by the need to consider the emotionalimpact of mastectomy and by the technical andtherapeutic advances of the past 15 years.

Patients undergoing immediate reconstruc-tion tend to incorporate the new breast intotheir body image, thereby maintaining greaterselfesteem, personal sexuality, and confidencein interpersonal relationships.5 They tend tohave less “cancer anxiety,” less recall, and

greater freedom in choosing clothing.6 Patientsundergoing mastectomy and immediate recon-struction demonstrate a similar psychosocialoutcome to that of breast conservation patients,having had lumpectomy with or without radia-tion.7 Body image may be adversely affecteddue to greater breast and donor site scarringcompared to patients having undergone breastconservation. Overall, psychologic morbidity issimilar, and clearly favorable compared to thatof patients having had delayed reconstruction.8

The opportunity to attain optimal estheticresults is enhanced with immediate reconstruc-tion. The newly raised mastectomy skin flapstend to preserve the shape of the natural breast,providing a structural template that determinesthe shape of the underlying volume, whether animplant or flap reconstruction. Skin flap fibro-sis associated with delayed reconstruction rep-resents inherent tissue loss and requires eithergreater tissue expansion or greater skin replace-ment at the time of autologous reconstruction.Fibrosis of the mastectomy skin flaps are animpediment in achieving a natural breast shape.Skin-sparing mastectomy in immediate recon-struction further increases the ability to attain asymmetric result, limits scarring to the periare-olar region, and minimizes the need for con-tralateral procedures such as reductions andmastopexies.9–13

Administration of adjuvant therapy is notdelayed in patients undergoing immediatebreast reconstruction, nor is the rate of compli-cations higher after immediate reconstruc-tion.14–15 The usual 3 to 4 week interval prior tothe initiation of adjuvant chemotherapy isample time for uncomplicated, postreconstruc-tive wound healing and patient recovery. Only1 to 2 percent of patients have their chemother-apy delayed beyond 3 to 4 weeks due to com-plications from immediate reconstruction, suchas delayed healing.16

Although neoadjuvant and adjuvant chemo-therapy have no relative impact upon immedi-ate reconstruction, adjuvant radiation is knownto unequivocally detract from the esthetic result

172 BREAST CANCER

Table 11–1. ADVANTAGES OF IMMEDIATE RECONSTRUCTION

Diminished psychologic traumaFacilitates coverage of radical defectsEases recipient pedicle dissectionSuperior esthetic resultsIncorporates skin sparing mastectomyMinimizes anemhesia and improves cost

and increase the local complication rate. This isinfluenced by many factors, including recon-structive technique and the type and dose ofradiation. Historically, radiation exaggeratedthe extent of fibrous capsular contracture pre-sent, to some extent, in all expander/implantreconstructions.17–19 Poor outcome paralleledthe need for substantial expansion and the useof large, smooth, silicone implants. The rate ofpoor cosmetic results in early series rangedfrom 18 to 40 percent, with a failure rate up to40 percent.20 Evans and colleagues reported a43 percent complication rate among radiatedimplant reconstruction patients, compared to a12 percent rate in nonradiated patients.18

Schuster reported a 55 percent complicationrate and unacceptable cosmesis in 24 percent ofpostreconstructive patients requiring adjuvantradiation.17 Of patients who had undergone acomposite autogeneous/implant reconstruction,40 percent of the radiated and 8.3 percent of thenonradiated patients had major complica-tions.18 Dickson reported an overall complica-tion rate of 70 percent for patients havingimmediate prosthetic reconstruction with radia-tion, and rates of 30 percent for skin necrosisand 67 percent for capsule contracture.21

Although the general consensus is to avoidprosthetic reconstruction in patients, an antici-pated need for adjuvant radiation, the regimentis most often recommended postoperatively.Use of textured saline prosthesis as well asimproved radiation techniques have demon-strated improved overall tolerance and dimin-ished complications in some early reports, butthere is no consensus.22–24

In contrast, tolerance of autologous tissue toradiation is generally good. Zimmerman reportedthe effect of postoperative radiation on immedi-ate free transverse rectus abdominis myocuta-neous (TRAM) reconstruction. He reported nototal or partial losses. Cosmesis, as rated bypatients, was excellent in 60 percent of cases,good in 30 percent, and fair in 10 percent.25

Although some variable degree of cutaneousfibrous contracture may occur, this can usually be

compensated for through surgical and designmodifications. It is interesting that the rate of fatnecrosis and volume loss in TRAM flaps, postra-diation, was higher in pedicled (33%) than in freeTRAMs (6%) reconstructions.26

Immediate postmastectomy reconstructionfor locally advanced disease has been reportedas encouraging. Sultan reported on 22 patientswith stage IIB or III disease who had under-gone neoadjuvant chemotherapy and comple-tion of chemotherapy 3 weeks subsequent tosurgery. Perioperative morbidity was 14 per-cent. Delay in resumption of chemotherapyoccurred in no instances, and patientsexpressed appreciation for having been offeredthis option.27 Styblo reported on 21 patientswith stage III disease who had undergoneimmediate TRAM reconstruction. There wereno delays in reinstitution of adjuvant treatmentand no increase in local relapse.28 It has beenshown that breast reconstruction may facilitateresection, without an increase in local compli-cations or relapse.

Immediate reconstruction also has eco-nomic advantages. Ablation and reconstructionare combined in one procedure, thereby limit-ing anesthetic risk and the time committed topostoperative recovery. Patients welcome theopportunity for a single procedure with lessimpact on occupational and domestic responsi-bilities. Avoidance of a staged second surgeryand hospitalization in delayed reconstructionhave obvious cost advantages.

SKIN-SPARING MASTECTOMY

Toth and Lappert first described skin-sparingmastectomy (SSM) in 1991.9 The technique isindicated for patients with early stage (I and II)breast cancer, patients managed with prophylac-tic mastectomy, and in attempts to facilitate ahighly esthetic outcome through maximal skinpreservation (Figure 11–1). Incisions are plannedthat will remove the breast, nipple-areolar com-plex, adjacent biopsy scars, and the skin overmore superficial tumors. Kroll and colleagues in

Breast Reconstruction 173

1991 reported only one local recurrence in 100cases with a follow-up of 23 months.10

Local recurrence is dependent upon tumorsize and locoregional nodal involvement.Despite variation in mastectomy technique,including SSM, the rate of local recurrence hasremained stable. Skin-sparing mastectomy ismore challenging for the oncologic surgeon,more time consuming, and requires delicatehandling of the skin flaps to avoid ischemiccomplications. These efforts to preserve theskin envelope and inframammary fold aregreatly appreciated by the patient and result ingreater symmetry, often diminishing the needfor a contralateral procedure. Subsequent areo-lar tatooing may completely camouflage thecentral incisions.

Newman reported a 6.2 percent local recur-rence rate in 372 patients who underwent SSMfor TI/II lesions. Ninety-six percent of theserecurrences presented as palpable skin flapmasses.11 Hidalgo reported on 28 patients whounderwent immediate reconstruction (92%receiving TRAM flaps) after SSM, with a meanfollow-up of 27 months. Complications at thereconstructive site were limited to cellulitis andmarginal periareolar skin loss. Esthetic resultswere judged as excellent in 75 percent ofpatients.12 Carlson compared 327 patients givenSSM to 188 non-SSM patients. After a meanfollow-up of 41 months, the local recurrence

rate was 4.8 percent in the former group and 9.5percent in the latter; native skin flap necrosisoccurred in 10.7 percent of the SSM patientsand in 11.2 percent of the non-SSM patients.13

Because local recurrence after SSM is low andthe likelihood of local control and survival arehigh, SSM with immediate reconstruction is anacceptable treatment for breast cancer.

BREAST IMPLANTS

The number of women with breast implantsranges between 1.5 and 2 million. The modernsilicone implant has been available since 1963and has undergone a multitude of subsequentmechanical and material improvements. Allimplants consist of a silicone elastomer shellthat may be single or double lumen, with asmooth or textured surface. Contents of singlechamber implants consist of either silicone gel,which is factory sealed and nonadjustable, orsaline, which may be adjusted intra- and/orperioperatively. Dual chamber implants weredevised to provide the benefits and camouflageof silicone texture (outer lumen), along withpostoperative saline adjustability (innerlumen). Various natural oils, triglycerides, andwater soluble hydrogels are currently underinvestigation but are not currently available inthe United States.

Silicone is ubiquitous in our environment.Individual exposure occurs through contactwith needles, syringes, medications (insulin,simethicone), lipstick, creams, cosmetics, andimplantable devices, such as pacemakers, jointreplacements, defibrillators, shunts, stents, andimplants.29 Extensive research undertakensince the FDA-directed silicone breast implantmoratorium in 1992 has confirmed thatimplantable medical grade silicone is amongthe least bioreactive, most inert substancesavailable for implantation.30–32 Studies havefailed to show linkage between connective tis-sue disease and silicone gel implants. The sili-cone elastomer shell and gel of breast implants,however, like all implanted devices, will trigger

174 BREAST CANCER

Figure 11–1. Skin-sparing mastectomy.

a foreign body inflammatory cell response,with giant cell formation and eventual scarring.The extent and impact of this fibrotic capsularresponse upon the fluid, and physical character-istics of breast implants is dependent upon cap-sular density, implant–tissue incorporation, thepresence of myofibroblasts, and/or the presenceof intracapsular silicone or sepsis.

Capsular contracture represents the mostcommon complication of breast implants. Itconsists of progressive fibrous constrictionaround breast implants and is unpredictable andvariable. It is graded according to a scale devel-oped by Baker (Table 11–2) and ranges fromvisually imperceptible (class I), to stone hardand painful (class IV). It may occur immedi-ately or years after implantation. There is agreater incidence associated with smooth sili-cone implants and with subglandular placementin cosmetic augmentation. Some theories sug-gest local contamination with Staphylococcusepidermidis as one inciting cause. The powderfrom gloves and inflammation from even lim-ited hematomas may play a role in some cases.

Capsular contracture may cause implantdeformation, migration, and rupture (Figure11–2) and may, on occasion, become calcifiedand detour from effective mammography. Indi-vidual perception is dependent upon severityand on the esthetic standard of the patient. Cap-sular contracture is not in itself a health risk.Twenty to 50 percent of reconstruction patientswho develop contractures require operativeintervention.

Contractures, historically, were releasedthrough aggressive manual compression. Thegoal was to “pop” the surrounding constrictingcapsule, leading to a softer breast. This tech-nique of closed capsulotomy resulted in occa-sional implant rupture, extracapsular siliconeextravasation, and surgeon injury (game-keeper’s thumb). In addition to long-term fail-ure, the technique could potentiate liability riskif future rupture was detected. Contracturestoday are more commonly corrected through alimited, outpatient, open capsulotomy, whereby

the fibrous capsule is surgically released and orexcised (capsulectomy) (Figure 11–3).

Silicone gel is composed of an amorphousmatrix consisting of silicone oils of varioussizes and weights. Smaller caliber oils areknown to diffuse through the elastomer shell(silicone gel “bleed”) and become incorporatedinto the fibrous capsule. Microscopic amountsmay percolate through lymphatic channels fol-lowing macrophage ingestion and migrate tothe regional lymph nodes. As with exposure toother medial grade silicones, there is no evi-dence to suggest that the minute quantitiestransgressing the elastomer shell have anymetabolic or long-term impact.

Silicone gel implant rupture occurs in up to63 percent of patients after 12 years, docu-mented during surgery in patients having theirimplants removed.33–36 Among asymptomaticpatients, the incidence of implant rupture isunknown but is believed to be significant.


Table 11–2. BAKER’S CLASSIFICATION OF CAPSULE CONTRACTURE

Class I Augmented breast feels as soft as an unoperated-upon breast

Class II Minimal; less soft, the implant can be palpated but is not visible

Class III Moderate; more firm, the implant can be easily palpated and is visible

Class IV Severe; the breast is hard, tender, painful, cold, and distorted

Reproduced with permission from Little G, Baker JL. Results ofclosed compression capsulotomy for treatment of contracted breastimplant capsules. Plast Reconstr Surg 1980;65:30.

Figure 11–2. Left breast class III capsule contracture.

Gradual attenuation of the elastomer shell, withimperceptable rupture, is well documented.Abrupt or premature rupture may be promptedby capsular contracture, implant shell infoldingleading to accelerated stress fractures, andtrauma. Once a gel implant shell ruptures, fromlongevity or trauma, the contents are usuallycontained within the surrounding fibrous cap-sule. This is likely to remain undetected and hasdemonstrated no systemic effects.

Post-traumatic change in the form of herni-ation, deflation, malposition, or deformationmay manifest extracapsular extravasation.When this occurs, free gel may infiltrate breastparenchyma and tissue planes, and/or elicit agranulomatous foreign body reaction. Thismay lead to regional silicone migration, sili-cone mastitis, and formation of irregular nod-ules that may, on physical examination andmammography, simulate a malignancy.37 Sus-pected implant rupture warrants evaluation.Magnetic resonance imaging has a greater sen-sitivity than do either mammography or ultra-sound and is the test of choice for detectingimplant rupture.38 Early removal of the free sil-icone and implant, with or without implantreplacement, will help to avoid these sequellaeand minimize subsequent confusion in mam-mographic screening.

Silicone and saline implants are radio-opaque on mammography and have led to con-cerns regarding potential delay in breast cancerdetection.39–40 Implant characteristics, whichmay affect the sensitivity of standard mammo-graphy, include implant size, the proportion ofoverlying breast tissue, implant placement(subglandular versus submuscular) and thepresence and immobility of capsular contrac-ture.41–42 As recommended by the AmericanCancer Society and the American Society ofPlastic and Reconstructive Surgeons, womenwith breast implants should maintain the sameschedule of mammography as all other women.They should secure a certified facility that hassufficient experience with breast implants andconfirm the availability of displacement mam-mography (Eklund) and ultrasound. Patientswith postmastectomy implant reconstructionare typically followed by physical examinationonly. One major epidemiologic study has con-firmed that the stage at breast cancer detectionin women with implants is identical or betterthan it is in the general population;43 a secondmajor study from the National Cancer Institutewill be addressing this question as well. Inaddition, there is no evidence that silicone iscarcinogenic in humans. In fact, in two largestudies women with implants exhibit 10 to 30percent less breast cancer than would be statis-tically expected when matched with the generalpopulation; the results, however, did not showstatistical significance.44–47 This issue needsfurther study with larger numbers of patients.The most recent large study, sponsored by theNCI (in press), shows an incidence no differentthan for the matched control group.

In 1992, a series of poorly documented casereports and the subsequent intense mediascrutiny, combined with a temporary suspen-sion of silicone gel implant usage by the FDA,led to lawsuits and an eventual multibillion dol-lar settlement with the major implant manufac-turers. Only one implant company (Mentor)was allowed to provide gel implants for recon-struction patients, with specific and rigid crite-

176 BREAST CANCER

Figure 11–3. Excised implant and enveloping capsule con-tracture.

ria on a highly monitored, investigational basis.There were a plethora of syndromes, autoim-mune diseases, and symptoms associated withsilicone breast implants, and intense litigationfollowed. Many of these proposed associations,such as rheumatoid arthritis, were, in fact,shown in subsequent, large retrospective stud-ies to occur in a lesser percentage of augmentedpatients than in the general population. Sclero-derma-like syndromes were not shown to beassociated with breast implants. The AmericanCollege of Rheumatology issued on October 22,1995, the following statement, based on accumu-lated data: “Studies provide compelling evidencethat silicone implants expose patients to nodemonstrable additional risk for connective tis-sue or rheumatologic disease.” None of the pos-tulated syndromes have withstood the scrutiny ofprospective epidemiologic testing.48–50 Resultsfrom a large National Cancer Institute studyare still pending.

The FDA has recently submitted its require-ments for submission of a “premarket approval”which will, once again, enable marketing of gelimplants. The protocol requires patient moni-toring during an 18-month follow-up, and sub-mission of a limited questionnaire.

PRIMARY IMPLANT RECONSTRUCTION

One-stage primary implant reconstruction, theworkhorse of breast reconstruction in the1980s, has become less frequently used due toimproved outcome with expander or autologousreconstruction. Certain patients with A- to B-sized breasts, having limited to no ptosis, suffi-cient soft-tissue coverage, and who desire anexpeditious and simplistic approach to breastrestoration, may remain candidates for eitherimmediate or delayed single-staged implantreconstruction. Even in this group, however, amore natural shape can be achieved by anexpander with a removable valve, that may alsoserve as a permanent implant.

Inherent to mastectomy are resection of thenipple-areolar complex, inclusion of adjacent

biopsy incisions, and a resultant, variable, ipsi-lateral skin deficiency. Immediate reconstruc-tion requires an initial assessment of skin flapvascularity, trauma, and tension. Only thehealthiest skin flaps should signify proceedingwith immediate implant reconstruction. Ques-tionable vascularity, or marginal necrosis, war-rants reappraisal and the choice of an alterna-tive option, such as an immediate expander orautologous flap reconstruction, or delayedreconstruction. Compromised flaps, and/orinsertion of a large implant under tension, risksdehiscence and implant exposure. Avoidance oftight compressive dressings and constrictingbras, prompt drainage of hematomas or sero-mas, and early revision of marginal necrosiswill minimize complications.

Implant position is determined by thedimension and esthetics of the contralateralbreast. The position of the inframammary fold,breast base width, volume, and overlying skinredundancy, or ptosis, are critical in attainingoptimal symmetry with the native breast. TheIMF may be lowered up to 2 cm when attempt-ing to simulate limited contralateral ptosis.Alternatively, a concurrent, or delayed con-tralateral reduction or mastopexy may maxi-mize esthetic outcome and symmetry.

Delayed implant reconstruction is a saferand more popular option. The well-healed skinflaps are elevated in the subpectoral plane andmay be stretched, thinned, and scored to pro-vide improved projection without regard to vas-cular compromise. The final outcome may besimilarly improved by a symmetry procedure.

One-stage implant reconstruction is anoption ideally suited to the rare patient withsmall to moderate sized nonptotic breast whopossesses sufficient soft tissue coverage andwho desires the simplest reconstructive option.Despite an initial desire to avoid a secondaryprocedure, many patients require future implantadjustments or symmetry procedures. This tech-nique has been largely supplanted by adjustableand permanent expanders/implants and the pop-ular, time-tested, two-stage expander technique.


ADJUSTABLE IMPLANT RECONSTRUCTION

Adjustable implants or permanent expanders/implants represent an option intermediate to thesingle-stage implant reconstruction and themore conventional two-stage technique (Figure11–4). Postoperatively, adjustable implantsenable precision in symmetry and the ability toattain a softer, often larger reconstruction withgreater ptosis. The technique offers protectionagainst tension-related wound complicationsand is generally considered preferable to pri-mary implant reconstruction. It offers an excel-lent alternative to patients with limited skindeficits, A- to B-sized contralateral breasts,and/or those patients who require only limitedexpansion. In addition, a second-stage implantexchange is often avoided.

Indications for use of an adjustable implantinclude patients with an immediate or delayedsoft tissue deficit, with mild to moderate ptosis,or who require salvage after failed primaryimplant reconstruction. Patients with asymmet-ric deformities from hypoplasia, trauma, burns,or congenital deformities (including pectusexcavatum and Poland’s syndrome) are alsoideal candidates. Adjustability is beneficial inaugmentation candidates with inherent paren-chymal asymmetries or tuberous breasts, or inpatients with an unpredictable or poorly com-municated esthetic standard.

Poor candidates for implant reconstructionare those with large, pendulous breasts. Thesewomen, often obese, represent a challenge withany technique and are unlikely to achieve satis-factory symmetry without a contralateral reduc-tion or mastopexy. Prior radiation treatment is astrong relative contraindication. The fibroticand relatively ischemic nature of radiated skinflaps resists expansion and tolerates an under-lying implant poorly, with a tendency towardscutaneous erosion and exposure. These patientsare better served with either an autologous orcomposite reconstruction.

There are two types of adjustable prosthesescurrently available. One is a round and anatomic,textured or smooth, postoperatively adjustable,saline implant. The implants are successivelyexpanded, with saline, by percutaneous injec-tion through a remotely positioned subcuta-neous injection port. Alternatively, Beckerexpander/prostheses are composed of a dualchamber system. The inner lumen, like theMentor implant, is filled and expanded withsaline through a self-sealing, removable injec-tion port. The outer lumen is factory sealedwith silicone gel. It provides patients with thetangible advantages and surface camouflage ofsilicone and the postoperative adjustability of asaline implant.

Once optimal size and shape have beenattained, as confirmed by patient and surgeon,and sufficient time for capsule maturation hasbeen allowed (4 to 6 months), the ports may beremoved under local anesthesia, usually througha short segment of the lateral mastectomy inci-sion. Vigorous retraction of the connecting tubeengages a self-sealing valve and prevents leak-age of intrinsic saline. Removal of a small vol-ume of saline prior to port removal may optimizeimplant softness and simulate ptosis.

Preoperative considerations include accu-rate assessment of size and base diameter.These determinations may be aided by the useof templates, sizers, and by the weight of themastectomy specimen. Preoperative markingsshould detail breast margins in the immediate

178 BREAST CANCER

Figure 11–4. Bilateral reconstruction with adjustable expander/prosthesis.

reconstruction setting or, in delayed reconstruc-tion, mirror the contralateral breast.

Total muscular coverage, in the immediatereconstruction setting, will reduce the inci-dence of implant exposure, infection, and cuta-neous complications. Sufficient coverage andimplant camouflage is provided by a submus-cular pocket composed of the pectoralis, serra-tus, and rectus muscles. The deflated implant isplaced, precisely, within the muscular pocketand cleared of redundant folds. The injectionport is connected, drawn through the lateral ser-ratus fibers, and fixed to the lateral chest wall.Patency of the filling system should be con-firmed. Skin flap viability must be assessedprior to closure and all questionable skinresected. The implants may be expanded toobliterate dead space but should not furtherstress the muscular or cutaneous closure.

Expansion is usually initiated 7 to 14 dayspostoperatively, following confirmation of skinflap viability. The frequency and extent of eachexpansion is dependent upon wound healing,skin sufficiency, and the patient’s tolerance andcomfort level. Typically, saline is injected to thepoint of tolerable skin tension, without blanch-ing, on a weekly basis. Maintenance of theimplant at maximum volume for a minimum of3 months allows for capsule maturation. Theimplant may then be adjusted, within a narrowrange, prior to port removal, to optimize con-sistency and ptosis.

IMMEDIATE TWO-STAGE BREASTRECONSTRUCTION

Tissue expansion in breast reconstruction waspioneered through the efforts of ChadomerRadovan and initially reported in 1976.51 Thepostmastectomy defect lacks both skin for cov-erage and the underlying breast mound. To bereconstructed, the skin envelope must have aadequate laxity to allow the breast mound toproject sufficiently, achieve symmetric ptosis,and remain soft in consistency. These goalsoften require the recruitment of substantial adja-

cent skin through temporary overexpansion.Expanders, presently available for immediatebreast reconstruction, enable focused expansionand simulation of a realistic inframammary fold,without the physiologic, donor site, and rehabil-itative demands of autologous reconstruction.Second-stage exchange with either saline or sil-icone implants is a simple, outpatient proce-dure. Tissue expanders remain the most popularmethod of immediate breast reconstruction.

The advent of textured surfaces has some-what lessened the incidence of capsule contrac-ture, has limited the incidence of perioperativemigration, and has resulted in more predictable,successful results. The introduction of anatomicexpanders and implants has enabled preferen-tial expansion of the lower pole, to better simu-late natural ptosis. Over longer follow-up, how-ever, there are still significant limitations in theability to consistently achieve a natural shapeand soft breast.

Although all patients who undergo a mastec-tomy may be considered candidates for expanderreconstruction, preferred patients are those withsmaller, minimally ptotic breasts. Some of thesepatients may exhibit sufficiently vigorous skinflaps to accommodate a primary implant recon-struction. Those with limited deficits, and par-ticularly those who increasingly undergo skin-sparing mastectomies, are candidates for the useof newer adjustable implants. The two-stagedapproach is a reliable, predictable reconstructionand has the ability to incorporate maximal adja-cent skin, achieve greater volumes and ptosis,and enable patient-directed modifications. Finalrefinements, including fold adjustments andcapsulotomy, are facilitated at the time ofimplant exchange, optimizing the esthetic result.Conversely, patients with large or pendulousbreasts require greater, more prolonged expan-sion and often a contralateral symmetry proce-dure to achieve an acceptable result.

Prior or anticipated chest-wall radiationafter breast conservation or mastectomy remainsa strong relative contraindication to immediateexpander reconstruction. Cutaneous radiation


fibrosis resists effective expansion, limits ulti-mate projection, and increases the risk of cap-sule contracture, skin flap necrosis, implantexposure, and infection. Patients with compro-mised wound healing ability, such as those withscleroderma and lupus, may also benefit fromalternative methods.

Textured, anatomic expanders with inte-grated ports are preferable. Smooth-surfacedexpanders have been shown to result in anunacceptable rate of capsule contracture, com-promising effective expansion.52 Anatomicexpanders enable preferential expansion of thelower pole, a more realistic shape, and greaterprojection. Use of an integrated port affordsgreater patient comfort, in that the overlyingmastectomy skin flaps are usually anesthetic.The integrated ports, when compared to remoteports, also incur a lower rate of malfunction.

Simulating contralateral base width and theheight of maximal projection are the key ele-ments in attaining optimal cosmetic outcome.The goal is to accomplish the major surgicalsteps at the first procedure, which requirescareful analysis of the contralateral breast. Sim-ulation of ptosis can be accomplished throughoverexpansion of the skin envelope and subse-quent deflation or secondary replacement withan implant of lower vertical profile. Moderateptosis can be simulated through the use of over-expansion and anatomic expanders, and bylowering of the inframammary fold. Moderateto severe ptosis can not be accurately matchedand usually necessitates either a contralateralreduction or mastopexy or composite recon-struction using the latissimus dorsi myocuta-neous flap. Contralateral procedures are usuallymore precise when based upon the quality andextent of expansion achieved and are thereforepreferentially performed during the secondstage. The end point of the expansion processoccurs when adequate projection is achieved inrelation to the contralateral breast, rather thanthe ultimate volume being attained.

Patients and their reconstructions are notadversely affected if concurrent expansion

adjuvant chemotherapy is imposed, pendingwound stability at initiation. When imple-mented, completion of a chemotherapeutic reg-imen and granulocyte recovery is usuallyrequired prior to the second stage.

If adjuvant radiation is deemed necessarysubsequent to expander placement, full, prera-diation expansion, with preferably a 15 to 20percent overcompensation, helps resist thefibrotic contracture associated with radiation.All patients undergoing postreconstructionradiation, however, risk radiation-inducedimplant complications. Radiation-induced cap-sule contracture affects the quality of theexpansion, often leads to local chest wall dis-comfort, and potentiates the risk of expanderextrusion or exposure. Patients should be mon-itored throughout their course and the expanderincrementally deflated if skin flap compromiseis noted. Treatment options for patients withradiation-induced complications include expand-er removal and delayed reconstruction, salvageby autologous replacement (TRAM or latis-simus), and, occasionally, delayed capsulotomy.Patients with large, smooth implants seem toshow the worst response.

Complications of expander reconstructionparallel those of primary and adjustable implantreconstruction53–57 (Table 11–3). Advantages andtheir ranges are illustrated in Table 11–1. Cap-sular contracture remains the single most trou-blesome complication and is reported in 10 to25 percent of patients.58–59 Progressive contrac-ture may lead to asymmetry, deformation, andpain, and require intervention, such as capsulo-tomy, in 20 percent of cases.

The advantages of prosthetic breast recon-struction include the ability to attain a reason-ably good esthetic result without a complex,prolonged operation and hospital stay. Compro-mising a donor site, with its potential compli-cating morbidity, is also avoided. Prostheticreconstruction remains appealing for bilateralcases in which symmetry is less of a problemand where bilateral autologous reconstructionwould impose substantial demands on the

180 BREAST CANCER

patient and surgeon. Similarly, in patients withsmaller breasts, in older patients, and in thoseless motivated, expander or implant reconstruc-tion remains a desirable option.

ADVANTAGES OF AUTOLOGOUSRECONSTRUCTION

Breast reconstruction with a silicone- or saline-based implant is technically the simplest optionavailable to mastectomy patients. Recentadvances enhancing the potential esthetic out-come include permanent expander prosthesisand postoperatively adjustable implants. Mostcompetent surgeons can insert a prosthesis,postmastectomy, in a wide range of patients.Consequently, prosthetic reconstruction is themost common mode of breast reconstructionavailable today.

Prosthetic reconstruction is safe and expedi-tious, with a limited recovery period. It is suitedto the patient desiring a simple approach towardbreast restoration. Candidates include thosewho wish to avoid an external prosthesis, thosewith limited expectations, those with smallerbreasts and limited-to-no ptosis, those withexisting medical risk factors and anxiouspatients who have difficulty comprehendingmore technical procedures, and those desiringan expeditious initiation of adjuvant treatment.Expander/implant reconstruction may alsopacify younger patients who wish to ultimatelyconvert to autologous reconstruction followinganticipated pregnancies.

Implant-based breast reconstruction, how-ever, has many disadvantages. The implant,which is clad only by a thin layer of skin andmuscle, is often poorly camouflaged and leadsto a round, “mechanical,” unnaturally aptoticand asymmetric replacement. Peri-implant cap-sule contractures may impose further distor-tion, migration, asymmetry, and discomfort.Capsular fibrosis limits the fluidity of bothsaline and silicone implants. It is noticeableupon palpation and in its inability to react nat-urally to positional changes. This is especially

apparent when lying supine, when the recon-structed breast remains fixed and projectingwhile the native breast falls naturally to theside. This represents the most common adversepostoperative development, occurring in 20 to40 percent of all mastectomy patients and requir-ing operative intervention in up to 20 percent ofcases.58–59 Implant-based reconstruction may,therefore, be a less strategic option for youngerpatients. Kroll reported on 325 postmastectomypatients who had undergone either expanderor autologous reconstruction. Complicationsoccurred in 23 percent of expander patients,compared to 9 and 3 percent in latissimus andTRAM flap reconstructions, respectively.60

Implants are devices and are susceptible todevice failure. It has been well demonstratedthat the silicone elastomer shell of both sili-cone and saline implants fatigue over time.This may manifest itself as either a silicone orsaline bleed and/or leak. An intracapsular sili-cone implant rupture is likely to remain unde-tected until some adverse event occurs. Mostcommonly, this may involve an increased ten-dency toward progressive capsular contracture.Blunt trauma resulting from a car, bicycle, orrollerblade incident, or even an overzealousmammogram, may convert a contained rup-ture into an extracapsular rupture. Patientstypically notice a change in the shape and/orvolume of the implant. This scenario warrantseither mammographic, ultrasonic, or MRIimaging to rule out rupture.38 Conversely, rup-ture of saline implants leads to implant defla-tion and a flat breast. In either case, implantreplacement is warranted.


Table 11–3. IMPLANT COMPLICATIONS

No XRT (%) XRT (%)

Implant loss/extrusion 3.4–18 4–10Deflation 3–4Infection 1.2–8 10Capsule contracture 2.9–31 20Skin necrosis 10–24 3–7Satisfaction 80–98 49–55

XRT = external radiation beam therapy

In contrast, autologous tissue has thewarmth, consistency, feel, and reactive mobilityof one’s own tissues. It is a malleable, con-formable, permanent medium that does notelicit a foreign body fibrotic response and ismore tolerant of adjuvant therapy, trauma, andinfection (Table 11–4). In contrast to the greatercontracture and rupture rates of implants, autol-ogous tissue softens and ages commensuratewith adjacent structures and is therefore anideal option for younger patients. An autolo-gous flap may be contoured to match a con-tralateral breast of almost any size and shape.Although the initial overall cost of the flapreconstruction is greater, the long-term costs ofautologous reconstruction have been shown tobe less than those of prosthetic reconstructiondue to subsequent secondary capsulotomies,revisions, and implant exchanges required withthe latter procedure.

Autologous reconstruction is inherentlymore complex from both a technical and anartistic standpoint. The functional and estheticoutcome of the initial procedure, which lastsfrom 4 to 5 hours, largely depends upon the sur-geon’s experience and/or microsurgical exper-tise. Although the initial procedure requires alonger hospitalization (3 to 4 days) and postop-erative recovery, the result is permanent andrarely requires a secondary adjunctive proce-dure. The TRAM flap is overwhelmingly theflap of choice when available. Alternativesinclude the latissimus dorsi, Ruben’s or peri-iliac, lateral thigh, and gluteal flaps.

CONVENTIONAL TRANSVERSE RECTUS ABDOMINIS

MYOCUTANEOUS FLAP

The transverse rectus abdominis myocutaneousflap, one of the most ingenious techniques inplastic surgery, has established itself over time asthe flap of choice for autogenous breast recon-struction. It presents the reconstructive surgeonwith the opportunity to a create a breast ofunsurpassed esthetic beauty, is unparalleled in itsability to simulate the opposite breast, and sec-ondarily improves the contour of the lowerabdomen. Attaining consistently good resultsrequires careful planning and technical profi-ciency. The lower abdomen consistently providesexceptional and sufficient tissue for unilateraland, in the majority of patients, bilateral breastreconstruction. The procedure is versatile andreliable when performed within its recognizedvascular and volumetric constraints. Hartrampf’slandmark introduction of the TRAM flap in1982, still the gold standard for autologousbreast reconstruction, provided the foundationfor the modern era of breast reconstruction.61

The conventional, unipedicled TRAM flap,as originally described, consists of a transverseellipse of skin and fat based on one rectusabdominis muscle and its intrinsic musculocu-taneous perforators from the superior deep epi-gastric pedicle. The pedicle branches as it trans-gresses through the substance of the ipsilateralrectus through a network of “choke” vessels,which reconstitute in the midabdomen.61–65

This inflow communicates with the periumbili-cal, myocutaneous perforators that supply thesuprafacial and subcutaneous plexuses. Bost-wick has determined that blood flow in the con-ventional TRAM is based upon pedicle caliber,number of perforators, integrity of thesuprafascial plexus across the midline, andvenous outflow.64–65 Perfusion has been gradedand is depicted as a sequence of zones, withzone VI, the most distal tissue, representingstrictly random perfusion (Figure 11–5). Flowin the conventional TRAM is, therefore, sec-

182 BREAST CANCER

Table 11–4. ADVANTAGES OF AUTOLOGOUS RECONSTRUCTION

SoftWarmPliablePermanentEnables wide resectionNo foreign body responseNatural consistency and appearanceTolerates adjuvant therapy wellDecreases need for symmetry procedureMore economic in the longterm

ondary and unpredictable beyond the midline.Patient selection is critical and is limited,among experienced surgeons, to those patientswho have tissue requirements met by the ipsi-lateral “hemi-TRAM.”

In the uncomplicated case, the flap extendsfrom the umbilicus to a point superior to thepubis. The incisions are beveled, after isolationof the umbilicus, to incorporate additional peri-umbilical perforators and subcutaneous fat. Theflap is elevated at the suprafascial level towardthe medial and lateral row of ipsilateral muscu-locutaneous perforators. The fascia is incised,immediately adjacent to the perforators, facili-tating subsequence closure, and the underlyingrectus is mobilized.

Most commonly, a full width muscle harvestwith a thin strip of fascia is performed. The rec-tus muscle is elevated beneath the superiorabdominal skin flap to the costal margin. Thesuperior epigastric pedicle is easily identified,enabling transection of the lateral rectus fibersas well as of the intercostal nerves. This facili-tates muscle atrophy and, thereby, minimizesthe central xiphoid bulge, common initiallyafter this procedure. The flap is transposedthrough a subcutaneous tunnel, which under-mines the medial IMF, and is inset into thebreast defect. Zones IV and II may be discardedprior to transposition to facilitate passage.

In an effort to preserve abdominal wallintegrity, an alternative “split-muscle” harvesthas been advocated.66–68 Pedicle (muscle)width is based upon the laterality of the medialand lateral row of perforators. It is usually pos-sible to preserve a substantial (one-third) widthof the lateral rectus and often a slip of infraum-bilical medial rectus. Although the muscle is,in most cases, denervated, it is thought touphold the muscular interface of the semilunarline and adds fibrous stability in the perioper-ative period.

It is common practice to include a segmentof skeletonized inferior epigastric pedicle in theevent additional perfusion is necessary to sus-tain the flap.69–70 This “lifeboat” enables sup-

plementary perfusion through a microvascularanastomosis, if intrinsic vascular insufficiencyis noted. The flap is “supercharged” through ananastomosis, most commonly to an axillaryrecipient pedicle.

Fascial donor site closure is achieved witheither interrupted figure-of-eight sutures or arunning, heavy, braided, synthetic. The patientis then flexed to 45° to facilitate abdominal clo-sure and ascertain breast symmetry. Typically,two drains are placed, both at the breast andabdominal sites. Postoperative flap monitoringis institution-specific and may encompass tem-perature probes, ultrasound or laser doppler,and clinical surveillance.

Optimal perioperative conditions are para-mount to early and late flap success. Patientcore temperature, intravascular fluid status,anxiety and pain level, and position may allhave an impact on final outcome. If the starttime is late in the day or there is minimal urine


Figure 11–5. TRAM flap zones.

upon foley placement, an initial fluid bolus maybe required. An intraoperative hourly urine out-put of 50cc should be maintained to ensureadequate flap perfusion. In addition, mainte-nance of normothermia may minimize vascularvasoconstriction, spasm, and shivering, all ofwhich may have an adverse impact on immedi-ate postoperative flap perfusion. The use ofheating blankets and fluid warmers is routine.

Use of the superiorly based unipedicleTRAM flap requires strict adherence to patientselection criteria. Obtaining consistent resultsdemands an assessment of potential risk fac-tors. It has been clearly demonstrated thatpatients who smoke, are obese, have significantabdominal scarring, or have had previous radi-ation have an increased risk of complications,including fat necrosis, partial flap failure, anddonor site complications.71–75 These risk factorsshould not eliminate patients from the proce-dure so much as indicate modification toenhance blood supply to the transferred tissue.For instance, nicotine from cigarette smokinghas been recognized as a potent vasoconstrictorof the microcirculation. Patients who smoke orare unable to abstain 4 to 6 weeks prior tosurgery are at extremely high risk for partialflap and donor site necrosis.76 These patients,likely to fail conventional reconstruction, oftensucceed under the preface of a delay, bipedicle,or free technique.

Alternatively, the midabdominal TRAMwas devised in response to a 20 to 60 percent ofpartial flap loss or fat necrosis and a high rateof hernia and abdominal wall weakness in high-risk patients.77–78 This flap is based on the per-forator-rich periumbilical region, and extendsinferiorly to a tangent parallel to the anteriorsuperior iliac spine (ASIS). Because muscleintegrity is preserved below the arcuate lines, alower incidence of hernia may be anticipated.Slavin’s review of 236 midabdominal flapsshowed a 2 percent rate of partial flap necrosisand a 1 percent incidence of fat necrosis.77 Theprimary disadvantage is an occasionally dis-pleasing high- or midabdominal scar and the

lack of the abdominoplasty effect inherent inconventional TRAM.

Although largely supplanted by microsurgi-cal advances, preoperative surgical delay of aconventional TRAM is another technique foraugmenting reliable flap dimensions.79–82 In1995, Codner demonstrated improved inflowand diminished congestion after surgicaldelay.82 Zones II and III proved more vigorousand reliable, especially in the high-risk patient,and lessened the need for a bipedicled approach.In 1997, Restifo documented greater pediclecaliber (1.3 versus 1.8 mm) and flow rates (7.5versus 18.2 mL/min) compared to controls aftersurgical delay.83 Staged interruption of the infe-rior epigastric pedicle, on a physiologic basis, ishighly effective in augmenting vascularity andis beneficial in the high-risk patient.

Scars within the confines of the harvestedflap, such as appendectomy, hernia, and mid-line scars, represent an ischemic boundary andwill diminish the volume tissue available, dueto variable vascular disruption. Regional inci-sions, such as paramedian, cholecystectomy, orextended Pfannenstiel’s incisions may directlydisrupt the pedicle and potentiate donor sitecomplications. Shaw and colleagues assessedcomplications among TRAM patients with pre-existing scars.75 In 43 percent of the patients,abdominal wall weakness, partial flap loss, fatnecrosis, and donor site morbidity developed.Paramedian scars precluded use of the freeTRAM in three of three patients. Cholecystec-tomy scars and multiple scars showed the high-est propensity toward skin-related complica-tions. Conservatism, and often an alternativeflap design, are warranted in patients with pre-existing scars.

Operative assessment of the contralateralbreast helps in formulating a reconstructivestrategy and optimizing symmetry. The locationand breadth of the IMF is a critical landmarkand serves as the basis for building a symmet-ric breast. Attention to the condition and vol-ume of retained skin (mastectomy skin flap),the size, shape, base width, and ptosis of the

184 BREAST CANCER

contralateral breast and how it relates to theIMF is necessary if symmetry is to be opti-mized. Patients with pre-existing macromastiamay elect to undergo concurrent or delayedcontralateral breast reduction, both to alleviateobjective symptoms (shoulder pain and groov-ing, intertrigo, lower back pain) and improvethe ultimate esthetic outcome. Patients withsubstantial glandular ptosis may elect toundergo mastopexy for similar reasons. It is thecurrent authors’ preference to perform thesecontralateral procedures as a second stage.Improved accuracy of a symmetry proceduremay be attained after resolution of flap edema,muscle atrophy, and skin retraction. Stagingalso enables concurrent refinements (SAL,IMF revision) on the recently restored breast.

The goal of reconstructive surgeons usingthe TRAM flap for breast reconstruction is toincrease the total efficiency, reduce operativemorbidity, and to be able to offer patients anabsolute minimum complication rate and hos-pital stay. Experience has demonstrated thevascular and volumetric constraints of thepedicled TRAM flap and led to technicalrefinements that are dependent upon individ-ual patient risk factors.

FREE TRANSVERSE RECTUS ABDOMINIS MYOCUTANEOUS FLAP

The free TRAM flap, based on the dominantinferior epigastric blood supply and requiring amicrovascular anastomosis, represents a reli-able, versatile, highly esthetic option for bothimmediate and delayed reconstruction. Therationale for employing this option stems fromits benefits in immediate reconstruction, wherethe ultimate goal of the reconstructive surgeonis to provide a highly desirable restoration,minimal complications, and an expeditiousrecovery that will remain invisible to the adju-vant therapeutic sequence (see Figure 11–5).Complications associated with conventionalTRAM reconstruction, occurring in up to 25percent of reported cases, are partial flap loss

and fat necrosis, and are inherent to the pro-cedure’s secondary blood supply and volumeconstraints.57,66,73–76,84–86 These complicationsmay impose prolonged wound healing and con-siderable delay in the therapeutic sequence.Although the free TRAM procedure requiresgreater technical proficiency and a slightlylonger operating time, the flap has unparalleledvascular reliability and versatility, and is theflap of choice in high-risk patients. Theseinclude obese patients, smokers, and thosepatients with prohibitive scars or who have hadprior radiation treatment.73–76

Suprafascial elevation is identical to thatemployed in the pedicled TRAM procedure.Widely dispersed perforators may be omitteddue to the dominant inflow, thereby limiting thefascial harvest. The lateral edge of the rectusmuscle is elevated to discern the path of theepigastric pedicle. This determines whether amedial and/or lateral muscular strip may bepreserved. The pedicle is ligated at the externaliliac origin. Axillary recipient vessels are nor-mally reliable, even if prior radiation therapyhas been implemented. Preference, in decreas-ing order, include the thoracodorsal, the cir-cumflex scapular, lateral thoracic, and internalmammary vessels. This last option requires aperisternal costochondrectomy at the third ribinterface for adequate exposure. An interruptedor running arterial microvascular anastomosisis typically performed with 9.0 nylon suture.Venous anastomosis may be similarly per-formed, or one may use an anastomotic cou-pling device (3M) for added speed. The occlud-ing clamps are removed, and the quality of flapperfusion is confirmed both clinically and withthe use of an intraoperative doppler.

Advantages of the free TRAM flap (Figures11–6A, 11–6B) include a more limited abdomi-nal harvest site, which corresponds to a moreexpeditious, often less uncomfortable postoper-ative recovery. The volume of rectus muscleharvested may be limited to a small cuff offibers surrounding the perforators transgressingthrough the rectus muscles to supply the


suprafascial and subcutaneous plexuses. Thus, amedial and/or lateral strip of rectus muscle maybe preserved, which benefits young, activepatients or those desiring future pregnancies.Use of the free TRAM enables preservation ofan inframammary fold, does not compromisethe marginal perfusion of the freshly elevatedmastectomy skin flaps, optimizing the estheticoutcome. Freedom upon insetting, due to theabsence of the conventional muscular “leash,”facilitates a quick, easy, and highly cosmeticand symmetric reconstruction (Figures 11–7A,11–7B). The improved blood supply may expe-dite wound healing and initiation of adjuvanttherapy, which may also be better tolerated thanin the conventional TRAM flap procedure.

The only absolute contraindications to freeTRAM reconstruction include prohibitive scar-ring, violation of the inferior epigastric bloodsupply from previous abdominoplasty, suctionlipectomy, extended Pfannenstiel’s incision, orprevious TRAM procedure. Pre-existing med-ical conditions may limit the patient’s ability totolerate 4 to 6 hours combined anesthesia time.This should be addressed preoperatively.

In a series of 211 free TRAM flaps, Schus-terman reported a flap thrombosis rate of 3.3percent and a flap loss rate of 1.4 percent.87

One study compared outcome among conven-tional and free TRAM reconstruction. It wasdemonstrated that despite a higher percentage

of high-risk patients (63 versus 28%), the freeTRAM group had fever complications (9 ver-sus 28%) than the conventional TRAM group.The advantages of the free TRAM procedureare outlined in Table 11–5.

The main disadvantage of TRAM flap recon-struction is the potential for weakening theabdominal wall. Questions remain as to the besttechnique of abdominal closure and the impactof free versus pedicled flap reconstruction onthe abdominal wall. Despite all the advantagesof the TRAM, it is a major surgical procedureand carries the risk of abdominal weakness,bulging, and hernia formation. True herniasresulting from the procedure are extremely rare(< 3% of cases). Abdominal wall bulges, indi-cating a separation and attenuation of the inter-nal and external oblique muscles, occur morefrequently (3 to 12% of cases).

Several studies have obtained objective mea-sures of abdominal muscle strength. Trunk mus-cle strength as measured by an isoknetic dyna-mometer demonstrated postoperative recoveryof 92, 96, and 98 percent at 3, 6, and 12 months,respectively, for unilateral free TRAM flappatients. Although the ability of one-half of thegroup to perform situps was not affected, theother half demonstrated mild impairment.88

Kind and colleagues compared the recoveryafter pedicled and free TRAM reconstruction.Flexion torque as measured by dynamometer

186 BREAST CANCER

Figure 11–6. A, Free TRAM reconstruction; B, 9 months postoperative.

A B

was 58 and 89 percent at 6 weeks and 6 months,respectively, for conventional TRAMs. In con-trast, the free TRAM showed 89 and 93 percentof preoperative flexion recovery at 6 weeks and6 months respectively. The investigators con-cluded that the pedicled TRAM caused a signif-icantly greater insult to the abdominal wall in theearly postoperative period but that the two tech-niques equilibrated to over 90 percent of preop-erative levels at 12 months. It was also deter-mined that muscle splitting techniques appearedto offer no functional advantage.89

The ability of patients to perform situps aftervarious modes of TRAM reconstruction hasbeen studied. Percentages of patients able toperform a situp were 63.0, 57.1, 46, and 27 per-cent for the single free TRAM, conventionalTRAM, bilateral free flap, and conventionalflap, respectively.90 Fitoussi reported that 47percent of single pedicle TRAM and 0 percentof bipedicle TRAM patients could performsitups postoperatively and concluded thatalthough the hernia rate did not vary betweenthe two groups, functional sequellae were statis-tically significant.91

Hernias occur in 5 to 6 percent of TRAMpatients, regardless of whether the procedurewas conventional or free, uni- or bipedicled.The incidence appears more closely associatedwith the technique and detail of abdominal clo-sure rather than with the number or extent ofmuscles harvested. Various modifications, out-

lined below, appear to have a positive impact.Approximation of the medial and lateral

remnants of the tendinous inscriptions appearsto “restore the ribs” of abdominal support andrelieve tension on the anterior rectus sheath.92

Primary repair of the fascial defect shouldinclude approximation of the underlying mus-cular support. Kroll advocates approximationof the anterior remnant of the internal obliqueto the linea alba and a reinforcing two-layerclosure. Fascial sheath closure is strengthenedwith sutures through the semilunar line.93

Abdominal wall complications are probablybest avoided through recognizing excessivetension during closure or undue attenuation ofweakened fascia. The threshold for mesh rein-forcement should correlate positively withthese findings. Mesh enables reinforcementwithout excess tension and may facilitate post-operative mobility, diminish pain, and expediterecovery. The infection rate after using mesh isreported to be < 2 percent and is usually a resultof other miscalculations that result in exposurethrough marginal dehiscence or necrosis.


Figure 11–7. A and B, Right free TRAM reconstruction prior to nipple/areolar reconstruction.

A B

Table 11–5. ADVANTAGES OF FREE TRAM

Primary and dominant blood supplyGreater available volumeLess muscle harvest/abdominal dissectionMore comfortable recoveryMore reliable in high-risk patientsGreater freedom in insettingGood tolerance to adjuvant therapy

The other complication that has led to con-siderable investigation and comparison betweenTRAM techniques is the incidence of fat necro-sis. Kroll and colleagues reported in 1998 on theincidence of fat necrosis among patients whohad had conventional versus free TRAM recon-struction. Of the 49 free TRAM patients, 8.2percent exhibited clinical fat necrosis, with onepatient showing mammographic evidence. Ofthe 67 pedicled TRAM patients, 27 percentdemonstrated fat necrosis on examination andnine patients on mammogram.94

BIPEDICLED TRANSVERSE RECTUSABDOMINIS MYOCUTANEOUS FLAP

A unipedicled conventional TRAM will reli-ably perfuse all of zone I, 20 percent of zone II,and 80 percent of zone III.95 An alternativetechnique or flap choice is warranted if tissuerequirements exceed these specifications. Indi-cations for a bipedicled TRAM include thosepatients who insist on autogenous reconstruc-tion, who require additional volume, and forwhom microsurgical reconstruction is not pos-sible due to an absence of reasonable recipientaxillary vessels. The indications parallel thosefor surgical vascular delay.

The lower abdominal pannus is isolated onthe medial and lateral row of perforators, bilater-ally. Once the upper abdominal apron is elevated,each superior epigastric pedicle is isolated withthe assistance of doppler mapping, and a splitbipedicle muscle harvest is performed. The flapis transposed and inset in much the same way asfor an unipedicled TRAM flap.

Multiple reports have investigated the long-and short-term impact of bilateral rectus har-vest. Hartrampf reported that 64 percent ofpatients could not perform a single situp afterbipedicled reconstruction, compared to 17 per-cent in the unipedicled group. Petit reported a20 percent incidence of subsequent severe backpain in bipedicled patients.96

The bipedicled flap has reduced the inci-dence of partial flap loss and fat necrosis in

much the same manner as the free technique.The use of mesh has markedly reduced the inci-dence of abdominal hernia formation andbulging. Although these patients have objectiveloss of abdominal function, subjective interfer-ence with daily activity is rare. There arereports of an increased incidence of long-termlower back pain.

Use of the bipedicled TRAM for unilateralreconstruction has invoked substantial contro-versy in the plastic surgery literature. Antago-nists claim the morbidity from bilateral muscleharvest, including abdominal wall weaknessand the propensity toward future back pain,“can no longer be defended” in the currentrealm of reliable microsurgical capability andsurgical delay.97 Conversely, proponents claimthat the split muscle technique and addition ofmesh reinforcement limit functional morbidityand that the resultant abdominal wall integrityis dependent upon the closure technique used.98

They adhere to its use as a reliable alternative inhigh-risk patients.

LATISSIMUS DORSI

The latissimus dorsi myocutaneous flap wasoriginally described by Tansini in 1906 andused to cover radical mastectomy defects.99 Ithas since demonstrated remarkable versatilityand is useful in providing purely autogenous,composite implant, and partial mastectomyreconstruction (Figure 11–8). The straightfor-ward anatomy, easy elevation, relative lack ofdonor morbidity, and ability to provide an addi-tional “curtain” of conforming tissue havemade it a reasonable adjunct to breast recon-struction, most commonly in healthy patientsconsidering expander reconstruction.

The indications for latissimus reconstruc-tion vary widely and depend upon the prefer-ences and capabilities of the surgeon. Severalsubsets exist, all governed by the assumptionthat a TRAM flap has been ruled out for eithermedical, anatomic, or personal reasons. Thefirst set includes those patients who are other-

188 BREAST CANCER

wise appropriate candidates for expanderreconstruction but for whom less than optimalcoverage is predicted. This may includepatients who have had a prior radical mastec-tomy and lack a pectoralis, those who have thinmastectomy skin flaps, or those who require alarge skin resection due to inclusive resectionof a remote biopsy site or to prior radiation.

The second set includes those patientsamenable to expander reconstruction who havesufficient coverage and high esthetic expecta-tions. The challenge in unilateral postmastec-tomy expander reconstruction is to provide abreast form which simulates the contralateralside. Prosthetic reconstruction provides a round,firm, relatively immobile breast form which isideally suited for patients with small to interme-diate sized breast and limited to no ptosis.Patients who are moderate or large in size anddevelop some degree of ptosis with age andchildbirth will demonstrate variable asymmetrywith unilateral prosthetic reconstruction. Thesenon-TRAM candidates may elect to undergoeither contralateral mastopexy or composite latis-simus-expander reconstruction for improvedsymmetry. The flap provides supplemental mus-cle and fat, which helps camouflage the under-lying prosthesis and replaces the resected skin,leading to a more natural ptotic breast form.

The third category includes those patientswho prefer an autogenous restoration but lackflap alternatives due to medical or surgical rea-sons. Most patients have a breast volume inexcess of their available flank tissue and requiresupplemental volume in the form of an implant.The resultant satisfaction in esthetic outcomeand greater projection and natural ptosis allaymost patient’s preoperative reluctance toward asupplemental implant.

The fourth and fifth sets involve autogenouslatissimus reconstruction without supplementalprosthesis and apply to two patient extremeswhere the available flank tissue volume simu-lates breast volume. Solely autogenous latis-simus reconstruction is routinely possible inheavier patients having substantial upper flank

tissue. These patients typically have redundantflank skin and additional subcutaneous bulkthat may be incorporated into the flap to pro-vide necessary volume and ptosis. Conversely,patients with marked breast hypoplasia mayalso attain sufficient volume, contour, and sym-metry from a purely autogenous latissimusmyocutaneous flap.

The flap has an extremely reliable bloodsupply and is versatile even in smokers and dia-betics. Partial flap necrosis has been reported inup to 7 percent of patients.100 The most com-mon nuisance is the persistence of seromas,which often requires prolonged drainage oraspiration. Implant-related complicationsinclude implant slippage and capsule contrac-ture. Use of textured, saline expanders andimplants has reduced these complications.

The latissimus dorsi flap represents a popu-lar, extremely reliable option for the mastectomypatient. The results are outstanding, when usedin conjunction with the newer textured, anatomicsaline expanders and implants, typically betterthan those achieved with expanders alone.

RUBENS FLAP

Peter Paul Rubens was known for his portraitsof females with particular fullness in thesuprailiac region. The skin and subcutaneoustissue in this region may be sustained by thedeep circumflex iliac artery, as originally


Figure 11–8. Latissimus dorsi myocutaneous flap recon-struction.

described by Taylor.101 Hartrampf coined thisperi-iliac fat pad the “Rubens flap.”102

The TRAM is the flap of choice in autolo-gous breast reconstruction. One of the benefitsof the TRAM flap is the performance of a con-current abdominoplasty, with resection of oftenlarge volumes of infraumbilical tissue. It is withsome bewilderment that patients complain of agreater lower abdominal circumference, not areduction, and have greater difficulty wearingtheir previously well-fitted clothing. Closing theanterior TRAM donor site leads to accentuationof the peri-iliac tissue and can cause an actualincrease in the peri-iliac circumference. Thisredundancy represents the tissue available forfree tissue transfer after a previous TRAM flap.The predominant indication for use of theRubens flap is therefore a prior TRAM harvestor abdominoplasty. Other indications for use ofthe Rubens flap include thin patients and pro-hibitive anterior abdominal scars.

Flap dissection requires a precise knowl-edge and familiarity with the intrinsic supportof the abdominal wall. The primary disadvan-tage of the flap is the occurrence of an occa-sional flank hernia. Compulsive closure of thedonor site is paramount to the success of thisprocedure and requires a dedicated surgeon todo so. Other potential morbidity includes long-standing seromas that require prolongeddrainage and compression garments. Patientswith this problem also have a higher incidenceof prolonged discomfort and often requiremonitored physical therapy.

The flap is oriented parallel to the iliac crest,with two-thirds of the skin paddle above andone-third below the crest (Figure 11–9A). Aninguinal incision lateral to the femoral pulsewill expose the external oblique. Once this isincised, the round ligament is identified andretracted superiorly to expose the inguinalfloor. An incision through the internal oblique,transversalis, and transversalis fascia willexpose the underlying deep circumflex iliacvessels. This helps guide the remainder of theflap dissection. Once the skin and fat are cut,

the muscle layers are cut immediately adjacentto the pedicle. The inferior skin flap is elevatedabove the tensor of fascia lata to the iliac crest.Subperiosteal dissection will ensure theintegrity of both the deep circumflex iliacartery (DCIA) pedicle and perforators. The lat-eral femoral cutaneous nerve runs inferiorly,within 1 cm of the anterior superior iliac spine,and may lie either above or below the DCIA.This nerve should be preserved.

Donor site closure is initiated by approxi-mation of the transversalis fascia to the ilio-psoas fascia. The remaining flank muscles aresecured to the iliac crest through drill holes andheavy suture or wire.

Deep circumflex iliac artery pedicle lengthfacilitates anastomosis to the preferred thora-codorsal vessels in the majority of cases. Thisflap tends to be less robust than the TRAM andmay exhibit a weak doppler signal, at best. Theflap provides excellent projection (Figure11–9B) and is an ideal option for bilateralreconstruction, which may be performed simul-taneously, concurrent with mastectomy.

SUPERIOR GLUTEAL FLAP

The superior gluteal flap was the first freeflap described for breast reconstruction.103

Microsurgical expertise is essential for suc-cess due to a tedious flap dissection, an inher-ently short vascular pedicle, and because themicroanastomoses are most commonly per-formed to the delicate and variable internalmammary vessels.104

Candidates include patients who fail qualifi-cation for implants due to prior chest-wall irra-diation or “implant anxiety” or who haveabdominal scars precluding TRAM reconstruc-tion. Such scars may have resulted from laparo-tomies, enterotomies, previous abdominoplas-ties, liposuction, or TRAM harvests. This flapmay represent the only autogenous option inthin patients who lack sufficient abdominal orlateral thigh tissue for unilateral or bilateralreconstruction.

190 BREAST CANCER

Like the TRAM, the gluteal flap offers a per-manent, soft, warm, and natural reconstruction.It has a more dense fat–septal network, provid-ing an intermediate size reconstruction withexcellent projection. It may be the flap of choicefor patients who have had a previous TRAM andrequire a staged contralateral mastectomy. Italso offers an inconspicuous donor site.

Flap dimensions typically extend from thelateral midsacrum to within 5 cm of the ASIS.The verticle height of the flap depends on thetissue needed but may vary from 10 to 15 cm.Flap dissection necessitates identification ofthe fragile superior gluteal vessels deep to thegluteus muscle. This pedicle emerges from thegreater sciatic foramen amidst numerousbranches and provides 1.5 to 2.0 cm of pediclelength. The internal mammary vessels areexposed and mobilized through a third perister-nal rib resection. The external jugular vein,rotated down from the mandibular angle, mayserve as a venous alternative if internal mam-mary vein integrity is lacking.

Like its counterpart, the inferior glutealflap is indicated in rare patients who refuseprosthetic reconstruction and who are not can-didates for either TRAM, lateral thigh, orlatissimus flaps. Although the length of thedonor inferior gluteal vessels enable anasto-mosis to the more forgiving thoracodorsalpedicle and the donor site scar is the least con-spicuous of any autogenous option, harvestnecessitates sacrifice of the gluteal motornerve, occasional sacrifice of the posteriorcutaneous nerve, and close dissection to thesciatic nerve, all of which may lead to tran-sient pain syndromes and weakness withambulation; prolonged rehabilitation may berequired. For these reasons, the gluteal flap isgenerally the least favored flap in the breastreconstruction algorithm.

LATERAL THIGH FLAPS

The lateral transverse thigh flap and tensor offascia lata flap are two reconstructive variants

that are based upon the lateral femoral circum-flex vessels and make use of the lateral “ridingbreeches” or “saddle bags.”105 The pedicle trans-gresses through and requires the sacrifice of themodest tensor muscle. Preservation of adjacentfascia lata helps to ensure lateral knee stabilitywithout functional compromise. More impos-ing, and representing the primary disadvantage,are the often disfiguring lateral thigh scars,which are long and remain poorly camouflaged.

Preoperative design requires experience andprecision. An excessive subcutaneous harvestwill result in objectionable lateral thigh contourdeficits. These are difficult to correct but do


Figure 11–9. Ruben’s flap A, Bilateral flap design; B, Imme-diate postoperative projection demonstrated. Reproducedwith permission from William W. Shaw, MD, Division of PlasticSurgery, UCLA.

A

B

benefit from delayed suction lipectomy. Patientsmay require prolonged drainage and garmentcompression to limit the tendency towardseroma formation. Advantages include a 7 to8 cm vascular pedicle, excellent flap projection,and the ability to perform concurrent bilateralsimultaneous harvests and reconstruction.

BILATERAL BREAST RECONSTRUCTION

Indications for contralateral prophylactic mas-tectomy include a strong family history ofbreast cancer, positive genetic testing, lobularcarcinoma in situ (LCIS), cancer anxiety, andequivocal or progressively difficult clinicaland/or radiographic examinations.

With improvements in breast cancer screen-ing, a greater number of early breast cancers arebeing detected in young, premenopausalpatients, many of whom have some degree offamilial cancer history. Patients with youngfamilies present with the intent to absolve breastcancer risk for the benefit of their young onesand represent a new indication for either pro-phylactic or bilateral mastectomy. Breast cancerawareness has elevated the level of sophistica-tion of all patients. Prosthetic and autologousreconstruction is a known entity that continuesto become more reliable, safe, and estheticallysatisfying. As this awareness becomes moreapparent and outcomes improve, it is not sur-prising that an increasing number of susceptiblewomen are at least questioning the option ofbilateral ablation and immediate reconstruction.

Esthetic outcome is often better in bilateralreconstruction than in unilateral reconstructiondue to the symmetry achieved. Macromastiaand pseudoptosis are not compounding factorssince skin redundancy may be addressed sym-metrically. Bilateral implants and/or permanentexpander implants, postmastectomy, usuallyprovide exceptional results, in contrast to unilat-eral procedures, which exaggerate implant char-acteristics. Postoperative adjustability ensures asymmetric result. This is ideal for the older

patient or the patient with marginal reserve whodesires to avoid an external prosthesis and couldnot tolerate a long, grueling procedure.

The TRAM flap, once again, is the flap ofchoice, providing reliability and minimal mor-bidity in bilateral autologous reconstruction(Figure 11–10). Sufficient tissue is present forbilateral reconstruction in 75 to 80 percent ofpatients. The majority of patients when advisedof the ability to perform an immediate, single-stage, highly esthetic and symmetric, perma-nent, bilateral autogenous reconstruction, andsimultaneously rid themselves of an often per-vasive lower abdominal pannus, are, mostoften, highly grateful and not overly concernedabout the possibility of having slightly smallerbreasts if less than profound amounts of tissueare available. Advantages of bilateral TRAMreconstruction include the ability to perform asimultaneous harvest in the supine position.

Either bilateral conventional or freeTRAMS may be performed. The vascular relia-bility of bilateral “hemi-TRAM” flaps is nor-mally adequate, because cross perfusion acrossthe midline is not necessary, depending on anabsence of excessive scarring, obesity, priorradiation, and history of smoking. Use of theconventional flaps is usually faster and techni-cally simpler than free TRAM reconstructionbut requires inherent sacrifice of both rectusmuscles, which may lead to objective and sub-jective abdominal wall weakness in the major-ity of more active patients. Extensive superiorabdominal dissection and tunneling is requiredand may prolong postoperative discomfort andrecovery. Transposition of bilateral flaps maylead to an upper abdominal bulge and violatesome aspect of both inframammary folds, com-promising final cosmesis.

The advantages of bilateral free TRAMreconstruction include limited muscle harvest,limited upper abdominal dissection (which min-imizes discomfort and expedites recovery),unparalleled esthetic outcome, and a greater tol-erance to adjuvant radiation if deemed neces-sary. Lateral extension of these free “hemi-flaps”

192 BREAST CANCER

may be incorporated to boost tissue volume inthin patients and is made possible by the excep-tional blood supply. Assuming the presence ofsuitable recipient vessels and sufficient experi-ence of the reconstructive team, free TRAMreconstruction for bilateral restoration is usuallypreferred for the ultimate benefit of the patient.The incidence of abdominal wall bulging andhernia formation is similar for free and conven-tional bilateral reconstruction and is dependentupon the security and detail of the abdominalclosure. Some surgeons perform a mild bowelprep preoperatively to facilitate closure in bilat-eral cases. The potential to exclude medial,diminutive, or outlying perforators in bilateralfree reconstruction facilitates fascial closurewithout the use of mesh. Although it is reportedthat mesh may be avoided in 60 to 80 percent ofpatients having bilateral TRAM reconstruction,the use of a more relaxed closure using meshmay facilitate postoperative comfort, recovery,and return of bowel motility. Prolene mesh iscurrently the authors’ preferred choice for rein-forcement. Closure may be facilitated by a pre-operative bowel prep, the appropriate use ofrelaxing agents, avoidance of nitrous oxide(which can lead to bowel dilitation) and the useof lateral external oblique relaxing incisions. Theincidence of true hernias is rare. Lower abdomi-nal attenuation or abdominal wall bulging occursin 4.4 to 20 percent of cases.

RECONSTRUCTION OF THE PARTIAL MASTECTOMY DEFECT

Breast conserving surgery combined with adju-vant radiation has been accepted as a regimeequivalent to modified radical mastectomy forearly stage (I and II) breast cancer. The tech-nique is popular due to its ability to eradicatebreast cancer while preserving a maximal vol-ume of breast tissue.

Skin incisions are designed directly over thelesion, and skin and subcutaneous tissues arepreserved unless involved in the lesion. Closureinvolves subcuticular closure only and the

avoidance of drains. The resulting deformityafter lumpectomy or quadrantectomy dependson initial breast size, tumor size and location,radiation dose, surgical technique, and adjuvantchemotherapy. The relative excision, in propor-tion to breast size, is perhaps the most importantfactor. Patients with large, pendulous breastsmay easily accommodate a 4 cm lumpectomy.The same resection in a smaller-breasted womanmay lead to an unacceptable cosmetic result.Radiation therapy exaggerates the tissue deficitin the form of ischemic fibrous contracture.

The treated breast is subject to edema, retrac-tion, fibrosis, calcification, hyperpigmentation,depigmentation, telangiectasia formation, andatrophy. It is not until 24 to 36 months postradi-ation that radiation-induced changes stabilize.Initial edema camouflages the initial deficit andis replaced with fibrosis and contracture thattends to worsen with time. Deficits within thelower pole tend to retract upward. Deficits alongthe superomedial aspect of the breast are diffi-cult to camouflage due to the paucity of availableadjacent tissue and are, unfortunately, sociallyconspicuous. Centrally located lesions are moreforgiving unless resection involves some aspectof the nipple-areolar complex.

An assessment of the patient’s overall onco-logic risk for recurrence should be consideredprior to any attempt at partial mastectomyreconstruction. Breast cancer history, the natureof the inciting lesion, and the patient’s familyhistory should be reviewed prior to an addi-tional procedure that may further affect subse-quent screening examinations. In any event,


Figure 11–10. Bilateral free TRAM reconstruction.

Figure 11–11. Reconstruction of the partial mastectomydefect. Superolateral reconstruction with latissimus dorsimyocutaneous flap.

stabilization of the breast appearance is a pre-requisite and occurs 1 to 3 years postradiation.

Investigators have attempted to classify thespectrum of partial mastectomy deficits andrelate them to specific treatment options. Clas-sification is based upon the localized deficit ofskin and glandular tissue, malposition and/ordistortion of the areola, and the extent offibrous contracture of the breast.106 Local flaptransposition is recommended for mild defor-mities, whereas myocutaneous flaps are reservedfor more extensive defects.

Approximately 15 percent of patients treatedwith BCT are not content with the esthetic out-come.107 These patients often seek consultationto improve selfesteem and body image. Carefulassessment of the actual and apparent tissuedeficits are crucial in the selection of the appro-priate reconstructive strategy. Contour deficitssignify substantial parenchymal loss, whereasradiation contracture represents extensive cuta-neous deficits. Nipple-areolar distortion neces-sitates a substantial increase in cutaneousreplacement, as central areolar support requiresdermal rather than subcutaneous support.

The majority of patients are poor candidatesfor implant reconstruction. Cutaneous fibrosisresponds poorly to implant displacement, andimplant radio-opacity impairs an already com-plex screening examination. Autologous tis-sues, conversely, are reliable, versatile, and pro-vide all the components necessary for partialrestoration. The inherent vascularity may actu-ally improve the quality of the relativelyischemic and radiated recipient tissue.

Large central excisions involving the nipple-areolar complex and primary closure take on aflat, attenuated appearance, lacking projection.These defects may be reconstructed in one oftwo ways. It may be possible to mobilize a skinglandular flap based on inferolateral perforatorsfrom the underlying pectoral fascia, which isthen mobilized into the defect. The curvilinearincision extends from the inferomedial aspect ofthe previous areola to the central inframammaryfold. All but a central skin paddle, rotated into

the areolar defect, is de-epithelialized. Primaryskin closure is facilitated by undermining at theparenchymal interface. The second techniqueparallels conventional mastopexy and enablessuperior advancement of an inferior dermoglan-dular pedicle. It is performed through a Wise orkeyhole pattern incision.106,108–109

Upper outer quadrant excisions are the mostfrequent and, fortunately, the most forgiv-ing.106,108–109 The great majority of these exci-sions do not require reconstruction. Occasion-ally, delayed augmentation, scar lengtheningvia Z-plasty, and areolar transposition are indi-cated. If a discrepancy between the medial andlateral breast quadrant is recognized due to asubstantial superolateral resection, immediatecentralization of the nipple-areolar complexover the point of maximal projection is war-ranted. This involves simple areolar transposi-tion after release of the dermal attachments.Wide excisions may require transfer of regionalor distant tissue. The latissimus dorsi myocuta-neous flap represents the ideal choice for thesedefects (Figure 11–11).

Partial inferior defects may be corrected onan immediate or delayed basis. The occasionalpatient, lacking significant radiation change,may benefit from delayed insertion of a smallround or custom (one-third) implant for volumereplacement. Most defects, however, benefitfrom a procedure that parallels a standard supe-rior pedicle reduction mammoplasty.108–109 Theresection and reconstruction are facilitatedthrough a standard keyhole pattern. Medial and

194 BREAST CANCER

lateral parenchymal flaps are mobilized fromthe pectoralis fascia and inframammary foldand mobilized into the inferior defect, whetherit be lateral, central, or medial.

Supra-areolar defects are socially conspicu-ous and necessitate local reconstruction due tothe paucity of available adjacent tissue and thetendency to develop a visible and depressed scar.These defects are corrected by superior advance-ment of the areolar complex, based on an infe-rior pedicle, in a procedure similar to an inferiorpedicle reduction mammoplasty.106,108–109

The latissimus dorsi myocutaneous flap rep-resents the flap of choice for the majority ofpartial mastectomy defects. Its regional loca-tion, malleability, ease of dissection, and lackof donor site morbidity are ideally suited forthis indication. All breast conservation defectsshould be reconstructed by overcorrecting theskin and soft tissue deficits. In general, twicethe apparent tissue loss should be inset to com-pensate for normal wound contracture, contin-ued retraction of the postradiation fibrosis, andanticipated muscle atrophy inherent in raisingmuscle flaps. The muscle may be folded andcontoured to accommodate the most irregulardefects. Although small skin paddles may beharvested to precisely accommodate the appar-ent skin deficit, a typical 4 ´ 6 cm skin paddlefacilitates flap harvest and replacement of com-promised or contracted radiated skin.

Although partial latissimus harvests arepossible, the majority of partial mastectomydefects warrant total flap elevation. Preserva-tion of the thoracodorsal nerve will maintaingreater muscle bulk but lead to early postoper-ative contractions. Compulsive fixation at therecipient site is necessary to avoid disruption.Transection or resection of the muscular inser-tion will help avoid the typical bulge within theanterior axilla. Finally, supporting the radiatednative breast skin with a de-epithelialized por-tion of the transposed skin paddle will improveultimate wound contour.

The TRAM flap represents a flap of sub-stantial bulk, typically incurring greater donor

site morbidity and a longer recovery. It wouldappear less economic in restoration of limitedtissue defects. It is indicated for the reconstruc-tion of large inferior pole deficits in large-breasted women.

Continued surveillance for recurrent cancerafter partial reconstruction should proceedunimpeded. Studies comparing pre- and post-operative mammograms after partial recon-struction have confirmed the radiolucency ofthese flaps. The development of new microcal-cifications, fat necrosis, and new lesions areeasily discernible. Some reports, interestingly,have noted improved mammographic visualiza-tion and resolution of breast density and fibro-sis as a result of improved local vascularity.

Immediate reconstruction of partial mastec-tomy defects is gaining popularity. The demandfor these techniques has evolved due to a ten-dency toward more aggressive resection in BCTand accumulated experience with unfavorabletumors. Petit and colleagues reported that imme-diate reconstruction of the partial mastectomydefect was performed in 25 percent of cases.They advocated close preoperative collaborationto optimize cosmetic results and enable“improved radicality” of the surgical breast con-servation.110 Thus, the potential for immediatepartial mastectomy reconstruction facilitates amore aggressive resection or marginal clearancein BCT and may lessen the need and/or fre-quency of re-excision. Also, it may lessen theneed for staged reconstruction following radia-tion-induced exaggeration of the defect.

NIPPLE-AREOLAR RECONSTRUCTION

Nipple-areolar reconstruction is a critical stagein breast reconstruction and may add remark-able realism to the new breast mound (Figure11–12). Areolar tattooing facilitates symmetryin color, may camouflage minor discrepanciesand scars, and lacks the morbidity associatedwith skin grafts. Nipple reconstruction is typi-cally performed at a second stage, at the time ofport removal, or breast mound revision.


Although single-stage reconstruction may beperformed, attaining symmetry of nipple-areo-lar position is crucial to esthetic outcome and ismost accurately attained at a second stage,when dermal edema and skin elasticity havenormalized.

The insensate, adynamic nipple remains sta-tic in size, contour, and projection and willlikely be visualized through undergarments,swimsuits, and clothing. The patient’s finalassessment and perception may closely parallelthe quality and symmetry of the newly con-structed nipple. Consideration of a symmetryprocedure should, therefore, be entertainedprior to final nipple reconstruction and shouldencompass whether the patient prefers support(bra) and to what extent. Simulation in a bra orsheer blouse preoperatively may help thepatient’s understanding of these issues.

Various techniques of nipple reconstructionare available and provide a range of caliberpotential projection. Modification, reduction,or composite grafting of the contralateral nip-ple may be considered as an option in thepatient with redundant nipples. Although thisrepresents the most realistic reconstruction, itnecessitates a procedure on the remainingintact nipple and is sensitive perioperatively.

Local flaps are the technique of choice fornipple reconstruction, most of which are vari-ants of the original skate flap. The skate flap

has proven to be a reliable workhorse, with thepotential for a long projectile nipple ifneeded.109 The donor site does require a skingraft, most commonly harvested from thegroin, inner thigh, or axilla. Precise demarca-tion of the central nipple complex is critical andserves as a basis for dermal flap elevation. Thelateral dermal wings are elevated, preservingthe central nipple core and an inferior extensionof fat. These components are elevated, preserv-ing the subcutaneous perforators, and then sur-faced by the lateral wings. The circular de-epithelialized harvest site is then covered witha full thickness skin graft.

The Star flap,111 C-V flap,112 fishtail flap(McCraw), and double opposing tab flap(Kroll) are additional flap options, most ofwhich are modifications of the skate flap.Although they provide less nipple projectionthan does the skate flap, they avoid the need fora skin graft. These are excellent alternatives forthe majority of patients with small to moderatesized contralateral nipples.

Intradermal areolar tattoo has greatly simpli-fied the final phase of restoration and addsabrupt and striking realism to the physical breastform. It remains an artistic challenge among sur-geons to simulate contralateral areolar pigment.This final phase enables the surgeon one ad-ditional opportunity to optimize symmetry.Nipple-areolar reconstruction may enhance thefocus of the reconstructed breast and improveoverall patient incorporation of the reconstructedbreast, both physically and psychologically.

REFERENCES

1. Ernster VL, Barclay J, Kerlikowske K, et al. Inci-dence of and treatment for ductal carcinoma insitu of the breast. JAMA 1996;275:913–8.

2. Grisotti A. Conservative treatment of breast can-cer. In: Spear S, editor. The breast: principlesand art. Philadelphia: Lippincott-Raven;1998. p. 137.

3. Goldwyn RM. Vincenz Czerny and the begin-nings of breast reconstruction. Plast ReconstrSurg 1978;61:673–81.

196 BREAST CANCER

Figure 11–12. Completed nipple-areolar reconstruction.

4. Georgiade G, Georgiade N, McKarty KS Jr, et al.Rationale for immediate reconstruction of thebreast following modified radical mastectomy.Ann Plast Surg 1982;8:20–28.

5. Rosenqvist S, Sandelin K, Wickman M. Patients’psychological and cosmetic experience afterimmediate breast reconstruction. Eur J SurgOncol 1996;22:262–6.

6. Wellisch DK, Schain WS, Noone RB, et al. Psy-chosocial correlates of immediate versusdelayed reconstruction of the breast. PlastReconstr Surg 1985;76:713–8.

7. Shover LR, Yetman RJ, Tuason LJ, et al. Partialmastectomy and breast reconstruction. A com-parison of their effects on psychosocial adjust-ment, body image, and sexuality. Cancer1995;75:54–64.

8. Noguchi M, Kitagawa H, Kinoshita K, et al. Psy-chologic and self-assessments of breast con-serving therapy compared with mastectomyand immediate breast reconstruction. J SurgOncol 1993;54:260–6.

9. Toth BA, Lappert P. Modified skin incisions formastectomy: the need for plastic surgery inputin pre-operative planning. Plast Reconstr Surg1991;87:1048–53.

10. Kroll SS, Ames F, Singletary SE, et al. The onco-logic risks of skin preservation at mastectomywhen combined with immediate reconstructionof the breast. Surg Gynecol Obstet 1991;172:17–20.

11. Newman LA, Keurer HM, Hunt KK, et al. Pre-sentation, treatment, and outcome of localrecurrence after skin sparing mastectomy andimmediate breast reconstruction. Ann SurgOncol 1998;5:620–6.

12. Hidalgo DA. Aesthetic refinement in breastreconstruction: complete skin sparing mastec-tomy with autogenous tissue transfer. PlastReconstr Surg 1998;102:63–70.

13. Carlson GW, Bostwick J, Styblo TM, et al. Skinsparing mastectomy: oncologic and reconstruc-tive considerations. Ann Surg 1997;225:570–75.

14. Yule GJ, Concannon MJ, Croll G, et al. Is thereliability with chemotherapy following immedi-ate breast reconstruction? Plast Reconstr Surg1996;97:969–73.

15. Grotting JC, Urist MM, Maddox WA, Vasconez LO.Conventional TRAM flap versus free microsur-gical TRAM flap for immediate reconstruction.Plast Reconstr Surg 1989;84(6):1005–6.

16. Elliott LF, Eskenazi L, Beegle PH Jr, et al. Imme-diate TRAM flap breast reconstruction: 128

consecutive cases. Plast Reconstr Surg 1993;92:217–27.

17. Schuster RH, Kuske RB, Young VL, Fineberg B.Breast reconstruction in women treated withradiation therapy for breast cancer: cosmesis,complications, and tumor control. Plast Recon-str Surg 1992;90:445–52.

18. Evans GR, Schusterman MA, Kroll SS, et al.Reconstruction and the radiated breast: is therea role for implants? Plast Reconstr Surg 1995;96:1111–5.

19. Spear S, Majidian A. Immediate breast recon-struction in 2 stages using textured integrated-valve tissue expanders and breast implants: aretrospective review of 171 consecutive breastreconstructions from 1989–1996. Plast Recon-str Surg 1998;101:53–63.

20. Jackson WB, Goldson AL, Staud C. Post-opera-tive irradiation following immediate breastreconstruction using a temporary tissueexpander. J Natl Med Assoc 1994;86:538–42.

21. Dickson MG, Sharpe BT. The complications of tis-sue expansion in breast reconstruction: a reviewof 75 cases. Br J Plast Surg 1987;40:629–35.

22. Jacobson GM, Sause WT, Thompson JW, PlenkHP. Breast irradiation following silicone gelimplants. Int J Radiat Oncol Biol Phys 1986;12:835–8.

23. Chu FC, Kaufman TP, Dawson GA, et al. Radia-tion therapy of cancer in prosthetically aug-mented or reconstructed breasts. Radiology1992;185:429–33.

24. Spear S. Prosthetic reconstruction in the radiatedbreast. In: Spear S, editor. The breast: princi-ples and art. Philadelphia: Lippincott-Raven;1998. p. 399–406.

25. Zimmerman RP, Mark RJ, Kim AL, et al. Radiationtolerance of transverse rectus myocutaneousfree flaps used in immediate breast reconstruc-tion. Am J Clin Oncol 1998;21:381–5.

26. Hunt KK, Baldwin BJ, Strom EA, et al. Feasibil-ity of post-mastectomy radiation therapy afterTRAM flap breast reconstruction. Ann SurgOncol 1997;4:377–84.

27. Sultan MR, Smith ML, Estabrook A, et al. Imme-diate breast reconstruction in patients withlocally advanced disease. Ann Plast Surg1997;38:345–9.

28. Styblo TM, Lewis MM, Carlson GW, et al. Imme-diate breast reconstruction for stage III breastcancer using transverse rectus myocutaneousflaps. Ann Surg Oncol 1998;3:375–80.

29. Brody GS. Safety and effectiveness of breast


implants. In: Spear S, editor. The breast: prin-ciples and art. Philadelphia: Lippincott-Raven;1998. p. 336–46.

30. Peters W, Keystone E, Snow K, et al. Is there a rela-tionship between autoantibodies and silicone gelimplants? Ann Plast Surg 1994;32:1–5.

31. Kossovsky N, Heggers JP, Robson MC. Experi-mental demonstration of the immunogenicityof silicone protein complexes. J Biomed MaterRes 1987;21:1125–33.

32. Heggers JP, Kossovsky N, Parsons RW, et al. Bio-compatibility of silicone implants. Ann PlastSurg 1983;11:38–45.

33. DeCamara DI, Sheridam SM, Kammer BA. Rup-ture and aging of silicone breast implants. PlastReconstr Surg 1993;91:828–34.

34. Greenwald WB, Randolph M, May JW. Mechani-cal analysis of explanted silicone breastimplants. Plast Reconstr Surg 1996;98:269–72.

35. Phillips JW, Decamara DL, Lockwood MD, et al.Strength of silicone breast implants. PlastReconstr Surg 1996;97:1215–25.

36. Robinson OG, Bradley EL, Wilson DS. Analysisof explanted silicone implants: a report of 300patients. Ann Plast Surg 1995;34:1–6.

37. Ahn CY, Shaw WW. Regional silicone gel migra-tion in patients with ruptured implants. AnnPlast Surg 1994;33:201–8.

38. Ahn CY, DeBruhl ND, Gorczyca DP, et al. Com-parative silicone breast implant evaluationusing mammography, sonography, and mag-netic resonance imaging: experience with 59implants. Plast Reconstr Surg 1994;94:620–7.

39. Leibman AL, Kruse BD. Imaging of breast cancerafter augmentation mammoplasty. Ann PlastSurg 1993;30:111–5.

40. Silverstein MJ, Gamagami P, Handel N. Missedbreast cancer in an augmented woman usingimplant displacement mammography. AnnPlast Surg 1990;25:210–3.

41. Carlson GW, Curley SA, Martin FE, et al. Thedetection of breast cancer after augmentationmammoplasty. Plast Reconstr Surg 1993;91:837–40.

42. Gumico CA, Pin P, Young VL, et al. The effect ofbreast implants on the radiographic detectionof microcalcifications and soft tissue masses.Plast Reconstr Surg 1989;84:772–8.

43. Birdsell DC, Jenkins H, Berkel H. Breast cancerdiagnosis and survival in women with andwithout breast implants. Plast Reconstr Surg1993;92:795–800.

44. Deapon DM, Bernstein L, Brody GS. Are breastimplants anti-carcinogenic? A 14 years follow-

up of the Los Angeles study. Plast ReconstrSurg 1997;99:1346–53.

45. Deapon DM, Pike MC, Casagrande JT, Brody GS.The relationship between breast cancer andaugmentation mammoplasty: an epidemiologicstudy. Plast Reconstr Surg 1986;77:361–7.

46. Engel A, Lamm SH. Risk of sarcomas of thebreast among women with breast augmenta-tion. Plast Reconstr Surg 1992;89:571–2.

47. Su CW, Dreyfuss DA, Krizek TJ, et al. Siliconeimplants and the inhibition of cancer. PlastReconstr Surg 1995;96:513–8.

48. Brody GS, Conway DP, Deapon DM, et al. Con-sensus statement on the relationship of breastimplants to connective tissue disorders. PlastReconstr Surg 1992;90:1102–5.

49. Gabriel SE, O’Faflon WM, Kurland LT, et al. Riskof connective tissue diseases and other disor-ders after breast implantation. N Engl J Med1994;330:1697–702.

50. Giltay EJ, Moens HJB, Riley AH, et al. Siliconebreast prostheses and rheumatic symptoms: aretrospective follow-up study. Ann Rheum Dis1994;53:194–6.

51. Radovan C. Breast reconstruction after mastec-tomy using the temporary expander. PlastReconstr Surg 1982;69:195–208.

52. Colen SR. Immediate two stage breast reconstruc-tion utilizing a tissue expander and implant. In:Spear S, editor. The breast: principles and art.Philadelphia: Lippincott-Raven; 1998. p. 375–86.

53. Forman DL, Chiu J, Restifo RJ, et al. Breastreconstruction in previously irradiated patientsusing tissue expanders and implants: a poten-tially unfavorable result. Ann Plast Surg 1998;40:360–3.

54. Mandrekas AD, Zambacos GJ, Katsantoni PN.Immediate and delayed breast reconstructionwith permanent tissue expanders. Br J PlastSurg 1995;48:572–8.

55. Francel TJ, Ryan JJ, Manson PM. Breast recon-struction utilizing implants: a local experienceand comparison of three techniques. PlastReconstr Surg 1993;92:786–94.

56. Yeh KA, Lyle G, Wei JP, et al. Immediate breastreconstruction in breast cancer: morbidity andoutcome. Am J Surg 64:1195–9.

57. Wickman M, Jurell G, Sandelin K. Technicalaspects of immediate breast reconstruction: 2year follow-up of 100 patients treated conserv-atively. Scand J Plast Reconstr Surg 1998;32:265–73.

58. Caffee HH. Textured silicone and capsule contrac-ture. Ann Plast Surg 1990;24:197–9.

198 BREAST CANCER

59. Pakium AI, Young CS. Submuscular breast recon-struction: a one stage method of tissue expan-sion. Ann Plast Surg 1987;19:312–7.

60. Kroll SS, Baldwin B. A comparison of outcomeusing three different methods of breast recon-struction. Plast Reconstr Surg 1992;90:455–62.

61. Hartrampf CR Jr, Scheflan M, Black PW. Breastreconstruction with a transverse abdominalisland flap. Plast Reconstr Surg 1982;69:216–9.

62. Boyd JB, Taylor GI, Corlett R. The vascular territo-ries of the superior and deep inferior epigastricsystems. Plast Reconstr Surg 1984;73:1–16.

63. Moon HK, Taylor GI. The vascular anatomy of thetransverse rectus myocutaneous flaps based onthe deep superior epigastric system. PlastReconstr Surg 1988;82:815–32.

64. Bostwick J. Plastic and reconstructive breastsurgery. St Louis: Quality Medical Publishing;1990.

65. Watterson PA, Bostwick J, Hester TR, et al.TRAM flap anatomy correlated with a ten yearclinical experience with 556 patients. PlastReconstr Surg 1995;95:1185–94.

66. Hartrampf CR, Bennett GK. Autogenous tissuereconstruction in the mastectomy patient: acritical review of 300 patients. Ann Surg1987;205:508–19.

67. Hartrampf CR, Michelow BJ. Breast reconstruc-tion with living tissue. Norfolk (VA): HamptonPress; 1991.

68. Little JW. Breast reconstruction by the unipedicleTRAM operation: muscle splitting technique.In: Spear S, editor. The breast: principles andart. Philadelphia: Lippincott-Raven; 1998.p. 521–34.

69. Takayanagi S. Extended transverse rectus abdomi-nus myocutaneous flap. Plast Reconstr Surg1993;92:757–8.

70. Yamamota Y, Nohira K, Sugihara T, et al. Superi-ority of the microvascularly augmented flap:analysis of 50 transverse rectus abdominusmyocutaneous flaps for breast reconstruction.Plast Reconstr Surg 1996;97:79–83.

71. Paige KT, Bostwick J, Bried JT, Jones G. A com-parison of morbidity from bilateral, unipedi-cled and unilateral, unipedicled TRAM flapbreast reconstructions. Plast Reconstr Surg1998;101:1819–27.

72. Williams JK, Bostwick J III, Bried JY, et al.TRAM flap breast reconstruction after radia-tion treatment. Ann Surg 1995;221:756–64.

73. Jacobson WM, Meland NB, Woods JE. Autolo-gous breast reconstruction with use of trans-verse rectus myocutaneous flap: Mayo clinic

experience with 147 cases. Mayo Clin Proc1994;69:635–40.

74. Berrino P, Campora E, Leone S, et al. The trans-verse rectus myocutaneous flap for breastreconstruction in obese patients. Ann PlastSurg 1991;27:221–31.

75. Takeishi M, Shaw WW, Ahn CY, et al. TRAMflaps in patients with abdominal scars. PlastReconstr Surg 1997;99:713–22.

76. Kroll SS, Gheradini G, Martin JE, et al. Fat necro-sis in free and pedicled TRAM flaps. PlastReconstr Surg 1998;102:1502–7.

77. Slavin SA, Goldwyn RM. The midabdominal rec-tus abdominus myocutaneous flap: review of236 flaps. Plast Reconstr Surg 1988;81:189–97.

78. Slavin SA, Hein KD. The mid-abdominal transverserectus abdominus myocutaneous flap. In: SpearS, editor. The breast: principles and art. Philadel-phia: Lippincott-Raven; 1998. p. 565–76.

79. Callegari PR, Taylor GI, Caddy CM, et al. Ananatomic review of the delay phenomenon. I.Experimental studies. Plast Reconstr Surg1992;89:397–407.

80. Morris SF, Taylor GI. The time sequence of thedelay phenomenon: when is a surgical delayeffective? An experimental study. Plast ReconstrSurg 1995;95:526–33.

81. Taylor GI. The surgically delayed unipedicledTRAM flap for breast reconstruction. AnnPlast Surg 1996;36:242–5.

82. Codner MA, Bostwick J, Nahai F, et al. TRAM flapvascular delay for high risk breast reconstruc-tion. Plast Reconstr Surg 1995;96:1615–22.

83. Restifo RJ, Ahmed SS, Isenburg JS, et al. Timing,magnitude and utility of surgical delay in theTRAM flap. I. Animal studies. Plast ReconstrSurg 1997;99:12–16.

84. Serletti JM, Moran SL. Free versus the pedicledTRAM flap: a cost comparison and outcomeanalysis. Plast Reconstr Surg 1997;100:1418–24.

85. Schusterman MA, Kroll SS, Weldon ME. Imme-diate breast reconstruction: why the freeTRAM over the conventional TRAM flap.Plast Reconstr Surg 1992;90:255–61.

86. Kroll SS, Netscher DT. Complications of theTRAM flap breast reconstruction in obesepatients. Plast Reconstr Surg 1989;84:866.

87. Schusterman MA, Kroll SS, Miller MJ, et al. Thefree transverse rectus myocutaneous flap forbreast reconstruction: one center’s experiencewith 211 consecutive cases. Ann Plast Surg1994;32:234–41.

88. Suominen S, Asko-Seljavaara S, Kinnunen J, et al.Abdominal wall competence after free trans-


verse rectus abdominus myocutaneous flapharvest: a prospective study. Ann Plast Surg39:299–34.

89. Kind GM, Rademaker AW, Mustoe TA. Abdomi-nal wall recovery following TRAM flap: afunctional outcome study. Plast Reconstr Surg1997;99:417–28.

90. Kroll SS, Schusterman MA, Reece GP, et al.Abdominal wall strength, bulging, and herniaafter TRAM flap breast reconstruction. PlastReconstr Surg 1995;96:616–9.

91. Fitoussi A, Le Taillandier M, Biffaud JC, et al.Functional evaluation of the abdominal wallafter raising a rectus abdominus myocutaneousflap. Ann Chir Plast Esthet 1997;42:138–46.

92. Hartrampf CR Jr, Bried JT. General considera-tions in TRAM flap surgery. In: Hartrampf CR,editor. Breast reconstruction with living tissue.New York: Raven Press; 1991. p. 33–70.

93. Kroll SS, Marchi M. Comparison of strategies forpreventing abdominal wall weakness afterTRAM flap breast reconstruction. Plast ReconstrSurg 1992;89:1045–51.

94. Kroll SS, Gherardini G, Martin JE, et al. Fatnecrosis in free and pedicled TRAM flaps.Plast Reconstr Surg 1998;102:1502–7.

95. Shestak KC. Bipedicle TRAM flap reconstruction.In: Spear S, editor. The breast: principles and art.Philadelphia: Lippincott-Raven; 1998. p. 535–46.

96. Petit JY, Rietjens M, Ferreira MA, et al. Abdom-inal sequellae after pedicled TRAM flap breastreconstruction. Plast Reconstr Surg 1997;99:723–9.

97. Jensen JA. Is double pedicle TRAM flap recon-struction of a single breast within the standardof care? Plast Reconstr Surg 1989;102:586–7.

98. Spear SL, Hartrampf CR Jr. The double pedicleTRAM flap and the standard of care. PlastReconstr Surg 1998;100:1592–3.

99. Maxwell GP. Iginio Tansini and the origin of thelatissimus dorsi musculocutaneous flap. PlastReconstr Surg 1980;65:686–92.

100. Hammond DC, Fisher J. Latissimus dorsi musculo-

cutaneous flap breast reconstruction. In: Spear S,editor. The breast: principles and art. Philadel-phia: Lippincott-Raven; 1998. p. 477–90.

101. Taylor GI, Townsend P, Corlett R. Superiority ofthe deep circumflex iliac vessels as the supplyfor free groin flaps. Clinical work. PlastReconstr Surg 1979;64:745–59.

102. Elliott LF, Hartrampf CR Jr. The Rubens flap.The deep circumflex iliac artery flap. ClinPlast Surg 1998;25:283–91.

103. Fugino T, Harashina T, Endomoto K. Primarybreast reconstruction after a standard radicalmastectomy by a free flap transfer. PlastReconstr Surg 1976;58:372–4.

104. Shaw WW. Superior gluteal free flap breast recon-struction. Clin Plast Surg 1998;25:267–74.

105. Elliott LF, Beegle PH, Hartrampf CR Jr. The lat-eral transverse thigh free flap: an alternativefor autogenous-tissue breast reconstruction.Plast Reconstr Surg 1990;85:169–78.

106. Slavin SA. Reconstruction of the breast conser-vation patient. In: Spear S, editor. The breast:principles and art. Philadelphia: Lippincott-Raven; 1998 p. 221–38.

107. Beadle F, Silver B, Botnick L, et al. Cosmeticresults following primary radiation therapy forearly breast cancer. Cancer 1984;54:2911–8.

108. Kroll SS, Singletary SE. Repair of partial mastec-tomy defects. Clin Plast Surg 1998;25:303–10.

109. Grisotti A. Immediate reconstruction after partialmastectomy. Oper Tech Plast Reconstr Surg1994;1:1–12.

110. Petit JY, Rietjens M, Garusi C, et al. Integrationof plastic surgery in the course of breast con-serving surgery for cancer to improve cosmeticresults and radicality of tumor excision. RecentResults Cancer Res 1998;152:202–11.

111. Little JW, Spear SL. The finishing touches in nip-ple–areola reconstruction. Perspect Plast Surg1988;2:1–17.

112. Bostwick J III. Creating a nipple. In: Berger K,Bostwick J III, editors. A women’s decision. St.Louis: Quality Medical Publishing; 1994.

200 BREAST CANCER

201

The modern era of breast cancer treatmentbegan over 100 years ago with the developmentof surgical techniques that emphasized the needfor total resection of tumor. Nevertheless,despite gross total excision, many patients withseemingly localized disease suffered relapse ordistant recurrence and died of their cancer. Thiswas presumably due to the growth of micro-scopic tumor unappreciated at the time of ini-tial therapy. The need for additional, adjuvanttherapy after surgery led to numerous random-ized controlled trials (RCTs) addressing theproblem. The role of adjuvant systemic therapyin increasing survival and decreasing mortalityhas been established by these studies and con-firmed by overview meta-analyses. The basisfor this success is the recognition of and adher-ence to the principles of adjuvant therapy,which are that (1) local treatments do not cureall patients with seemingly localized cancer; (2)populations at high risk of relapse can be iden-tified; (3) patterns of relapse and failure areunderstood; (4) palliative therapy against overtmacroscopic tumor can potentially eradicateoccult microscopic disease; (5) the value ofadjuvant therapy has been validated in RCTs;(6) the choice of adjuvant therapy considers thebiology of the tumor and of the patient and; (7)the benefits of treatment outweigh the toxicityand risks of therapy.

Halstead in the late 19th century appreci-ated that not all patients with early breast can-

cer were cured by surgery. Regrettably this factremains so even today. It is only in the last 30years, however, that the other principles ofadjuvant therapy were applied to the treatmentof early breast cancer leading to the improve-ment of outcome.

IDENTIFICATION OF POPULATIONS AT HIGH RISK OF RELAPSE

Although not every women with early stage,“localized” breast cancer is rendered cancerfree by local treatment, many women are. It iscritical, if additional treatments are to be used,that they be directed at women at highest risk ofrecurrence. Indiscriminate administration ofadjunctive therapies runs the risk of unnecessar-ily exposing those already cured to toxicity, mor-bidity and, with some aggressive approaches,even mortality from treatment. Choice ofpatients to be treated is critical in the develop-ment of successful adjuvant therapy. Forwomen with early stage, operable breast cancer,the single most important prognostic feature forrecurrence and death is the presence or absenceof tumor metastases in the axillary lymphnodes.1,2 Patients with no axillary nodal metas-tases have a 70 to 75 percent chance of long-term disease-free survival (DFS) when treatedwith surgery alone. Patients with any numberof nodal metastases have a 25 to 30 percentchance of long-term survival.3 The risk of

12

Adjuvant Systemic Therapy of Early Breast CancerGERSHON Y. LOCKER, MD

recurrence and death increases with the numberof involved nodes. Women with more than 10nodes involved have an extremely poor progno-sis, less than a 20 percent cure rate as shown inmost studies. From the beginning of the era ofadjuvant therapy trials, patients with involvedaxillary nodes were identified as the highest-risk group and were the subjects of adjuvantapproaches. As systemic adjuvant therapies fornode-positive disease became accepted, atten-tion turned to improving the prognosis of the 30percent of node-negative cases destined to haverecurrences and die. Multiple studies haveshown that in these women, the size of the pri-mary tumor is the most important predictor ofsurgical outcome.3 Women with tumors smallerthan 1 cm in diameter and no involved nodeshave better than a 90 percent chance of long-term DFS. Adjuvant therapy is not a major pri-ority in this group and is not routinely given.On the other hand, women with tumors > 5 cm,despite having no nodal metastases, have aprognosis comparable to patients with axillarynodal spread. Clearly, women with largetumors, whatever the axillary status, are candi-dates for additional therapy beyond surgery.Other than nodal status and tumor size, the useof prognostic features to determine who shouldor should not be considered for adjuvant ther-apy is very controversial. While tumor grademay be helpful,1 the variability of gradingbetween pathologists makes it problematic.Hormone receptor status, proliferation markers(%S phase, thymidine labeling, Ki67), onco-gene expression/mutation (c-erbB-2, p53) maybe more reproducible and have also been corre-lated with outcome of local therapy.4 Neverthe-less, to date, none has been universallyaccepted as an independent predictor for theneed for adjuvant treatment. The AmericanSociety of Clinical Oncology (ASCO) TumorMarker Expert Panel, using criteria of evi-dence-based medicine, could not endorse anyof these tumor markers by themselves as ade-quate to determine the need for additional ther-apy beyond breast surgery.4 Some of the mark-

ers, however, may be of value in determiningthe type of adjuvant therapy to be used (hor-mone receptors, (c-erbB-2). Studies in the pastand current practice use nodal status and tumorsize as primary determinants of need for treat-ment. The other prognostic factors are used(often together) only when the need for, or thepotential benefits of, systemic therapy afterlocal treatment is unclear.

UNDERSTANDING PATTERNS OFRELAPSE AND FAILURE

There are multiple effective treatments of breastcancer. Which one is most appropriate toincrease survival and cure rate after primarytherapy of early disease is dependent on thenature of the failure. Is it due to inadequate con-trol of the primary tumor, regional spread, ordistant metastases? The Halstedian view ofbreast cancer spread was that of a prolongedperiod of local/regional disease before systemicdissemination. By implication, failure to curemight be due to incomplete surgery. As dis-cussed in Chapter 11, multiple studies havefailed to confirm the survival benefit of moreaggressive versus less aggressive surgery. Arecent meta-analysis of 3,400 women in ran-domized trials of more versus less extensivesurgery found no difference in 10-year survivalbetween the two approaches.5 This was truewhether or not the patients had axillary nodalinvolvement. Radiation therapy is anotherapproach to local and regional control. If theHalstedian paradigm were true, local regionalirradiation by killing residual tumor shouldincrease the overall cure rate and survival.Again, most studies have failed to confirm sur-vival benefit with irradiation of the chestwall/draining nodes, despite significantlydecreasing local recurrence5 (see also Chapter16). In studies of local modalities, the decreasein local recurrence did not improve survivalbenefit, in part because of the development ofdistant recurrences. This refutation of the Hal-stedian theory strongly argues that failure to

202 BREAST CANCER

cure localized breast cancer with local therapiesis due to the presence of unappreciated distantmicrometastases, that are destined to grow andkill the patient. To improve surgical results, addi-tional therapy must be directed to the systemicnature of breast cancer in high-risk women.

PALLIATIVE THERAPY AGAINSTOVERT MACROSCOPIC TUMOR AND

POTENTIAL ERADICATION OF OCCULTMICROSCOPIC DISEASE

A basic assumption underlying adjuvant sys-temic therapy of early breast cancer is thatthose therapies that may be only palliativeagainst bulky macroscopic metastases might becurative against microscopic disseminatedtumor that is assumed to be present in high-riskpatients. To develop effective adjuvant systemictherapy of early disease, one must first identifyeffective and safe therapies for advanced-stagebreast cancer.

Systemic therapy for overt advanced breastcancer began 100 years ago when Beatsonobserved shrinkage of locally extensive breastcancers after oophorectomy in premenopausalwomen.6 This phenomenon is based on thetrophic effect of estrogen on approximately halfof all breast cancers studied. Removal of theovaries leading to a drop in endogenous estro-gen levels in younger women can arrest cancergrowth and result in regression. Anotherapproach to depriving breast cancers of estro-gen effects is to block estrogen binding to theprotein, estrogen receptor (ER), in the breastcancer cell cytoplasm. The receptor-estrogencomplex mediates much of the effect of the hor-mone on the cell, and blocking the interaction,such as by removing estrogen, leads to regres-sion of the hormone-dependent cancer.7 Tamox-ifen (Nolvadex) is the prototype competitiveinhibitor of estrogen binding at its receptor. Theability to measure estrogen and progesteronereceptors and thus predict responsiveness toendocrine therapy made hormonal treatment ofovert, metastatic breast cancer a standard

approach.8 Oophorectomy in premenopausalwomen and tamoxifen in women of all agescause significant regressions of clinicallyadvanced estrogen and or progesterone recep-tor–containing breast cancers with generallyacceptable toxicity. They were obvious candi-dates to be tried in the adjuvant setting afterlocal therapy of early disease to treat occultmicrometastases.

Unfortunately, not all breast cancers areestrogen dependent and responsive to hormonalmanipulation. After the introduction of chemo-therapeutic drugs in the 1940s and 1950s, mul-tiple agents were identified that were able tocause temporary shrinkage of tumor in womenwith disseminated breast cancer. Melphelan(PAM), thiotepa (T), cyclophosphamide (C),methotrexate (M), fluorouracil (F), or vinblas-tine (V) have a 20 to 30 percent chance of caus-ing transient regression of metastatic breast can-cer when used alone.9 In the late 1960s Coopercombined five drugs and reported higherregression rates.10 This regimen was the fore-runner of the cyclophosphamide methotrexatefluorouracil (CMF) combination regimen,which was in the 1970s the standard chemo-therapeutic treatment of advanced disease(Table 12–1).11 In the late 1970s, doxorubicin(Adriamycin[A]) was introduced and found tobe the most active single agent against advancedbreast cancer. It was soon combined with otherdrugs11 (ex-cyclophosphamide adriamycin fluor-ouracil [ex-CAF]) and, in the 1980s, became thenew standard treatment of overt metastatic dis-ease (see Table 12–1). Finally, another class ofdrugs, the taxanes (paclitaxel–Taxol and doc-etaxel–Taxotere) were introduced in the lastdecade with activity comparable with that ofdoxorubicin in the palliation of advanced mam-mary cancer. All these drugs can be given safely,though they have significant toxicity. All werecandidates for use in the adjuvant setting.

A recent trend in the treatment of metastaticbreast cancer is the use of high-dose chemother-apy, either with cytokine support of the bonemarrow or with stem cell or bone marrow stor-

Adjuvant Systemic Therapy of Early Breast Cancer 203

age and reinfusion (“bone marrow transplanta-tion”). These therapies are based on the assump-tion of a dose-response curve for drug-inducedcancer cell death. While still controversial inadvanced breast cancer (and very toxic), suchapproaches are also candidates for evaluationagainst the poorest-risk, early-stage disease.

ADJUVANT THERAPY VALIDATED INRANDOMIZED CONTROLLED TRIALS

Effective therapies of advanced breast cancerwere long known and anecdotally used aftersurgery for early breast cancer, but their value asadjuncts to primary therapy was difficult toassess. Unlike in advanced disease, where tumorshrinkage after systemic treatment can be deter-mined directly by observation or radiographi-cally, when used to increase survival after treat-ment of early stage disease, there are no directdeterminants of effectiveness. Cure in an indi-vidual patient with early-stage disease may havebeen achieved as a result of the systemic therapyor might have occurred even if it had not beengiven as a result of the local treatment. Recur-rence may be a sign of failure of adjunctive ther-apy but might have been delayed because of it.The only way to truly evaluate the usefulness ofadditional treatments after primary therapy ofbreast cancer is by large RCTs. Initially, thosetrials randomized women with high-risk, early-stage disease to surgery alone or to surgery plusan experimental adjuvant therapy. As adjuvanttreatments were proved effective, the next gener-ation of trials randomized women to primarytherapy plus “standard” adjuvant systemic treat-ment versus primary therapy plus “experimen-tal” adjuvant treatment. The end point of suchtrials were disease-free interval (DFI) and DFS(the time to recurrence and percentage ofpatients alive without recurrence of cancer at anytime point) and overall survival (time to deathand percentage of patients alive at any timepoint). In general, the benefits of adjuvant sys-temic chemotherapy or hormonal therapies aremore pronounced on DFS than overall survival

because once a patient relapses, there are effec-tive palliative treatments to prolong life inadvanced disease. Perhaps the most convincingeffect of adjuvant treatment, however, is on thepercentage of patients alive long after treatment.Multiple randomized trials have shown that sys-temic treatment of early-stage breast cancer sig-nificantly and reproducibly decreases the risk ofdeath years after local therapy. This benefit per-sists with time, suggesting the likelihood of cure.

CHEMOTHERAPY

Single Agents

As in the case of advanced disease, the earliesttrials of adjuvant chemotherapy were of singleagents. The National Surgical Adjuvant Breastand Bowel Project (NSABP) conducted ran-domized trials in the 1950s and 1960s looking ata short course of thiotepa or fluorouracil aftersurgery versus surgery alone. Overall, there wasno survival benefit to the chemotherapy, butthere was a suggestion that thiotepa might havesome benefit in premenopausal women.12 Therationale for these studies was a belief thatmanipulation of the tumor during surgery mightpromote detachment and spread of its cells andthat chemotherapy might kill the scattered cells.It was only with the realization that failure maybe due to distant metastases already present atthe time of surgery that trials of protractedchemotherapy were undertaken. The risk of tox-icity of these more prolonged approaches led torestriction of the trials to women with node-pos-itive disease. The critical single-agent study wasdone by the NSABP in the early 1970s. It ran-domized 349 women with node-positive diseaseto surgery alone or surgery plus 2 years of inter-mittent oral melphalan.13 At 10 years, there wassignificant improvement in DFS and a trendtoward improved overall survival in thechemotherapy group. The benefits, however,were confined to women under 50 years. Atabout the same time, randomized studies wereshowing the superiority of combination

204 BREAST CANCER

chemotherapy over single-agent therapy in thetreatment of metastatic breast cancer. Would thesame be true for adjuvant therapy?

Combination Cyclophosphamide Methotrexate Fluorouracil Therapy

and Meta-analyses

In the early 1970s Bonadonna at the NationalCancer Institute in Milan evaluated CMF given 2weeks on and 2 weeks off for 12 months (“clas-sic CMF”).14 Three hundred and eighty-sixwomen with node-positive disease were random-ized to surgery ± CMF. As with the PAM trial,patients on CMF had a statisfically significantimprovement in DFS and a trend towardimproved overall survival. Again the benefit wasconfined to premenopausal women with no sig-nificant benefit for postmenopausal patients.14

Treatment with CMF (and its variants) becamethe standard adjuvant therapy for premenopausalwomen with node-positive disease (see Table12–1). Although there are several ways of admin-istering CMF (eg, classic monthly: po C ´ 14days + IV MF day 1 and 8; all IV every 3 weeks),several studies support the classic 28-day pro-gram as being the most effective.15,16 On theother hand, in a subsequent randomized study,Bonadonna found that 6 months of CMF to be asefficacious as 1 year of drugs.17 Six months of

CMF remains a standard adjuvant systemic ther-apy for early-stage breast cancer today.

Almost immediately after the adoption ofadjuvant chemotherapy for node-positive pre-menopausal women, several questions arose:was combination chemotherapy truly betterthan single-agent therapy? Is there really only adisease-free but not overall survival benefit toadjuvant chemotherapy? What is the optimalduration of therapy? Is it of any benefit in poor-risk, node-negative disease? Was it truly of novalue in postmenopausal women? Each of theseissues was evaluated in individual randomizedtrials. Nevertheless, it became increasingly dif-ficult, even with relatively large studies, todefinitively come up with answers, as resultswere often conflicting.

In 1985, 1990,18 and 1995,19 a consortium ofbreast cancer researchers, the Early Breast Can-cer Trialists’ Collaborative Group (EBCTCG)conducted meta-analyses, using primary datafrom multiple randomized trials of adjuvantchemotherapy to address these issues. The 1990overview looked at 13 studies (enrolling ~3,400women) comparing single-agent versus combi-nation chemotherapy.18 There was greater bene-fit with polychemotherapy, a relative 17 ± 5 per-cent decrease in yearly risk of death due tobreast cancer compared with single-agent ther-apy, reinforcing the widely held clinical impres-


Table 12–1. ADJUVANT SYSTEMIC THERAPY REGIMENS USED IN THE TREATMENT OF EARLY BREAST CANCER

Regimen Indication Frequency Drugs Dose

CMF Poor prognosis node – Every 28 days Cyclophosphamide 100 mg/M2 po qd ´ 14 dor 1 to 3 + nodes Methotrexate 40 mg/M2 IV dl and d8(erbB-2 –) ´ 6 months Fluorouracil 600 mg/M2 IV dl and d8

AC Poor prognosis node – Every 21 days Cyclophosphamide 600 mg/M2 IV dlor 1 to 3 + nodes Doxorubicin 60 mg/M2 IV dl(erbB-2 + or –) ´ 4 treatments

CAF 4 or more + nodes Every 28 days Cyclophosphamide 100 mg/M2 po qd ´ 14dDoxorubicin 30 mg/M2 IV dl and d8

´ 6 months Fluorouracil 600 mg/M2 IV dl and d8AC® Paclitaxel 4 or more + nodes AC given ´ 4 as above

Every 21 days Followed by´ 4 treatments Paclitaxel 175 mg/M2 3 h IV infusion

Tamoxifen Receptor Daily for 5 years 20 mg po daily(ER or PR)-containing tumors• Postmenopausal node ±• Premenopausal node –

sion. The 1995 overview, consequently, lookedonly at combination chemotherapy regimens, Itreviewed prolonged chemotherapy versus nochemotherapy in 47 trials encompassing 18,000women; longer versus shorter chemotherapy in6,100 patients in 11 trials; CMF versus anthra-cycline (doxorubicin or epirubicin)–based com-binations in 6,000 women in 11 trials.19 For allstudies, there was a statistically significant ben-efit in polychemotherapy versus no chemother-apy in terms of recurrence (relative decrease of24%, p < .00001) and also survival (relativedecrease in mortality of 15%, p < .00001).Chemotherapy was not just delaying recurrence.Comparisons of standard durations of adjuvantCMF-like regimens (6 months) with more pro-longed durations found no significant survivalbenefit to more prolonged therapy,19 confirmingBonadonna’s results.17 The benefits ofchemotherapy in the 1995 overview was seen inall nodal groups.19 In trials of combinationchemotherapy versus control, mortality was cutby the same relative amount in both node-posi-tive and node-negative patients. Given the dif-ferences in the risk of death in the two groups,the absolute benefit, however, was different. Inwomen under age 50 years, the absolutedecrease in death at 10 years was 12.4 percent(p < .00001) in the node-positive group and 5.7percent (p < .02) in the node-negative group.Finally, despite the many seemingly negative tri-als, polychemotherapy was found to be of valuein postmenopausal women but only in the 50- to69-year-old age group.19 For node-positivewomen over 50 years, chemotherapy decreasedthe rate of death from any cause by an absolute2.3 percent (p = .002), far less than in youngerwomen. For node-negative older women, how-ever, the absolute decrease in rate of death(6.4% p < .005) was comparable with that inyounger women. The paradox of chemotherapyhaving a greater absolute benefit in post-menopausal node-negative women than in node-positive ones is unexplained. As a result of themultiple studies that comprised the EBCTCGoverviews, by 1990, in premenopausal women,

adjuvant chemotherapy, particularly CMF,became standard for both node-positive andhigh-risk, node-negative breast cancers. In post-menopausal women, chemotherapy was reservedfor receptor-negative patients because of theperceived lesser benefit and increased toxicity.It was also being used for high-risk operablemale breast cancer.20

Doxorubicin

The suggestion that doxorubicin-based chemo-therapy was more effective than CMF in thepalliation of advanced breast cancer11 led to theuse of combinations containing the drug asadjuvant therapy particularly in poor prognosticgroups, such as those with multiple positivenodes. The 1995 overview analyzed 11 CMFversus anthracycline (doxorubicin or epiru-bicin) polychemotherapy trials.19 At 5 yearsafter surgery, anthracycline combinationsdecreased recurrence by an absolute 3.2 percent(p = .006) and mortality by an absolute 2.7 per-cent (p = .02). Individual trials themselves havebeen contradictory, but there are some generaltrends. The NSABP Protocol B15 showed thatin node-positive women, a short four-treatmentcourse of IV doxorubicin/cyclophosphamide(AC) (see Table 12–1) was equivalent to theeffect of classic CMF for 6 months.21 Further-more, the AC regimen was effective and welltolerated in postmenopausal women. It is nowwidely used as an alternative to CMF for poorprognosis node-negative and 1- to 3-node–pos-itive adjuvant therapy.21 Its popularity is basedon the briefer duration of treatment (3 versus 6months), less overall toxicity and lower risk ofpermanent menopause in younger women.Studies comparing CMF with CAF (doxoru-bicin substituted for methotrexate) have beenharder to interpret. They have suffered fromusing the inferior all-IV 3-weekly CMF in thecontrol arm or using epirubicin as their anthra-cycline drug. The Southwest Oncology Group(SWOG), however, did conduct a study of stan-dard CMF versus CAF (± tamoxifen) in node-

206 BREAST CANCER

negative women. Their preliminary report wasof a small but statistically significant 2 percentimprovement in survival (p = .03) for the CAFgroups compared with the CMF groups,22 sim-ilar to the 1995 meta-analysis results.19

Whether such a small benefit in a relativelygood prognosis group warrants the excess toxi-city is questionable. The difference, however,would be clinically significant in the high-riskmultinode-positive patient. If the final resultsof this study continue to show the difference, itwould support CAF ´ 6 as an alternative foradjuvant chemotherapy for women with four ormore positive lymph nodes.

Investigational Approaches

Unfortunately, even with aggressive doxoru-bicin-based chemotherapy, the prognosisremains poor for women with many axillarynodes involved.19 Another approach to improv-ing the results was the use of multiplenon–cross-resistant agents as adjuvant therapy.Studies looking at CMF variants alternatingwith doxorubicin regimens have not consis-tently shown benefit.21,23,24 The introduction ofthe taxanes, which are highly active and notcross-resistant with doxorubicin, offer greaterhope of improving outcomes. Several studiesare evaluating doxorubicin based regimens ±paclitaxel or docetaxel. The Cancer and AcuteLeukemia Group (CALGB) reported the pre-liminary results of CALGB 9344, which foundthat the addition of four doses of paclitaxel afterAC ´ 4 (see Table 12–1) in node-positivewomen decreased recurrence rate by 22 percentand death by 26 percent. The absolute decreasein mortality at 18 months was 2 percent(p = .039).25 While it is too soon to state that AC´ 4 ® paclitaxel ´ 4 is the standard therapy fornode-positive disease (or if it is better than CAF´ 6), the data suggest a role for the taxanes. It isreasonable to consider AC® paclitaxel as analternative to CAF in high-risk situations wherethe toxicity of prolonged doxorubicin adminis-tration is a concern (Table 12–2).

The assumption that there is a dose-response relationship to tumor cell kill and thedevelopment of cytokine and stem cell bonemarrow support has lead to a series of studieslooking at higher-or more-intense-dose adju-vant chemotherapy regimens. These studies areof three types. Some escalate drugs two to fourtimes the conventional dosage and use cytokinesupport (escalated conventional dose). Othersescalate drug doses minimally but give drugs atbriefer intervals (dose-dense therapy). Thereare few studies that address this concept. Thethird approach uses massive (5- to 10-fold)dose escalation and requires stem cell support(“bone marrow transplant”). Only for the esca-lated conventional dose approach are finalresults of randomized trials available. They arevery disappointing. The NSABP trials B-2226

and B-2527 randomized node-positive patientsto conventional AC ´ 4 or to regimens contain-ing higher doses of cyclophosphamide (up tofour-fold escalation with GCSF). There was nodisease-free or overall survival benefit to thehigher-dose schedules. The CALGB 9344 trialrandomized conventional AC ´ 4 (± paclitaxel)versus AC ´ 4 (± paclitaxel) with the doxoru-bicin escalated by 25 or 50 percent. There wasno benefit to the higher-dose doxorubicin regi-mens compared with those containing conven-tional AC.25 To date, there is no convincing evi-dence to support escalated standard-doseadjuvant chemotherapy.

While high-dose chemotherapy with stemcell transplant is widely employed in some cen-ters as adjuvant therapy for women with 10 ormore involved nodes, data to support it areincomplete. Proponents of the approach citevastly superior survival results in womenundergoing the technique compared with his-torical controls receiving conventional chemo-therapy. Critics believe the superiority com-pared with historical controls is due to patientselection and cite a small randomized studythat found no benefit in the approach.28 TheEastern Cooperative Oncology Group (ECOG)has completed a study in women with 10 or


more positive nodes of CAF ´ 6 versus CAF´ 6 followed by high-dose chemotherapy withstem cell reinfusion. The CALGB trial is alsoconducting a study in 10 or more node-positivepatients. They are randomized to CAF followedby low-dose consolidation chemotherapy or bythe same drugs given in high doses with stemcell support. For women with four to ninenodes involved, SWOG is conducting a study ofsequential high conventional dosedoxorubicin® paclitaxel® cyclophosphamidetherapy versus AC ´ 4 followed by high-dosechemotherapy with stem cells. The latter trial,unfortunately, suffers from not having a trulystandard control arm. Hopefully, when thelongterm results of these large studies arereported, we will know what role (if any) high-dose chemotherapy with stem cell support hasin the adjuvant treatment of early breast cancer.

HORMONAL THERAPY

Tamoxifen

Multiple events in the late 1970s led to thedevelopment of tamoxifen as the most widelyprescribed drug for the adjuvant systemic ther-apy of early breast cancer. Tamoxifen wasshown to be an effective and safe therapy of

metastatic breast cancer in postmenopausalwomen.29 Estrogen receptor could be used topredict response of metastatic disease.8 Initialreports of randomized trials of adjuvantchemotherapy in early breast cancer found min-imal benefit in postmenopausal women com-pared with younger women.13,14 Trials in theUnited Kingdom and Scandinavia began look-ing at adjuvant tamoxifen after surgery in node-positive postmenopausal breast cancer patients.In 1983, Baum reported preliminary results ofone of the studies. Tamoxifen given aftersurgery decreased recurrence and increasedsurvival.30 Since then over 50 randomized trialshave been conducted looking at the role ofadjuvant tamoxifen in 37,000 women. As withchemotherapy, the EBCTCG conducted meta-analysis overviews of the tamoxifen trials, mostrecently in 1995.31 Tamoxifen was beneficial innode-positive and node-negative disease, inpostmenopausal and premenopausal patients.Although any duration of therapy was benefi-cial, longer durations were more beneficialthan briefer duration. Only for women withtumors not containing ER was there no bene-fit.31 For all women given tamoxifen, there wasa 26 percent relative decrease in recurrence, a14 percent relative decrease in death, and, at 10years after surgery, an absolute decrease of 3.7

208 BREAST CANCER

Table 12–2. GUIDELINES FOR THE CHOICE OF ADJUVANT SYSTEMIC THERAPY

Node-Tumor-Receptor Status Premenopausal Postmenopausal Over Age 70 Years

Axillary node negative: tumor < 1 cm No Rx*‡ No Rx*‡ No Rx*‡

Axillary node negative: tumor > 1 cm or Tamoxifen or CMF† or Tamoxifen or TamoxifenPoor prognosis receptor positive AC or tamoxifen + AC tamoxifen + AC*

Axillary node negative: tumor > 1 cm or CMF† or AC AC or CMF† Chemo Rx on Poor prognosis receptor negative‡ individual basis*

Axillary node positive: 1 to 3 + nodes CMF† or AC ± Tam Tamoxifen ± AC* TamoxifenReceptor positive (or CMF†)

Axillary node positive: 1 to 3 + nodes CMF† or AC AC (or CMF†) Chemo Rx on Receptor negative‡ individual basis

Axillary node positive: ³ 4 + nodes CAF or AC® paclitaxel Tamoxifen + CAF Tamoxifen (± chemo Rx Receptor positive ± tamoxifen or AC® paclitaxel on individual basis)

Axillary node positive: ³ 4 + nodes CAF or AC® paclitaxel CAF or Chemo Rx on Receptor negative‡ AC® paclitaxel individual basis

* If multiple prognostic markers adverse consider Rx† If c-erbB-2 overexpressed doxorubicin based chemotherapy should be used‡ Consider Tamoxifen as chemopreventive agentRx = therapy; CMF = cyclophosphamide, methotrexate, fluorouracil; AC = adriamycin, cyclophosphamide; CAF = cyclophosphamide, adriamycin,fluorouracil.

percent in death from any cause, comparedwith women not getting the drug (p < .00001).31

For women with tumors containing hormonereceptor and receiving 5 years of tamoxifen, theabsolute decrease in death at 10 years was 5.6percent for node-negative tumors and 10.9 per-cent for node-positive tumors.31 In the 1990s,tamoxifen became standard adjuvant therapyfor all postmenopausal women with node-posi-tive and poor-prognosis, node-negative breastcancers containing hormone receptor.

Several controversies persist in the use ofadjuvant tamoxifen. One is the duration of ther-apy; the other is its role in premenopausalwomen. While the meta-analysis and individualrandomized trials favor 5 years of therapy,31 twostudies of 5 years versus more than 5 years inpredominantly node-negative women found nobenefit to additional years of therapy.32,33 TheECOG, however, found that more than 5 yearsof tamoxifen added to adjuvant chemotherapy inwomen with receptor containing tumors wasmore beneficial than only 5 years of the drug.34

In general, the standard approach is to givenode-negative women 5 years of tamoxifen; fornode-positive women, 5 years is suggested, butindividualizing duration of therapy on the basisof risk of recurrence and toxicity is widely done.The Oxford Group is conducting a trial(ATLAS), which is directly addressing the opti-mal duration of adjuvant tamoxifen therapy.

In the United States, chemotherapy is thepredominant form of adjuvant systemic therapyused in premenopausal women. This is not sur-prising since many younger women have breastcancers that do not contain hormone receptorand are unlikely to benefit from adjuvanttamoxifen. Yet NSABP trial B-14 found tamox-ifen to be an effective therapy in premeno-pausal axillary node-negative women.35 Themeta-analysis looking at 5 years of tamoxifenfound no significant difference in the benefit oftamoxifen between younger and olderpatients.31 This emphasizes that in hormonereceptor-positive breast cancer (particularlynode-negative disease), tamoxifen can be con-

sidered an alternative to chemotherapy in pre-menopausal women. Tamoxifen is also of valuein receptor-positive male breast cancer.36

Oophorectomy and GonadotropinReleasing Hormone Analogue

The sporadic use of adjuvant oophorectomy afterbreast cancer surgery in younger women wascontinued by surgeons for many years in thehope of preventing recurrence. Randomized tri-als looking at its value date back 50 years.37

Unfortunately, these early trials suffer from thelack of hormone responsiveness of most breastcancers in premenopausal women and they pre-date our ability to predict responsiveness withhormone receptor measurements. The EBCTCGconducted overview meta-analyses of adjuvantoophorectomy in 1985, 1990, and 1995. Themost recent overview encompassed 12 trials ran-domizing 2,100 women to surgical or radiationoophorectomy versus no castration.38 In womenunder the age of 50 years, oophorectomyresulted in an 18 percent relative decrease inrecurrence, an 18 percent relative decrease indeath, and, at 15 years after surgery, an absolutedecrease of 6.3 percent in death from any cause,compared with women not getting the procedure(p < .001).38 The relative benefit was the same innode-negative and node-positive patients.38

These results are very similar to the chemother-apy meta-analysis results. Furthermore, in aScottish trial, adjuvant CMF was compared withoophorectomy in premenopausal women withnode-positive disease.39 There was no differenceseen in the overall result. In women with recep-tor-positive tumors, the trend was the superiorityof castration; in receptor-negative disease CMFappeared better.

In the United States, adjuvant surgical cas-tration is rarely done these days, with tamox-ifen the preferred adjuvant hormonal approachin younger women (if any hormonal approachis used). The introduction of gonadotropin-releasing hormone (GnRH) analog has thepotential to change this practice. These drugs,


when given by slow-release depot injection,continually stimulate the pituitary, eventuallydepleting it of FSH and LH. This results in thecessation of ovarian function, achieving a bio-chemical oophorectomy.7 The GnRH analog,goserelin (Zoladex) and leuprolide (Lupron)have antitumor activity comparable withoophorectomy and with tamoxifen againstovert hormone-responsive metastatic diseasein premenopausal women.40 Furthermore, theiraction on the ovary is reversible. Currently, theEuropean “ZEBRA” study is comparing adju-vant goserelin with chemotherapy in pre-menopausal women. Despite its proven effi-cacy as adjuvant therapy in younger women,oophorectomy is unlikely to be widelyaccepted in the United States.

Other Hormonal Approaches

Several other hormonal therapies have activityagainst metastatic breast cancer and have beenevaluated as adjuvant therapy in early disease.Toremifene (Fareston) is a derivative of tamox-ifen with a similar mechanism of action andactivity against disseminated disease.7 It isbeing evaluated in randomized trials againsttamoxifen as adjuvant therapy in older women.Progestins lower endogenous estrogen levels inpostmenopausal women and cause tumorregression in many women with advanced dis-ease.7 Medroxyprogesterone has been studiedas adjuvant therapy in randomized trials, withnegative results.41,42 Aromatase inhibitors (AI)inhibit the enzyme that catalyzes the conver-sion of androgen to estrogen. They, too, lowerserum (and intracellular) estrogen levels inolder women and are effective hormonal thera-pies of metastatic breast cancer.7 Aminog-lutethimide, one of the first-generation aro-matase inhibitors, however, was no better thanplacebo in an adjuvant trial after surgery inpostmenopausal women.43 New classes ofmore potent and selective aromatase inhibitorshave been recently introduced for the treatmentof advanced disease and are being evaluated as

adjuvant therapy. The ATAC trial is a multina-tional randomized double-blinded study inpostmenopausal women of adjuvant tamoxifenversus the new AI anastrozole (Arimidex) ver-sus the combination for 5 years. Another triallooks at women that have received 5 years oftamoxifen and are then being randomized to nofurther therapy or treatment with the AI letro-zole (Femara). The results of the toremifene,arimidex, and letrozole trials and a Europeantrial of the AI formestane are not yet known.To date, other than tamoxifen and oophorec-tomy, there are no standard hormonal adjuvanttherapies.

COMBINED CHEMOHORMONAL THERAPY

The 1995 EBCTCG overviews looked at therelative benefits of adjuvant combined chemo-hormonal therapy versus single-modality treat-ment.19,31,38 There was a suggestion that inwomen aged 50 to 69 years, tamoxifen pluschemotherapy decreased the annual risk ofdeath by 10 percent compared with tamoxifenalone.31 The issue was prospectively studied innewer trials. The NSABP, SWOG, and theInternational Breast Cancer Study Group, eachfound benefit in their studies of combined ther-apy versus tamoxifen alone in postmenopausalwomen;44–46 the National Cancer Institute(NCI) of Canada did not.47 While it is prema-ture to suggest that all postmenopausal womenwith receptor-positive cancer should receivechemotherapy and tamoxifen; certainly, it isappropriate in selected high-risk women underthe age of 70 years. The best way to combinethe two, simultaneously or sequentially, stillremains unresolved.

For women under the age 50 years, the 1995overview suggested that the addition ofoophorectomy to adjuvant chemotherapy wasof borderline benefit, with a nonstatisticallysignificant decrease of 10 percent in the annualmortality.38 Furthermore, the ECOG recentlyreported the preliminary results of a random-

210 BREAST CANCER

ized study in premenopausal women with hor-mone-sensitive, node-positive breast cancer.Women were randomized to receive CAF ´ 6 orCAF ´ 6 + 2 years of goserelin or CAF ´ 6 +goserelin + tamoxifen. The results are prelimi-nary. It seems unlikely that adding goserelin toadjuvant chemotherapy improves results inpremenopausal women in whom the chemo-therapy already achieved a chemical castrationbut may be of value in younger women who arestill menstruating. Unfortunately, the meta-analyses did not address these issues.31 Theaddition of tamoxifen in the ECOG trial seemedbeneficial in older women in whom chemother-apy led to menopause. Although not conclu-sively proven effective, chemotherapy plustamoxifen is often given to premenopausalwomen with receptor-positive early breast can-cer, usually because the chemotherapy has ren-dered them menopausal or more recently forchemopreventive reasons (vide infra).

NEOADJUVANT THERAPY

Increasingly, lumpectomy plus radiation hasbecome the desired standard of local therapyfor early breast cancer. Unfortunately, tumor orbreast size in many women make lumpectomycosmetically or technically not feasible. Thegratifying results with initial chemotherapy inthe treatment of locally inoperable breast can-cer (such as inflammatory carcinoma) led totrials of preoperative chemotherapy in the hopeof shrinking operable but large tumors to thepoint that breast conservation could be accom-plished. Bonadonna’s group in Milan, in twotrials using several preoperative chemotherapyregimens, was able to perform breast preserva-tion surgery (quadrentectomy) in 66 percent ofwomen with tumors > 5 cm.48 No one preoper-ative chemotherapy regimen was clearly supe-rior.48 Building on these findings, the NSABPinvestigated whether preoperative chemother-apy, besides shrinking the primary tumor,might also be more effective as an adjuvantsystemic therapy than postoperative chemother-

apy. The NSABP trial B-18 randomized womento receive four cycles of preoperative AC orfour cycles of postoperative therapy.49,50 Inwomen with tumors > 5 cm in diameter, therewas a near-doubling of the breast conservationrate with neoadjuvant chemotherapy.50 Never-theless, there was no difference in overall dis-tant recurrence and survival.49 A new NSABPtrial (B-27) is asking the same questions butlooking at AC ± docetaxel given in variousneoadjuvant and/or adjuvant combinations. Inpatients with large hormone receptor-positivebreast cancers, neoadjuvant tamoxifen hasbeen shown to be as effective as chemotherapyin shrinking the tumor to facilitate breast con-servation.51 It should be considered in patientsthat are not candidates for chemotherapybecause of age or infirmity. For now, neoadju-vant therapies remain an effective way to facil-itate breast conservation but not to improvesurvival over that achieved with postoperativeadjuvant treatment.


Although the primary role of adjuvant systemictherapy is to treat occult distant disease, theremay also be a local benefit, at least in ductalcarcinoma in situ trial (DCIS). The NSABPtrial B-24 randomized 1,804 women with DCIStreated with lumpectomy plus radiation to nofurther therapy or to 5 years of tamoxifen.52 At5 years, tamoxifen decreased the risk of thedevelopment of invasive breast cancer in thetreated breast by 47 percent (2.1% versus 3.4%in control, p = .04) and of all breast cancerevents (ipsilateral and contralateral) by 34 per-cent (8.8% versus 13% in control, p = .007).52

These results apply only to women who hadlumpectomy/radiation for DCIS. They are notrelevant to women treated with mastectomy.While the absolute magnitude of the benefitwas small, certainly tamoxifen should be con-sidered in many women with DCIS treated withbreast conservation, particularly those withhigh-risk pathology.


THE CHOICE OF ADJUVANT THERAPYAND THE BIOLOGY OF THE TUMOR

AND OF THE PATIENT

Despite the successes of adjuvant chemother-apy and hormonal therapy, many patients thatreceive such treatments have recurrences anddie. To optimize results, when choosing treat-ment, tumor and patient biology must be takeninto account. This is most apparent in the use ofhormonal therapy. Although there is a small butreal response rate with tamoxifen in metastatichormone receptor-negative breast cancer, the1995 meta-analysis found no significant bene-fit to adjuvant tamoxifen in women with recep-tor-poor tumors.31 It is, therefore, critical thathormone receptor be measured in any patientwith newly diagnosed breast cancer.4 Althoughthere was some controversy in the past as towhether hormone receptor status is also ofvalue in predicting response to chemotherapy, itis now known that no such relationship exists.19

On the other hand, there is preliminary datathat suggest the presence of an overexpressedc-erbB-2 oncogene in breast cancers may havepredictive value in the choice of chemother-apy.53,54 The NSABP trial found that melphalanwith fluorouracil + doxorubicin was more effec-tive adjuvant therapy than melphalan with fluo-rouracil alone. However, when the data werereanalyzed, the benefit was only seen in patientswhose tumors overexpressed c-erbB-2.53 TheCALGB trial randomized women to three dif-ferent dose-intense CAF regimens. The twohigher-dose regimens (both within the range ofstandard dosage) were superior to the low-doseregimen; however, an analysis of the data byc-erbB-2 status found the difference to be pre-sent only in the subset of oncogene overexpres-sor.54 These data suggest that if adjuvantchemotherapy is to be given in the presence ofc-erbB-2 overexpression, it should contain dox-orubicin at full dose. What is not clear iswhether lack of c-erbB-2 overexpression pre-dicts the effectiveness of nondoxorubucin-con-taining regimens. While there are suggestions

that c-erbB-2 overexpression may predict unre-sponsiveness to adjuvant tamoxifen,55,56 recentstudies are not supportive.57,58

Though the biology of the tumor is impor-tant, so too is the biology of the patient in thechoice of adjuvant systemic therapy. Age is themost important factor. There is little data onthe efficacy of chemotherapy in patients overthe age of 70 years; of the 19,000 reviewed inthe EBCTCG chemotherapy overview, therewere only 600 women over 70 years. Theycould draw no conclusion as to the value ofchemotherapy in that age group.19 Chemother-apy should be reserved for women over 70years with poor prognostic tumors containingno hormone receptor, with a reasonable lifeexpectancy, and that are physiologically inexcellent health. It is important to emphasize,however, that being over 70 years should not apriori exclude a woman from consideration ofchemotherapy. In the younger postmenopausalgroups, chemotherapy is clearly beneficialalone or when added to tamoxifen. Neverthe-less, it should not be universally given. TheEBCTCG trial analyzed the same randomizedtrials where they found a survival benefit tochemotherapy plus tamoxifen in women 50 to69 years but found no increase in “quality (oflife)–adjusted survival” compared with tamox-ifen alone.59 The implication is not thatchemotherapy should not be given to womenage 50 to 69 years but rather that its use inaddition to tamoxifen should be limited tohigh-risk, poor-prognosis patients, despitecalls to the contrary. The administration ofadjuvant systemic therapy cannot be consid-ered a standardized process. It must always beindividualized.

BENEFITS VERSUS TOXICITY ANDRISKS OF THERAPY

The acute toxicities of adjuvant systemic ther-apy of early breast cancer are significant butgenerally well tolerated and are easily justifiedgiven the potential benefit. The toxicities of

212 BREAST CANCER

chemotherapy can be divided into those of theCMF-like regimens and those of the doxoru-bicin regimens. All chemotherapies used asadjuvant treatment cause significant myelosup-pression, with leukopenia generally clinicallymore significant than anemia or thrombocy-topenia. In the NSABP trials of classic CMF´ 6, the incidence of neutropenia less than2,000 was ~ 10 percent and severe infectionabout 1 percent.21 With AC ´ 4, it is 4 percentsevere neutropenia and 2 percent severe infec-tion.21 With 6 months of CAF, the risk ofleukopenia and infection is higher. Thrombocy-topenia is seen in less than 1 percent of patientsin most regimens.21 Doxorubicin-containingregimens are more emetogenic than CMF; how-ever, the incidence of severe vomiting is rapidlydropping with the introduction of serotoninantagonists. Alopecia is nearly universal withdoxorubicin and is seen in about 40 percent ofCMF patients.21 Diarrhea is rarely seen witheither regimens; the use of serotonin antagonistantiemetics is associated with constipation (andmild headache). Cystitis is seen in about 1 per-cent of patients receiving cyclophosphamide-containing regimens and correlates with longerdurations of therapy.21 Other rare side effects ofboth regimens include mucositis, thromboem-bolic events, and, for doxorubicin, extravasa-tion skin ulceration.

The most common chronic chemotherapytoxicity is the cessation of menses and induc-tion of menopause in premenopausal women.This is more common with 6 months of CMF(and CAF) than with AC ´ 4. In one study,amenorrhea was seen in 68 percent of womenon CMF and 34 percent of women on AC.60

Symptomatic cardiomyopathy is a rare compli-cation seen with doxorubicin-containing regi-mens. The risk is less than 1 percent withcumulative doxorubicin doses less than 350mg/M2.61 The cumulative dose with AC ´ 4 is240/m2; with CAF, it is 360/m2. The risk isincreased with age, left chest wall irradiation,and prior heart disease. Chemotherapy agentsare carinogenic in experimental systems. Nev-

ertheless, the incidence of second malignancieshas been low. The ECOG estimated the risk ofsecondary leukemia or myelodysplasia after itsCMF adjuvant regimens to be less than 0.2 per-cent similar to that of the general population.62

Bonadonna could find no increased risk ofmalignancy in long-term follow-up of his adju-vant CMF patients.63 The M.D. Anderson Hos-pital, in reviewing its adjuvant doxorubicin pro-grams, found the risk of secondary leukemia ormyelodysplasia to be 0.2 to 0.5 percent.64 Thereis some suggestion that concomitant high-dosecyclophosphamide may increase the doxoru-bicin leukemia risk.65 Other than menopause inyounger women, long-term complications ofadjuvant chemotherapy are infrequent.

Tamoxifen is a selective estrogen receptormodular (SERM) and so may be antiestrogenicor estrogenic, depending on its interaction withthe individual tissue receptor. Its toxicity pro-file reflects this duality. The most commonacute tamoxifen side effects are menopausalsymptoms. In the NSABP trial B-14, hotflashes were seen in about two-thirds ofpatients, about a third had weight gain, fluidretention, and vaginal discharge, and a quarterexperienced nausea, and weight loss.35 Irregu-lar menses were seen in a fourth of pre-menopausal women.35 The only significantacute toxicities were rare thromboembolicevents: deep vein thrombosis in 0.8 percent andpulmonary embolus in 0.4 percent. Moodswings and depression are unusual. Very-high-dose tamoxifen may cause retinal changes, butthese are rarely seen with conventional doses.There are reports of cataracts in patients on thedrug.66 In a large review of ocular toxicity fromthe NSABP, there were no cases of vision-threatening eye toxicity with tamoxifen.66

There is an increased risk of developinguterine cancer in women receiving tamoxifen(2/1,000/y of therapy versus 1/1,000/y in con-trol).67 In the 1995 meta-analysis, 10 years afterbreast surgery, women on 5 years of the drughad a 1.1 percent risk of uterine malignancycompared with 0.3 percent for those who did


not receive the drug.31 Furthermore, mostreported cases were early stage and highly cur-able,68,69 although fatal cases of uterine cancerhave been reported.70 In contrast to the uterinecancer effects are thoses of tamoxifen on thedevelopment of contralateral breast cancer.Multiple studies have found that tamoxifengiven to prevent recurrence of previous breastcancer significantly decreases the risk of devel-oping new contralateral breast cancer.68 The1995 overview found a 47 percent decrease inthe risk of contralateral malignancy in womenreceiving 5 years of the drug.31 This effect oftamoxifen was confirmed by the Breast CancerPrevention Trial (NSABP P-1), which found avirtually identical decrease in the developmentof new breast cancers in high risk women with-out a history of the disease.71 This effect wouldsuggest that even in women with a history ofreceptor-negative breast cancer, tamoxifenmight be considered not to prevent recurrencebut to decrease the risk of new malignancy.

Other beneficial effects of adjuvant tamox-ifen include an estrogenic-like decrease in boneloss in postmenopausal women,72–74 decrease incholesterol,75 and, in some studies, decreasedcardiac mortality.76–78 Overall, the risk/benefitratio strongly favors tamoxifen’s use as adju-vant therapy for hormone receptor-containingearly-stage breast cancer.

SUMMARY

The addition of adjuvant chemotherapy or hor-monal therapies to local treatment signifi-cantly decreases recurrence and mortality inwomen with axillary node-positive or high-risk, node-negative, operable breast cancer.The choice of therapy should be individualizedon the basis of the perceived risk of recurrence,particularly as determined by nodal status andtumor size, patient age and general health, andthe presence or absence of hormone (estrogenor progesterone) receptor in the tumor. Somegeneral guidelines are reasonable (see Table12–2). For all postmenopausal women with

receptor-containing tumor warranting adjuvanttherapy (node + or –), tamoxifen should begiven for 5 years. If at particularly high risk,such as with multiple positive nodes or anextremely large tumor, the addition ofchemotherapy (AC ´ 4) should be consideredif the patient is in good health and has a rea-sonable life expectancy. For receptor-negativepostmenopausal women with node-positive orhigh-risk, node-negative disease, chemother-apy (CMF or AC ´ 4 or for multiple nodesAC ´ 4® paclitaxel ´ 4 or CAF) should beadministered. The borderline efficacy of CMFin postmenopausal women make doxorubicin-containing regimens preferable if cardiac func-tion permits. Chemotherapy should be usedcautiously in women over 70 years. In theUnited States, for premenopausal women withpositive nodes, adjuvant chemotherapy is stan-dard. For patients with 1 to 3 positive nodes,AC or CMF is widely used. For women withfour or more involved nodes, CAF ´ 6 is thestandard. Recent data suggest that the additionof taxanes to doxorubicin regimens also has arole in this group (AC ´ 4® paclitaxel ´ 4) andmay supplant CAF.25 If the patient’s tumor isreceptor positive, tamoxifen is frequentlyadded to chemotherapy, but the data to supportthis are limited. Nevertheless, if the chemother-apy renders a woman menopausal or there isgreat concern about contralateral breast can-cer, few would fault the addition of tamoxifento chemotherapy for node-positive, premeno-pausal, receptor-positive, breast cancer adju-vant therapy. In Europe, such women mightundergo oophorectomy as an alternative tochemotherapy ± tamoxifen.39 For high-risk,node-negative premenopausal women, thereare several alternatives. If the tumor is hor-mone receptor positive, tamoxifen, CMF, orAC (or chemotherapy + tamoxifen) are accept-able. If receptor negative, chemotherapy withCMF or AC is used. Although the SWOG trialsuggested the superiority of CAF in this groupof patients,22 the added toxicity is likely to limitthe use of this combination in node-negative

214 BREAST CANCER

women. The use of high-dose chemotherapywith stem cell support should be consideredinvestigational, even for patients with 10 ormore involved nodes. Neoadjuvant therapiesare appropriate in an attempt to achieve breastconservation but have not been proved a moreeffective adjuvant systemic therapy than post-operative treatment.49 Delay of local treat-ment (surgery and/or radiation) for theadministration of neoadjuvant49 or adjuvanttherapy79 does not negatively impact out-come. For women with high-risk DCIS treatedwith lumpectomy plus radiation, tamoxifentherapy should be considered on the basis of the preliminary results of NSABP trial B-24.52 Whenever possible, women withearly-stage breast cancer should be encour-aged to enroll in RCTs of adjuvant therapies.Only with such studies will further progressbe made in reducing mortality in women withbreast cancer.

REFERENCES

1. Contesso H, Mouriesse H, Friedman S, et al. Theimportance of histologic grade in long-termprognosis of breast cancer: a study of 1,010patients, uniformly treated at the Institut Gus-tave-Roussy. J Clin Oncol 1987;5;1378–6.

2. Fisher ER, Gregorio RM, Fisher B, et al. Thepathology of invasive breast cancer. A syllabusderived from findings of the National AdjuvantBreast Project. Cancer 1975;36:1–85.

3. Donegan WL. Tumor-related prognostic factors forbreast cancer. CA Cancer J Clin 1997;47:28–51.

4. Tumor Marker Guideline Panel, American SocietyClinical Oncology. Clinical practice guidelinesfor the use of tumor markers in breast and col-orectal cancer. J Clin Oncol 1996;14:2843–77.

5. Early Breast Cancer Trialists’ Collaborative Group.Effects of radiotherapy and surgery in earlybreast cancer. An overview of the randomizedtrials. N Engl J Med 1995;333:1444–55.

6. Beatson GT. On the treatment of inoperable casesof carcinoma of the mamma: suggestion for anew method of treatment with illustrativecases. Lancet 1996;2:104–7.

7. Locker GY. Hormonal therapy of breast cancer.Cancer Treat Rev 1998;24:221–40.

8. Desombre ER, Carbone PP, Jensen EV, et al. Spe-cial Report. Steroid receptors in breast cancer.N Engl J Med 1979;301:1011–2.

9. Hoogstraten B, Fabian C. A reappraisal of singledrugs for advanced breast cancer. Cancer ClinTrials 1979;2:101–98.

10. Cooper RG. Combination chemotherapy in hor-mone resistant breast cancer. Proc Am AssocCancer Res 1963;10:15.

11. Bull JM, Tormey DC, Li SH, et al. A randomizedcomparative trial of adriamycin versus metho-trexate in combination drug therapy. Cancer1978;41:1649–57.

12. Fisher B, Ravdin RG, Ausman RK, et al. Surgicaladjuvant chemotherapy in cancer of the breast:results of a decade of cooperative investiga-tions. Ann Surg 1968;168:337–56.

13. Fisher B, Carbone P, Economou SG, et al. L-Pheny-lalanine mustard (L-PAM) in the managementof primary breast cancer. A report of early find-ings. N Engl J Med 1975;292:117–22.

14. Bonadonna G, Brusamolino E, Valagussa P, et al.Combination chemotherapy as an adjuvanttreatment in operable breast cancer. N Engl JMed 1976;294:405–10.

15. Engelsman E, Rubens RD, Klign JGM. Compari-son of the classical CMF with a three weeklyintravenous CMF schedule in postmenopausalpatients with advanced breast cancer: anEORTC study. Proceedings of the 4th EORTCBreast cancer Working Conference 1987;1:1.

16. Goldhirsch A, Colleoni M, Coates AS, et al.Adding adjuvant CMF chemotherapy to eitherradiotherapy or tamoxifen: are all CMF’s alike?the International Breast Cancer Study Group(IBCSG). Ann Oncol 1998;9:489–93.

17. Tancini G, Bonadonna G, Valgussa P, et al. Adju-vant CMF in breast cancer: comparative 5 yearresults of 12 versus 6 cycles. J Clin Oncol1983;1:2–10.

18. Early Breast Cancer Trialists’ CollaborativeGroup. Systemic treatment of early breast can-cer by hormonal, cytotoxic, or immune ther-apy: 133 randomized trials involving 31,000recurrences and 24,000 deaths among 75,000women. Lancet 1992;339:1–15, 71–85.

19. Early Breast Cancer Trialists’ CollaborativeGroup. Polychemotherapy for early breast can-cer: an overview of the randomized trials.Lancet 1998;352:930–42.

20. Bagley CS, Wesley MN, Young RC, Lipmann ME.Adjuvant chemotherapy in males with cancerof the breast. Am J Clin Oncol 1987;6:45–50.


21. Fisher B, Brown AM, Dimitrov NV, et al. Twomonths of doxorubicin-cyclophosphamidewith and without interval reinduction therapycompared with 6 months of cyclophosphamide,methotrexate, and fluorouracil in positive-nodebreast cancer patients with tamoxifen-nonre-sponsive tumors: results from the NSABP B-15.J Clin Oncol 1990;8:1483–96.

22. Hutchins L, Green S, Ravdin P, et al. CMF versusCAF with and without tamoxifen in high risknode-negative breast cancer patients and a nat-ural history follow-up study in low-risk nodenegative patients: first results of IntergroupTrial INT 102. Proc Am Soc Clin Oncol1998;17:1a.

23. Moliterni A, Bonadonna G, Valagussa P, et al.Cyclophosphamide, methotrexate and fluo-rouracil with and without doxorubicin in theadjuvant treatment of resectable breast cancerwith one to three positive axillary nodes. J ClinOncol 1991;9:1124–30.

24. Tormey DC, Gray R, Abeloff MD, et al. Adjuvanttherapy with a doxorubicin regimen and longterm tamoxifen in premonopausal breast can-cer patients: an ECOG trial. J Clin Oncol 19921992;10:1848–56.

25. Henderson IC, Berry D, Demetri G, et al.Improved disease-free and overall survivalfrom the addition of sequential paclitaxel butnot from the escalation of doxorubicin dose inthe adjuvant chemotherapy of patients withnode-positive primary breast cancer. Proc AmSoc Clin Oncol 1998;17:101a.

26. Fisher B, Anderson S, Wickerham DL, et al.Increased intensification and total dose ofcyclophosphamide in a doxorubucin-cyclo-phosphamide regimen for the treatment of pri-mary breast cancer: findings from NationalSurgical Adjuvant Breast and Bowel Project B-22. J Clin Oncol 1997;15:1858–69.

27. Wolmark N, Fisher B, Anderson S. The effect ofincreasing dose intensity and cumulative doseof adjuvant cyclophosphamide in node positivebreast cancer: results of NSABP B-25. BreastCancer Res Treat 1997;46:26.

28. Rodenhuis S, Richel DJ, van der Wall E, et al. Arandomized trial of high-dose chemotherapyand haematopoetic progenitor-cell support inoperable breast cancer with extensive axillarynode involvement. Lancet 1998;352:515–21.

29. Mouridsen H, Palshof T. Tamoxifen in advancedbreast cancer. Cancer Treat Rev 1978;5:131–41.

30. Baum M. Brinkley DM, Dosset JA, et al. Improvedsurvival among patients treated with adjuvant

tamoxifen after mastectomy for early breastcancer [letter]. Lancet 1983;2(8347):450.

31. Early Breast Cancer Trialists’ CollaborativeGroup. Tamoxifen for early breast cancer: anoverview of the randomized trials. Lancet1998;351:1451–67.

32. Fisher B, Dignam J, Bryant J, et al. Five versusmore than five years of tamoxifen therapy forbreast cancer patients with negative lymphnodes and estrogen receptor-positive tumors. JNatl Cancer Inst 1996;88:1529–42.

33. Stewart HJ, Forrest AP, Everington D, et al. Ran-domized comparison of 5 years of adjuvanttamoxifen with continuous therapy for oper-able breast cancer. The Scottish Cancer TrialsBreast Group. Br J Cancer 1996;74:297–9.

34. Tormey DC, Gray R, Falkson G. Postchemother-apy adjuvant tamoxifen therapy beyond fiveyears in patients with lymph node-positivebreast cancer. Eastern Cooperative OncologyGroup. J Natl Cancer Inst 1996;88:1828–33.

35. Fisher B, Costantino J, Redmond C. A random-ized clinical trial evaluating tamoxifen in thetreatment of patients with node negative breastcancer who have estrogen receptor positivetumors. N Engl J Med 1989;320:479–84.

36. Ribeiro G, Swindell R. Adjuvant tamoxifen formale breast cancer. Br J Cancer 1992;65:252–4.

37. Cole MP. A clinical trial of an artificial meno-pause in carcinoma of the breast. INSERM1975;55:143.

38. Early Breast Cancer Trialists’ CollaborativeGroup. Ovarian ablation in early breast cancer:an overview of randomized trials. Lancet1996;348:1189–96.

39. Scottish Cancer Trials Breast Group and ICRFBreast Unit. Adjuvant ovarian ablation versusCMF chemotherapy in premenopausal womenwith pathological stage II breast carcinoma: theScottish trial. Lancet 1993;341:1293–8.

40. Davidson NE. Ovarian ablation as treatment foryoung women with breast cancer. J Natl Can-cer Inst Monographs 1994;16:95–9.

41. Focan C, Beaudin M, Salamon E. Adjuvant highdose medroxyprogesterone for early breastcancer: 13 years update of a multicenter ran-domized trial. Eur J Oncol 1998;(Suppl 1):34.

42. Pannuti F, Martoni A, Cilenti G, et al. Adjuvanttherapy for operable breast cancer withmedroxyprogesterone acetate alone in post-menopausal patients or in combination withCMF in premenopausal patients. Eur J CancerClin Oncol 1988;24:423–9.

43. Jones AL, Powles TJ, Law M, et al. Adjuvant

216 BREAST CANCER

aminoglutethimide for postmenopausal patientswith primary breast cancer: analysis at 8 years.J Clin Oncol 1992;10:1547–52.

44. Fisher B, Dignam J, DeCillis A, et al. The worth ofchemotherapy and tamoxifen (TAM) over TAMalone in node negative patients with estrogen-receptor positive invasive breast cancer: firstresults from NSABP B-20 [abstract]. Proc AnnMeet Am Soc Clin Oncol 1997;16:1a.

45. Albain K, Green S, Osborne CK, et al. Tamoxifen(T) versus cyclophosphamide, adriamycin and 5-FU plus either concurrent or sequential T in post-menopausal receptor (+), node (+) breast cancer:a Southwest Oncology Group phase III inter-group (SWOG-8814, INT-0100) [abstract]. ProcAnn Meet Am Soc Clin Oncol 1997;16:128a.

46. International Breast Cancer Study Group. Effec-tiveness of adjuvant chemotherapy in combina-tion with tamoxifen for node-positive post-menopausal breast cancer patients. J Clin Oncol1997;15:1385–94.

47. Pritchard KI, Paterson AH, Fine S, et al. Random-ized trial of cyclophosphamide, methotrexateand fluorouracil chemotherapy added to tamox-ifen as adjuvant therapy in postmenopausalwomen with node-positive estrogen and/orprogesterone receptor-positive breast cancer: areport of the National Cancer Institute ofCanada Clinical Trials Group. J Clin Oncol1997;15:2302–11.

48. Bonadonna G, Valagussa P. Primary chemother-apy in operable breast cancer. Semin Oncol1996;23:464–74.


50. Fisher B, Brown A, Mamounas E, et al. Effect ofpreoperative chemotherapy on local-regionaldisease in women with operable breast cancer:findings from NSABP B-18. J Clin Oncol1997;15:2483–93.

51. Willsher PC, Robertson JFR, Chan SY, et al.Locally advanced breast cancer: early results ofa randomized trial of multimodal therapy ver-sus initial hormone therapy. Eur J Cancer1997;33:45–9.

52. Wolmark N, Dignam J, Fisher B. The addition oftamoxifen to lumpectomy and radiotherapy inthe treatment of ductal carcinoma in situ(DCIS): preliminary results of NSABP protocolB-24. Breast Cancer Res Treat 1998;50:227.

53. Paik S, Bryant J, Park C, et al. ErbB-2 andresponse to doxorubicin in patients with axil-

lary lymph node-positive, hormone receptor-negative breast cancer. J Natl Cancer Inst 1998;90:1361–70.

54. Thor AD, Berry DA, Budman DR, et al. ErbB-2,p53, and efficacy of adjuvant therapy in lymphnode-positive breast cancer. J Natl CancerInstitute 1998;90:1346–60.

55. Borg A, Caldetorp B, Ferno M, et al. ErbB-2amplification is associated with tamoxifenresistance in steroid-receptor positive breastcancer. Cancer Lett 1994;81:137–44.

56. Carlomagno C, Perrone F, Gallo C, et al. C-erbB-2 overexpression decreases the benefit of adju-vant tamoxifen in early-stage breast cancerwithout axillary node metastases. J Clin Oncol1996;14:2702–8.

57. Soubeyran I, Quenel N, Coindre JM, et al. PS2 pro-tein: a marker improving prediction of responseto neoadjuvant tamoxifen in post-menopausalbreast cancer. Br J Cancer 1996;74:1120–5.

58. Berns EM, Foekens JA, van Staveren IL, et al.Oncogene amplication and prognosis in breastcancer: relationship with systemic treatment.Gene 1995;159:11–8.

59. Gelber RD, Cole BF, Goldhirsch A, et al. Adjuvantchemotherapy plus tamoxifen compared withtamoxifen alone for postmenopausal breastcancer: meta-analysis of quality-adjusted sur-vival. Lancet 1996;347:1066–71.

60. Bines J, Oleske DM, Cobleigh MA. Ovarian func-tion in premenopausal women treated withadjuvant chemotherapy for breast cancer. JClin Oncol 1996;14:1718–29.

61. Von Hoff D, Layard MW, Basa P, et al. Risk fac-tors for doxorubicin-induced congestive heartfailure. Ann Int Med 1979;91:710–7.

62. Tallman MS, Gray R, Bennett JM, et al. Leuke-mogenic potential of adjuvant chemotherapyfor early-stage breast cancer: the ECOG expe-rience. J Clin Oncol 1995;13:1557–63.

63. Valagussa P, Tancini G, Bonadonna G. Secondmalignancies after CMF for resectable breastcancer. J Clin Oncol 1987;5:1138–42.

64. Diamandidou E, Buzdar AU, Smith TL, et al.Treatment-related leukemia in breast cancerpatients treated with fluorouracil-doxorubicin-cyclophosphamide combination adjuvantchemotherapy: the University of Texas M.D.Anderson Cancer Center experience. J ClinOncol 1996;14:2722–30.

65. DeCillis A, Anderson A, Bryant J, et al. Acutemyloid leukemia and myelodysplastic syn-drome on NSABP B-25: an update. Proc AmSoc Clin Oncol 1997;16:130a.


66. Gorin MB, Day R, Costantino JP, et al. Long-termtamoxifen citrate use and potential ocular toxi-city. Am J Ophthalmol 1998;125:493–501.

67. Jordan VC, Assikis VJ. Endometrial carcinomaand tamoxifen: clearing up a controversy. ClinCancer Res 1995;1:467–72.

68. Rutqvist LE, Johansson H, Signomklao T, et al.Adjuvant tamoxifen therapy for early stagebreast cancer and second primary malignan-cies. J Natl Cancer Inst 1995;87:645–51.

69. Fisher B, Costantino JP, Redmond CK, et al.Endometrial cancer in tamoxifen-treated breastcancer patients: findings from the NationalSurgical Adjuvant Breast and Bowel Project(NSABP) B-14. J Natl Cancer Inst 1994;86:527–37.

70. Magriples U, Naftolin F, Schwartz PE, CarcangiuML. High-grade endometrial carcinoma intamoxifen-treated breast cancer patients. J ClinOncol 1993;11:485–90.

71. Fisher B, Costantino JP, Wickerham DL, et al.Tamoxifen for the prevention of breast cancer:report of the NSABP P-1 study. J Natl CancerInst 1998;90:1371–88.

72. Love RR, Mazess RB, Barden HS, et al. Effects oftamoxifen on bone mineral density in post-menopausal women with breast cancer. N EnglJ Med 1992;326:852–6.

73. Kristen B, Ejlertsen B, Dalgaard P, et al. Tamox-ifen and bone metabolism in postmenopausal

low-risk breast cancer patients: a randomizedstudy. J Clin Oncol 1994;12:992–7.

74. Powles TJ, Hickish T, Kanis JA, et al. Effect oftamoxifen on bone mineral density measuredby dual-energy x-ray absorptiometry in healthypremenopausal and postmenopausal women. JClin Oncol 1996;14:78–84.

75. Love RR, Wiebe DA, Feyzi JM, et al. Effects oftamoxifen on cardiovascular risk factors inpostmenopausal women after 5 years of treat-ment. J Natl Cancer Inst 1994;86:1534–9.

76. McDonald CC, Alexander FE, Whyte BW, et al.Cardiac and vascular morbidity in womenreceiving adjuvant tamoxifen for breast cancerin a randomized trial. The Scottish Cancer Tri-als Breast Group. Br Med J 1995;311:977–80.

77. Costantino JP, Kuller LH, Ives DG, et al. Coro-nary heart disease mortality and adjuvanttamoxifen therapy. J Natl Cancer Inst 1997;89:776–82.

78. Rutqvist LE, Mattsson A. Cardiac and throm-boembolic morbidity among post menopausalwomen with early-stage breast cancer in a ran-domized trial of adjuvant tamoxifen. TheStockholm Breast Cancer Study Group. J NatlCancer Inst 1993;85:1398–406

79. Recht A, Come SE, Henderson IC, et al. Thesequencing of chemotherapy and radiation ther-apy after conservative surgery for early-stagebreast cancer. N Engl J Med 1996;334:1356–61.

218 BREAST CANCER

219

The role of radiation therapy in the treatment ofbreast cancer has been affected by shiftingtrends in treatment. Historically, mastectomywas the surgical treatment of choice for allstages of disease. In later times, radiation ther-apy was found to significantly reduce the riskof local recurrence. Pioneers in radiotherapysuch as Gilbert Fletcher published importantdose response data for the control of regionaladenopathy.1 As the treatment of breast cancerhas changed, the role of radiotherapy hasevolved for all stages of disease.

Breast conservation therapy for ductal carci-noma in situ (intraductal) and early-stage inva-sive cancer has been shown to provide equiva-lent survival to mastectomy in properly selectedpatients. The importance of cosmesis in breastconserving treatment can not be underesti-mated. Issues of self-esteem and sensuality addcomplexity to the treatment decision. Breastconservation therapy (BCT) is clearly estab-lished as a standard treatment for early-stagebreast cancer. In fact, a consensus panel of theNational Institutes of Health determined BCTto be the preferred treatment for patients eligi-ble for the procedure.2

Prognostic factors have been evaluated todetermine those patients at high risk for recur-rence after BCT. There are many new patho-logic and cytologic markers that may be used topredict local or distant recurrence after stan-

dard treatment. On the other hand, there may besubsets in the population of breast cancerpatients whose risk of recurrence is low enoughafter local excision alone to preclude radiother-apy after lumpectomy. Future investigations inBCT will be needed to individualize treatmentbased on these evolving parameters.

Routine postmastectomy radiotherapy hasfallen out of favor in recent years, despite provenefficacy in reducing local recurrence. Althoughlocal recurrences were reduced, positive impacton overall survival was difficult to document.Thus, the potential for distant metastatic diseaseis thought to be the single most important prog-nostic factor for survival. Chemotherapy hasemerged as an effective treatment for metastaticand potentially metastatic disease. Adriamycin-based chemotherapy is probably the most effec-tive regimen, despite well-documented cardiacmorbidity associated with adriamycin. Cardiacmorbidity due to the late effects of radiationbegan to be appreciated at the same time theseeffective chemotherapy regimens were devel-oped.3 Additionally, systemic and local toxicityprohibits concomitant delivery of radiotherapyand systemic therapy for breast cancer. Radio-therapy following mastectomy has thereforebeen infrequently used.

Recent data have emerged showing a sur-vival benefit when radiotherapy is added tomastectomy and chemotherapy as treatment for

13

Breast Cancer and Radiation TherapyMICHAEL A. L ACOMBE, MDWILLIAM D. BLOOMER, MD

jimingshuo

高亮

selected patients. Older studies of radiotherapyhave been criticized for poor technique as wellas employing low-energy equipment.4 How-ever, historic studies of patients treated withmodern equipment showed a survival advan-tage. Thus, it is important to define thosepatients with locally advanced breast cancerwhose risk of local recurrence is significantenough to warrant radiotherapy.

Confounding any evaluation of the treat-ment results of breast cancer is the long naturalhistory of the disease, the interval evolution ofsurgical, radiotherapeutic, and systemic treat-ment, and the long latency of late-radiationsequelae developing. The long natural historyof breast cancer requires years of follow-up todissect results of comparative trials. The devel-opment of more effective therapies obscuresresults as local and distant recurrences may notdevelop for many years. During these long fol-low-up periods, the standards of treatmentchange. Current studies in breast cancer ther-apy may be virtually obsolete before the datamatures. Trials have been criticized for usingchemotherapy regimens that are no longerstandard.5 The toxicity of radiotherapy is mostoften defined by late effects on normal tissue.These late effects may take years to developand are related to technique. For example, therisk of second malignancy may not be evalu-able for decades.6

The epidemiology of breast cancer is chang-ing. There was a large increase in incidenceseen during the 1980s, and a constant increaseof about 1 percent per year has been seen untilrecently.7 This is undoubtedly because of theacceptance of screening mammography. In1979, 22 percent of women over the age of 40years had had at least one mammographicstudy; presently, 74 percent of women over theage of forty have undergone at least one study.The incidence of ductal carcinoma in situ(intraductal) breast cancer increased from 7.4percent in 1985 to 14.3 percent in 1995,8 aresult attributable almost entirely to bothimprovements in mammographic techniques

and increased screening. More patients are pre-senting with stage I disease, fewer with stage IIdisease, with the number presenting with moreadvanced disease remaining the same. Studiescomparing screened populations with unscreen-ed patients have demonstrated a reduction inoverall breast cancer mortality in the screenedpatients.9,10 The logical conclusion is that theacceptance of a screening modality leads todetection of tumors at an earlier stage as well asinitially increasing the incidence of the disease.Earlier staged tumors are more curable, andtheir detection decreases mortality related tothe disease. The incidence of the disease mayeventually decline as a generation of patientscompletes screening.

Acceptance of the equivalency of BCT tomastectomy is not widespread. Although it isestimated that 70 to 80 percent of patientsshould be eligible for BCT, the actual numberof patients undergoing BCT is as low as 10 to15 percent.11,12 There is a trend away from thishowever. One study showed the number ofpatients treated with mastectomy for stage Ibreast cancer to have decreased from 56 to 43percent from 1985 to 1995.8 The most commonmedical contraindication to BCT is multicen-tric disease involving more than one quadrantof the breast.13 Intraductal cancer presents withmulticentricity in a higher percentage ofpatients than does invasive cancer. On the otherhand, 20 percent of patients eligible for BCTchoose mastectomy.13

How does this apply to the radiotherapeutictreatment of breast cancer? As more early-stage and noninvasive cancers are detected,more patients will be eligible for breast con-servation therapy. There appears to be growingacceptance of BCT as the standard treatment ineligible patients. Optimizing individual treat-ment based on predictive indicators for localrecurrence may lead to subsets of patients whodo not require radiotherapy after excision.Likewise, those at highest risk for local recur-rence may benefit from additional or alterna-tive treatment.

220 BREAST CANCER

EARLY-STAGE BREAST CANCER

Mastectomy and Breast ConservationTherapy: Equivalent Survival

Several randomized clinical trials have con-firmed equivalent survival data when lumpec-tomy and breast radiotherapy are compared tomodified radical mastectomy for early-stageinvasive breast cancer. Local recurrence ratesafter conservative surgery and radiotherapyrange from 3 to 19 percent (Table 13–1).14–20 Itshould be noted that the study from the NCIhad the highest rate of local recurrence but didnot require negative pathologic resection mar-gins, the significance of which is discussedbelow. This same study was the only trial toshow a significant advantage to mastectomywith respect to local recurrence. Despite abroad range of entry criteria (acceptable tumorsize ranged from 2 to 5 cm), the data areremarkably consistent. The most importantfinding was the equivalent overall survivalbetween treatment arms in all studies. Localrecurrence in the mastectomy-treated patientsranged from 4 to 14 percent. Equivalent sur-vival between extensive surgery and limitedsurgery plus radiotherapy was confirmed in ameta-analysis of randomized trials in breastcancer. Also, a threefold reduction in localrecurrence was demonstrated with the additionof radiotherapy to local excision.21

There are large retrospective reviews corre-lating these results.22 These studies have alsobeen useful in identifying prognostic features

from patients with breast cancer treated withconservative surgery and radiotherapy. It isvery safe to conclude that the data supports theuse of breast conservation therapy for the treat-ment of early-stage invasive breast cancer.

Risk Factors for Local Recurrence

Among patients treated with breast conserva-tion therapy, several risk factors have beenidentified that predict for local recurrence.These have been classified as patient-related,tumor-related, and treatment-related factors. Ifa risk factor was associated with a prohibitivelyhigh recurrence rate, this would be a con-traindication to BCT only if the same factorwas not a risk factor for recurrence after mas-tectomy. For example, several series haveshown age of diagnosis to be a predictive indi-cator for local recurrence after BCT, withyounger patients being at higher risk.23–26 How-ever, one cannot conclude that young age is acontraindication for BCT, because other studieshave shown the same population of youngpatients to be at higher risk for local recurrenceafter mastectomy.27,28

PATHOLOGIC VARIABLES

Table 13–2 lists several factors that are or at onetime were thought to be factors predictive forlocal recurrence. Multicentricity in more thanone quadrant of the breast is a contraindicationto BCT. Several authors have concluded that therisk for local failure is significantly higher in

Breast Cancer and Radiation Therapy 221

Table 13–1. BREAST CONSERVATION THERAPY COMPARED WITH MASTECTOMY: RANDOMIZED TRIALS

Follow-upLocal Recurrence (%) Survival (%)

Trial (years) Mastectomy BCT p Value Mastectomy BCT p Value

NSABP B-0614 12 8 10 ns 59 63 nsNCI15,16 10 6 19 .01 75 77 nsMilan17 18 4 7 ns 65 65 nsEORTC18 14 14 17 ns 61 54 nsDanish19 — 4 3 ns 82 79 nsIGR20 15 14 9 ns 65 73 ns

NS = not significant.

patients with documented disease in more thanone quadrant of the breast.29–31 This likely indi-cates the presence of residual tumor burdenthroughout the breast after local excision thatcannot be controlled with radiotherapy doses of45 to 50 Gy. Since mastectomy for multicentricdisease is not associated with an increased riskof local recurrence,32 it is the standard of carefor patients with defined multicentricity.

Historically, it was thought that patientswith an extensive intraductal component (EIC)associated with their invasive tumor were atsignificant risk for recurrence.33,34 This condi-tion has been defined as intraductal carcinomaoccupying ³ 25 percent of the area encom-passed by invasive tumor as well as the sur-rounding stroma of the resected specimen.However, when this risk factor was re-evaluatedin association with negative surgical resectionmargins, it was found that if negative marginsfor both invasive and noninvasive componentscould be obtained, no increase in local recur-rence was observed.35,36

The presence of lobular carcinoma in situ(LCIS) within the tumor specimen likewise hasno impact on the probability of local recur-rence.37 However, these patients may be athigher risk for development of a second subse-quent cancer in the contralateral breast. Infil-trating lobular histology behaves differentlythan infiltrating ductal histology. Although thenatural history of lobular carcinoma may differ

(larger size at presentation, higher hormonalreceptor positivity, decreased axillary involve-ment), there is no difference in local control inpatients with infiltrating lobular histologytreated with local excision and radiotherapy.38

Patients with centrally located tumors wereinitially thought to be unsuitable for BCT due tocosmetic concerns about tumors near the nipple-areolar complex.30 There was also a concern thatinvolvement of major breast ducts might be asso-ciated with more diffuse disease. However, recentanalysis of tumors within 2 cm of the nipple-areolar complex show no increased risk of localfailure when treated with breast conservationtherapy, provided negative resection margins areachieved.39 Good to excellent cosmesis can beachieved in this scenerio.40,41

Patients with Paget’s disease of the nipplewere previously considered poor candidates forBCT because of its significant association withseparate palpable infiltrating malignancies. Forpatients with Paget’s disease and no palpablemalignancy, the concern was the possibility ofmicroscopic invasive disease. This concern isprobably less valid today with newer mammo-graphic technologies. Moreover, it has beenshown that if patients presenting with Paget’s dis-ease of the breast without an associated palpablemass are treated with local excision and radio-therapy, there is no increased risk of recurrence.42

Tumor size and histologic grade have prog-nostic significance for survival, risk of axillaryinvolvement, and the development of distantmetastatic disease. However, local control maynot be impacted by either the size or histologicgrade of the primary lesion. Most retrospectivestudies find no significant difference in localcontrol for T1 or T2 tumors. Recent studies failto correlate higher grade tumors with increasedlocal recurrence.43

PATIENT VARIABLES

Significant family history of breast cancer is nota contraindication for BCT. There is no higherrisk of local failure in patients with a history of

222 BREAST CANCER

Table 13–2. RISK FACTORS FOR LOCAL RECURRENCEAFTER BREAST CONSERVATION THERAPY

Current Risk Factor Literature Support

Young age at diagnosis23–27 YMulticentricity29–32 YExtensive intraductal component35,36 NLCIS found in specimen37 NInfiltrating lobular histology38 NSubareolar location39–41 N Paget’s disease42 NPathologic grade43 NFamily history44,45 NPositive surgical margins47,51 Y

LCIS = lobular carcinoma in situ.

breast cancer in a first-degree relative. In fact,several studies have shown that patients withbreast cancer who have a first-degree relativewith the disease may have an increased overallsurvival.44,45 This observation may correlatewith evolving data showing a good prognosiswith BRCA1 genetic overexpression.46

Young age at diagnosis is associated with anincreased local recurrence rate. The studies dif-fer in the age at which risk of recurrence is sig-nificantly higher, but all conclude young age tobe a risk factor. Because these same patients areat higher risk for local failure after mastectomy,young age is not a contraindication to BCT.24–27

In addition, there is no impact on overall sur-vival for younger patients treated with breastpreservation therapy.

THERAPEUTIC VARIABLES

Surgical Factors

The principal surgical factor correlated withincreased local–regional recurrence after BCTis an inability to resect the lesion with clearpathologic margins. Definition of a “clear mar-gin” varies. While the absence of tumor at theinked margin of the resected specimen isdefined by some authors as a negative margin,others have differentiated between negative andclose margins by using 1 to 3 mm to define anegative margin. These distinctions notwith-standing, there appears to be a consistently sig-nificant increase in local recurrence with dif-fusely positive margins compared to “negative”margins. The recurrence rate for diffusely posi-tive margins approaches 30 percent, comparedto 2 to 10 percent for negative margins. Thus,the presence of diffusely positive pathologiclumpectomy margins is an indication for re-excision. If re-excision is not possible, mastec-tomy should be performed. Surgical marginsclassified as negative, close, focally positive, ormore than focally positive correlate with a 2, 3,9, and 28 percent risk of local recurrence,respectively (Table 13–3).47–49

Focally positive and close margins appear tobe associated with a risk of local recurrencethat is intermediate between diffusely positiveand negative margins. In some studies, thisintermediate risk of recurrence is statisticallysignificant, implying re-excision should be per-formed.50 However, other studies do not showsignificance related to close or focally positivemargins.51 The presence of two or more positivemargins has been shown to be predictive ofhigher local failure compared to one.52 Of par-ticular note, proper evaluation of resection mar-gins is essential, as one-third of patients withfocally positive shaved margins are negativewhen inked margins are properly evaluated.

The amount of tumor at the margin of a sur-gical specimen correlates with the risk of dif-fuse involvement in the breast.53 There is astrong correlation between the extent ofinvolvement of surgical margins and residualtumor burden in the subsequently re-excisedspecimen. The risk of residual disease is a con-tinuum based upon distance from the tumor,with margin distance determined within thecontext of cosmetic outcome.54 Standard dosesof radiotherapy are less likely to control abreast with a significant microscopic tumorburden. Although boost doses of radiotherapyto the surgical bed are believed to increase localcontrol, definitive evidence is meager. Onestudy looked at the effect on local control withboth re-excision and boost doses of radiother-apy.55 There was a statistically significantincrease in local control with re-excision forclose, indeterminate, or positive margins com-pared to no re-excision. There was no differ-ence in local control for these same patientswhen evaluated by total dose delivered to thetumor site, indicating that re-excision for ques-tionable margin status was more important thanboost irradiation. This supports the notion thatthe best radiation in the world cannot compen-sate for inadequate surgery. The presence of afocally positive margin appears to be signifi-cant and should also be re-excised. Althoughthe significance of close resection margin prob-


ably depends more on the underlying defini-tions, re-excision does not appear necessary.Ideally, negative margins should be obtained tooptimize local control.

The anticipated cosmetic outcome afterlumpectomy is critical to the surgical selectionof patients for BCT. The pressure of a largetumor-to-breast ratio is likely to lead to a lesssatisfactory cosmetic outcome from the pointof view of both patient and physician. Excisionof a major proportion of breast tissue to obtainnegative margins (usually greater than a quad-rant) is a relative contraindication to BCT. Cos-metic evaluation of the breast is very subjec-tive, with many layers of complexity in terms ofself-esteem and sensuality. While studies com-paring lumpectomy to quadrantectomy haveshown better local control with quadrantectomyfollowed by radiotherapy, in general lumpec-tomy plus radiotherapy provides equivalentlocal control, with better cosmesis from thepoint of view of both patient and physician.56,57

Radiation Therapy Factors

Certain collagen vascular diseases, such asscleroderma and systemic lupus erythematosus,have been associated with an increase in acuteskin and subcutaneous toxicity to standarddoses of radiotherapy and may be related toinadequate repair of sublethal radiation injury.However, a retrospective review of patientswith collagen vascular disease showed that only

patients with scleroderma exhibited prohibitivetoxicity. Patients with other collagen vasculardiseases did not experience prohibitive toxicityand could be considered candidates for BCT.58

The potentially deleterious effects of radio-therapy on the developing fetus prohibit patientsin the first two trimesters of pregnancy frombeing candidates for BCT. Estimated cumula-tive fetal doses of 3 to 4 Gy are delivered withtangential radiotherapy.59 Because there is noknown threshold dose for mutagenesis, radio-therapy is an absolute contraindication duringthe first two trimesters. Long-term effects ofradiation during the third trimester are unclear.

Boost to Lumpectomy Site

Two questions exist regarding the use of irradi-ation to the lumpectomy site in conjunctionwith whole breast irradiation. First, is a boostnecessary? Most centers treat the whole breastwith fraction sizes of 180 cGy to a total dose of45 to 50 Gy followed by a boost to the tumorsite of an additional 10 to 15 Gy. Protocolsestablished by the National Surgical AdjuvantBreast Project (NSABP) routinely treat at200 cGy per day to a total dose of 50 Gy withno boost to the tumor site. The principal ratio-nale for boost irradiation is that it can be deliv-ered safely without major cosmetic detriment.In a large retrospective study, 17 percent ofpatients who did not receive a boost showedlocal failure, compared to 11 percent for those

224 BREAST CANCER

Table 13–3. MARGINS OF EXCISION CORRELATED TO RATE OF LOCAL RECURRENCE

Margin Status (%)

Study Negative Positive Close Indeterminate Focally More thaninvolved Focally involved

Anscher et al47 10 2 10Gage et al51 2 16 3 9 28Heimann et al49 2 11Ryoo* et al48 6 13 8Smitt et al63 2 18†

Solin* et al62 10 8 14 13Spivack et al50 3.7 18

*Boost radiation delivered based on margin status.†Combined data on close or positive margins.

who did.60 On the other hand, another studyrandomizing patients to a boost after wholebreast radiation versus no boost foundincreased local control in boosted patients.61

However, patients treated with a boost had aworse cosmetic outcome, with a statisticallysignificant increase in telangiectasia formation.Unfortunately, the dose per fraction in the studywas 250 cGy for both the whole breast andboost portions of treatment—this represents ahigher fractional dose than is routinely used inthe United States.

The second question is whether an increasedradiation boost can compensate for close orpositive resection margins. The data on thisshow mixed results. Retrospective data areoften confounded by the routine use of anincreased boost dose to patients with close mar-gins. On the one hand, there are studies show-ing that the increasing dose used to boost closemargins increased local control.62 On the otherhand, studies demonstrate that patients withpositive margins have higher local recurrencerates with or without a boost.63 The issue maybe resolved by a current EORTC study ran-domizing patients with inadequate surgicalmargins to boost doses of 10 and 25 Gy.

Systemic Therapy Factors

Two conclusions can be gleaned from data on theaddition of chemotherapy to BCT. Chemotherapyadded to lumpectomy and radiotherapy inpatients at high risk for development of metasta-tic disease reduces the risk of ipsilateral breastrecurrence to as low as 2.6 percent.64 On theother hand, chemotherapy cannot substitute forradiotherapy in BCT. Local recurrence in patientstreated with local excision and chemotherapy hasbeen shown to be significantly higher than thatfor standard lumpectomy and radiotherapy.65

Tamoxifen and Local Control

Tamoxifen benefits estrogen-receptor positivepatients both in terms of overall survival and

local control. Three randomized trials evaluat-ing the addition of tamoxifen to BCT haveshown local control and event-free survival tobe significantly improved in tamoxifen-treatedpatients. Tamoxifen reduces ipsilateral and con-tralateral recurrences.66–69 There is currently arandomized trial is underway to evaluate theomission of breast radiation in elderly womenwith estrogen-receptor positive tumors treatedwith excision and tamoxifen.

Cytologic Factors and Local Recurrence

New cytologic and genetic factors are beingidentified and associated with breast cancerprognosis. The majority of these studies to dateare small retrospective series and have notinfluenced the choice of therapy. Case controlstudies looking at the overexpression ofinsulin-like growth factor-I receptor (IGF-IR)and HER-2/neu indicate that overexpressionmay predict for an increased risk of local recur-rence.70,71 In these studies, the tumor tissue ofall patients who had experienced local recur-rence at a single institution were evaluated foroverexpression of these newly described mark-ers. Case controls were drawn from thosepatients treated who did not experience localrecurrence. Overexpression of IGF-IR andHER-2/neu was found significantly more fre-quently in patients who experienced recurrence.Although HER-2/neu overexpression negativelyimpacted disease-free survival in patientstreated with tamoxifen and radiotherapy,72 thiseffect was not seen in patients treated withchemotherapy.

The impact on local control, however, isunclear. Several studies show that patients withthe germ-line mutations BRCA1 or BRCA2 mayexperience statistically significant improvementin survival.73 Overexpression of p53 has beenassociated with poor response to tamoxifen, butno data exist on its role with radiotherapy orlocal control.74 The “tumor suppressor” genep53 is thought to confer radioresistance by a lossof apoptosis in response to radiation. The possi-


bility of radioresistance has been raised in a pre-liminary study of patients treated with BCT ormastectomy and postoperative radiotherapy.75

Finally, the presence of angiogenesis as mea-sured by microvessel count has been shown to beprognostically significant for survival. Thosepatients with node-negative cancers and a lowmean vessel count (MVC) had excellent overallsurvival, but the impact on local recurrence wasnot indicated.76 However, laboratory studieshave shown a synergistic effect between angio-genesis inhibitors and radiotherapy.77

Lumpectomy Alone

The NSABP B-06 trial determined that the riskof local recurrence after local excision alonewas as high as 35 percent, compared to 10 per-cent when combined with radiotherapy. Analternative view of this data is that 65 percentof patients will not experience local recurrenceafter lumpectomy alone. Therefore, the major-ity of patients treated with radiotherapy wouldnot have local recurrence in the absence of radi-ation. Attempts have been made to identify sub-sets of patients with early stage breast cancerwho do not require radiotherapy. Six studieslisted in Table 13–4 prospectively randomizedpatients to excision versus excision plus radio-therapy.78–81 Although all studies have shown asignificant increase in local recurrence in thepatients who did not receive radiation, noneshowed a statistically significant survivaladvantage with radiotherapy.

The stress of tumor recurrence on thepatient as well as the low morbidity of tangen-tial breast radiation must be considered whenevaluating these results. Although no survivaladvantage was documented, local recurrencemay be predictive of subsequent increased mor-tality. Also, a significant number of patientswho experience recurrence after local excisionalone choose mastectomy at the time of recur-rence, regardless of survival data.

Within these studies, subset analyses wereperformed to identify patients with favorable

prognostic features, since these patients maybenefit least from the addition of radiother-apy.82,83 The two favorable prognostic factorsidentified were advancing age of the patientand small tumor size. In patients with T1tumors and advancing ages, the local recur-rence risk ranged from 4 to 16 percent. Of note,the Uppsala study compared patients > 50 yearsold with T1 tumors to all others and found arecurrence rate of “only” 15.9 percent.84 How-ever, such a local recurrence is probably unac-ceptable since the addition of radiation likelyreduces the risk of recurrence to < 5 percent. Ofnote, some studies were unable to identify sub-sets of patients who did not benefit from radio-therapy.85 As indicated above, there is currentlya prospective trial for patients > 70 years oldwith T1 tumors evaluating the omission of radi-ation in patients receiving tamoxifen, based onthe subset analysis of the prior studies andknowledge of the impact of tamoxifen on localcontrol. Outside of a trial setting, the omissionof radiation therapy from breast conservingtherapy is not presently indicated. Interestingly,practice patterns in the treatment of patients> 65 years old have been shown to differ fromyounger patients, with this population morelikely to be treated with local excision and noradiotherapy.86,87 The ongoing study may justifythese practice patterns.


The increasing acceptance of screening mam-mography has led to a dramatic increase in theincidence of ductal carcinoma in situ (DCIS),or intraductal carcinoma. Historically, patho-logic evaluation of intraductal carcinomashowed focality in the majority of cases. How-ever, Holland showed intraductal foci of intra-ductal disease remote from the primary site in40 percent of patients.88 Thus, the rationale forradiotherapy to the remaining breast followinglocal excision appears appropriate for themajority of patients with intraductal cancer.There was a subset of patients on NSABP B-06

226 BREAST CANCER

identified as having in situ histology. The localrecurrence rate in patients undergoing excisionwithout radiotherapy was 43 percent, versus 7percent in patients receiving radiation.89,90

The largest prospective study of DCIS is theNSABP randomized study B-17 comparinglocal excision alone with local excision plusradiotherapy.91,92 There was a significantdecrease in local recurrence in the radiationarm but no difference in overall survival. Theaddition of radiation reduced the noninvasivecancer recurrence from 10.4 to 7.5 percent.Invasive cancer recurrences were reduced from10.5 to 2.9 percent. Importantly, of those withrecurrence in the local excision-only arm, halfthe recurrences were invasive. Other retrospec-tive series confirm these findings.93–95 Thus,the standard of care for patients with DCIS whoare eligible for breast sparing treatment is localexcision followed by radiotherapy.

Several centers have attempted to identifysubgroups of patients with an extremely lowrisk for recurrence after local excision. Tumorgrade, particularly the presence of comedocar-cinoma, is a significant predictor of local recur-rence. The Van Nuys Prognostic Index (VNPI)has been proposed, using 155 patients treatedwith local excision alone.96,97 There is a scoregiven to each of three factors: tumor grade, sizeof disease, and extent of negative surgical mar-gins. The result of a low VNPI appears to bepredictive for decreased local recurrence.98

There is presently a single-arm nonrandomizedstudy underway evaluating low-risk patients forlocal recurrence after local excision alone. As

an entry criteria, a minimum of 3 mm negativesurgical margins are required.

The impact of local recurrence for DCISdoes not appear to have an impact on sur-vival.99,100 However, the potentially devastatingemotional trauma the patient experiences withrecurrence should be balanced against the verylow risk of radiation complications. Decreasedmortality from breast cancer is likely the resultof earlier detection of earlier-stage disease, par-ticularly DCIS. The reported 75 percent reduc-tion in subsequent development of invasivebreast cancer for patients with DCIS treatedwith postexcisional radiotherapy supports thisconclusion.

RADIOTHERAPY FOR LOCALLYADVANCED BREAST CANCER

Early studies of local-regional radiotherapy fol-lowing mastectomy for locally advanced dis-ease showed reduced recurrences in the axilla,supraclavicular fossa, and chest wall. However,these patients experienced an increase in non-breast cancer-related mortality that negated thesurvival advantage.101,3

Cardiac toxicity due to chest wall radiother-apy is purported to be the causative factor in theincrease in nonbreast cancer-related mortality.Concerns about cardiac morbidity from radio-therapy in the face of potentially cardiotoxicadriamycin-based chemotherapy led to decreaseduse of postmastectomy radiotherapy. Technicalfactors in historically quoted studies mayexplain the high incidence of radiation cardiac


Table 13–4. COMPARISON OF LUMPECTOMY VERSUS LUMPECTOMY PLUS RADIOTHERAPY: RANDOMIZED TRIALS

Follow-upLocal Recurrence (%) Survival (%)

Trial (years) No XRT XRT p Value No XRT XRT p Value

NSABP B-0614 12 35 10 + 58 62 nsScottish trial69 5 28 6 + 85 88 nsOntario78 8 35 11 + 90 91 nsMilan79,80 5 18 2 + 92 92 nsUppsala81 5 18 2 + 90 91 ns

XRT = external beam radiation therapy.

toxicity. The principle factor is the use of an enface internal mammary and supraclavicularportal, often referred to as a “hockey stick”port. Other considerations are related todosimetry, such as the use of orthovoltageenergy radiotherapy. The energy delivered withorthovoltage radiation is lower in energy, lesspenetrating, therefore leading to greater inho-mogeneity of dose. Comparative dosimetricanalysis has shown the dose to the heart to besignificantly higher with orthovoltage thanwith modern megavoltage radiotherapy.102,103

When patients treated with orthovoltageradiotherapy were excluded from analysis, a sur-vival benefit of approximately 10 percent wasseen in the patients receiving chest wall radio-therapy.104 Recently, two prospective random-ized studies reported results showing survivaladvantage from the addition of local–regionalradiotherapy after mastectomy and chemother-apy for node-positive premenopausal breast can-cer patients. Both studies show an improvementin survival of 8 to 10 percent.105,106 Both studieshave shown a benefit to radiating all patients,regardless of the number of lymph nodesinvolved. Although patients with one to threenodes positive and four or more nodes positiveboth showed benefit from radiation in subgroupanalysis, routine regional lymph node radiother-apy in all patients with positive nodes is not gen-erally accepted practice (Table 13–5).

There has been criticism of the studies forinadequate evaluation of the axilla. Many patientshad small numbers of axillary lymph nodesexcised. The patient with only one positive nodeexcised may have had other involved nodes in theundissected axilla. Thus, a proportion of thepatients thought to have one to three nodes posi-tive may have actually had four or more involvednodes. Prior studies have shown this populationof patients to be at lower risk for local–regionalrecurrence. There was also criticism that thelocal recurrence rates in these studies were inor-dinately high compared to prior studies.107

Prior retrospective studies have evaluatedthe role of radiotherapy following mastectomy

for locally advanced breast cancer. Studies havelooked at recurrence patterns in patients treatedwith mastectomy alone.108 Results have shownthat patients with four or more nodes positive,large primary tumors > 5 cm, pectoral fascia, orskin involvement are at greatest risk of localrecurrence after mastectomy, as high as 25 to30 percent. Chest wall radiotherapy in thesepatients reduces local recurrence to < 10 per-cent. Recurrence patterns show the chest walland supraclavicular fossa to be the most com-mon sites of recurrence, with axillary recur-rence seen less frequently. Patients with tumors< 2 cm and negative axillary lymph nodes havelocal recurrence risk of < 10 percent and wouldnot appear to benefit from radiotherapy.109

Patients with tumors between two and five cen-timeters or one to three involved axillary lymphnodes are at intermediate risk, the role of radio-therapy is controversial in these patientsbecause of modest potential benefit.110

Extracapsular extension (ECE) of tumor inaxillary lymph nodes is a potential risk factorfor regional recurrence. It is likely to be associ-ated with other poor prognostic indicators, suchas multiple positive axillary nodes. Whileshown to be significant on univariate analysis,

228 BREAST CANCER

Table 13–5. POSTOPERATIVE CHEST WALL RADIOTHERAPY FOLLOWING MASTECTOMY

Treatment indicated108–110

Tumor greater than 5 cm orPositive surgical margins orFour or more positive axillary lymph nodes

Treatment to be considered112

Premenopausal patients with one to three positive nodesInclusion of internal mammary nodes controversial

Treatment not indicatedAll patients with

Tumor less than five centimeters andNegative margins andNo positive lymph nodes

Postmenopausal patients withOne to three positive lymph nodes

Treatment not indicated based on recent literature113

Patients with extracapsular extension of lymph node involvement

Treatment guidelines differMale patients with breast cancer

ECE was not found to predict for regional fail-ure on multivariate analysis.111,112

The recently published trials showing a sur-vival benefit to radiotherapy after mastectomyincluded the internal mammary nodes. The argu-ment for treatment is that persistent microscopicinvolvement after chemotherapy is still poten-tially curable with regional radiotherapy. Theoret-ically, untreated microscopically involved nodesmay lead to development of distant disease. Cur-rently, there is an EORTC study underway evalu-ating the efficacy of treating internal mammarynodes; efficacy and toxicity data, however, willnot be available for 15 or more years.

Detecting involved internal mammarynodes is difficult. The region is not clinicallyevaluable as are the axilla or supraclavicularfossa. Computed tomography scanning candetect grossly enlarged nodes but not micro-scopic disease in normal-sized nodes. Sentinellymph node mapping, on the other hand, mayresolve this issue.113

MULTIMODALITY THERAPY FOR LOCALLY ADVANCED

BREAST CANCER

Neoadjuvant chemotherapy is the subject ofintense clinical investigation for tumor down-staging before surgery. Patients who otherwisewould not be candidates for breast conservationtherapy may experience a significant reductionin tumor burden after neoadjuvant therapyallowing for local excision with negative mar-gins. Response rates are reported to be approx-imately 65 to 75 percent.114–116 Studies haveshown no detriment to overall survival whencompared to standard postoperative treatment.Complete clinical response may define a morefavorable subgroup than a partial responsealthough this has not been universally noted.117

What is the optimal therapy for womenexperiencing a complete response? In onestudy, those patients who had a complete clini-cal response to chemotherapy who thenreceived radiotherapy to the breast without

surgery had a higher local recurrence rate thanthose with partial response, local excision, andsubsequent radiotherapy.118 Residual disease,although not clinically detectable, may still beextensive in these patients. In a prospectiveevaluation of mastectomy specimens in patientswho had a complete clinical response to neoad-juvant chemotherapy and radiation therapy,residual disease ranging from 0.6 to 6.5 cm wasconfirmed in all specimens.119

What is the role of postmastectomy radiother-apy in patients who can not undergo BCT afterchemotherapy? Clinical staging often under-estimates the extent of disease when compared topathologic staging. Neoadjuvant therapy onlyconfounds the issue. Should regional lymphaticirradiation be used in all patients with locallyadvanced disease? There is currently an inter-group trial evaluating neoadjuvant therapy thatprohibits chest wall radiotherapy after mastec-tomy, regardless of pathologic findings. In lightof the recent data showing a survival advantagein selected patients treated with postmastectomyradiotherapy, this approach is difficult to justify.

The sequencing of chemotherapy and radio-therapy in all stages has been studied in detail.Current practice focuses on systemic therapy asthe top therapeutic priority since it has the great-est impact on survival. Concurrent chemother-apy and radiotherapy are prohibitively toxic.Although there are data showing a detrimentalimpact on local control from delaying radiother-apy more than 6 months after surgery,120 a ran-domized trial comparing upfront chemotherapyfollowed by radiation versus upfront radiother-apy followed by chemotherapy concluded thatdelaying chemotherapy increased distant metas-tases and adversely affected survival. Althoughdelayed radiotherapy increases local failure, asignificant percentage of patients experiencinglocal recurrence can be salvaged.121

TECHNIQUES OF RADIOTHERAPY

Breast irradiation after local excision is gener-ally administered with megavoltage photon


energies of 4 to 10 MV to deliver 45 to 50.4 Gyto the whole breast at 180 cGy per fraction. Tan-gential portals are established from midline tomidaxilla using wedge filters and a half beamblock or independent jaws to minimize pul-monary radiation. Many centers deliver a boostto the tumor bed using electron beam radiationfor an additional 900 to 1500 cGy. Alternatively,NSABP protocols prescribe 5000 cGy at 200cGy per fraction, with an optional boost. Superi-orly, the tail of the breast is encompassed whileexcluding the humoral head. Inferiorly, a reason-able margin of 1.5 to 2.0 cm below the infra-mammary fold is standard. The chest wall cur-vature necessitates treatment of a small volumeof lung tissue. If the maximum width of lung tis-sue treated is < 2 cm, the risk of pulmonaryinjury is exceptionally low (Figure 13–1).122

The boost treatment volume generallyencompasses the surgical bed with margins.Some centers advocate the use of clips at thetime of surgery to outline the surgical bed forboost treatment.123 Electron energies of 6 to15 MeV are used, depending on the depthneeded to encompass the tumor bed. The use ofinterstitial implants for the boost treatment hasbeen supplanted in recent years.

Three-dimensional treatment planning is amajor development in the delivery of radiother-apy. As the contour of the breast is irregular,there is potential for significant inhomogeneityof dose. Using CT scanning of the treatmentvolumes, technology exists to determine dosedelivery in three dimensions with modifica-

tions to maximize homogeneity of dosethroughout the breast tissue.

RADIOTHERAPY COMPLICATIONS

Dosimetric analyses of historic treatment tech-niques show an average of 25 percent of the car-diac volume received at least 50 percent of theprescribed dose. With modern treatment ener-gies, 5.7 percent of the cardiac volume receivesthis same dose.124 Serum troponin measure-ments during radiotherapy, a measure of cardiacinjury, revealed no significant elevation in onestudy.125 Patients treated for left-sided tumorswho also had internal mammary nodes treatedwith modern techniques were evaluated afterten years for possible late effects to the heart.Thallium stress tests revealed no statistical evi-dence of increased abnormality when comparedto the general population.126 No increased rateof myocardial infarction or cardiac-relateddeaths was seen in retrospective reviews ofpatients treated with tangential radiation.127,128

Other reported complications from breastradiation include lymphedema, rib fracture,brachial plexopathy, pulmonary fibrosis, car-cinogenesis, and contralateral breast cancer.Standard surgical therapy for invasive breastcancer includes level I and II axillary dissection,which has a finite risk of arm edema of 10 to 15percent.129,130 The addition of radiotherapy mayincrease the risk to as high as 20 percent. Themeasured incidence of arm edema likely differsfrom the incidence of clinically significant armedema. One study showed a direct relationshipbetween the incidence of arm edema and thenumber of lymph nodes dissected.131 There isless risk of lymphedema in a level I and level IIselective dissection than in a full axillary dis-section. It is hoped that sentinel node biopsywill eliminate this complication for patientswith histologically negative nodes. Any risk oflymphedema after sentinel node biopsy has yetto be determined. Brachial plexopathy is a com-plication in patients receiving supraclavicularradiation. The delivered dose is the significant

230 BREAST CANCER

Figure 13–1. Cross-sectional representation of tangentialradiation portals for breast conservation therapy.

risk factor as the incidence is reported to be < 2percent if 50 Gy is delivered, versus 5 percent if> 50 Gy is given.122 Rib fracture possibly relatedto beam energy has been reported after breastand chest wall radiotherapy. The incidence isvery low (< 0.5 percent), but 4 MV photons maybe associated with a higher incidence (2.2percent) than are higher energy beams.122

The chief rationale for the use of BCT overmastectomy is cosmesis. Poor cosmesis shouldbe considered a complication of treatment. Ingeneral, 75 to 85 percent of patients and treatingphysicians report good to excellent cosmesisafter lumpectomy and radiotherapy. Factors thatmay affect cosmesis include delivery of concur-rent chemotherapy, type and extent of surgery,and the dose of radiotherapy delivered.132,133

Two series have shown that radiotherapydelivered after breast augmentation can lead tocapsular contracture in 30 percent of patients.Although radiation following breast recon-struction is very well tolerated,134,135 thereported contracture rate for the general popu-lation is 5 to 10 percent. Subcutaneous implan-tation is associated with a higher risk of con-tracture than are subpectoral implants.Postmastectomy radiotherapy after transverserectus abdominus myocutaneous (TRAM) flapbreast reconstruction can be delivered effec-tively without cosmetic compromise.136

The carcinogenic potential of ionizingradiotherapy is rare and the latent period long.Follow-up intervals > 20 years are usuallynecessary to see carcinogenic effects. Mostlarge retrospective series report scatteredcases of sarcoma developing in the treatedfield years after treatment. Lymphangiosar-coma has been reported in patients who expe-rience significant lymphedema in the treatedbreast or ipsilateral arm. Osteosarcomas arefound to be most common within the treatedfield, with an overall incidence of sarcomadevelopment of 0.2 percent and a medianlatency of 11 years.137,138

Patients with breast cancer are at higher riskfor developing a contralateral breast cancer

than is the general population. Whether radio-therapy further increases that risk is the subjectof debate. Current technology virtually elimi-nates dose to the contralateral breast with tan-gential radiotherapy. However, the half-beamblock technique described earlier allowed forup to 400 to 500 cGy to be delivered to the con-tralateral breast. Hazard analysis of historicalseries revealed a small increased risk in irradi-ated patients compared to patients treated withsurgery alone.139

LOCAL RECURRENCE

Local recurrence has been shown to predict forsubsequent distant metastases, having animpact on survival.140 These data must be con-sidered in “very low risk” patients evaluated foromission of radiotherapy. Salvage treatmentwith mastectomy after ipsilateral breast recur-rence in patients treated with BCT offers sub-sequent disease-free survival in 40 to 70 per-cent of affected patients. Axillary recurrencesare rare but subsequent survival is lower (25 to30 percent).141 Failure in the supraclavicularfossa has a dismal prognosis.142 Short intervalto recurrence is likely to be predictive of subse-quent distant disease.143

There is a question whether development ofcancer in a treated breast many years after BCTrepresents relapse or a new primary cancer. Ithas been postulated that the location of therecurrence in relation to the primary, the timeinterval between incidents, and the presence orabsence of diffuse disease in the surroundingstroma may distinguish between recurrence anda new primary. The presence of a lesion nearthe original tumor site with disease in the sur-rounding stroma presenting relatively shortlyafter initial diagnosis more likely representsrecurrent disease than does a remote lesion pre-senting in normal surrounding tissue after along interval.144 Differences in DNA contentdetectable by flow cytometry are more likely ina second primary malignancy than in a recur-rent tumor. The subsequent treatment of the dis-


ease may differ, particularly in patients exposedto systemic treatment.

Standard therapy for patients experiencingrecurrence in an irradiated breast is mastectomy.However, there are limited data on the retreat-ment of patients with further excision and reirra-diation; 62 percent of patients were salvaged andtreatment was well tolerated, as 15 out of 16patients received 5000 cGy with no reporteddeleterious effect.145

The majority of patients who experiencerecurrence after local excision and no radio-therapy are treated with and often choose mas-tectomy over re-excision and radiotherapy assalvage treatment. If the overall goal of breastconserving treatment is breast retention, ther-apy should optimize the patient’s wish, even atthe expense of potential over-treatment.

CONCLUSION

Breast conservation therapy consisting of localexcision and breast radiotherapy should be thetreatment of choice for the majority of patientswith early invasive and in situ carcinoma.Radiotherapy to the breast has been shown to beeffective in improving local and regional con-trol, with minimal complications. The incidenceof intraductal and early stage breast cancer maycontinue to increase as mammographicallydetected lesions become more predominant.Radiotherapy delivery may become more selec-tive as methods are improved to identifypatients at high and low risk for local recur-rence. However, with the very small risk ofcomplication from treatment, the number ofpatients not receiving radiotherapy shouldremain very small. Recent data regarding post-mastectomy radiotherapy indicate that sub-groups of patients with locally advanced diseaseachieve a survival benefit from the addition ofpostmastectomy radiation. Modern radiotherapytechniques minimize cardiac injury. Identifyingpatients with significant risk for local failure,optimizing local treatment in these patients, andevaluating the impact on overall cure are areasfor current and future investigation.

REFERENCES1. Fletcher GH. Local results of irradiation in the

primary management of localized breast can-cer. Cancer 1972;29:545–51.

2. National Institutes of Health Consensus Develop-ment Panel. Consensus statement: treatment ofearly-stage breast cancer. J Natl Cancer InstMonogr 1992;11:15.

3. Cuzick J, Stewart H, Peto R, et al. Overview ofrandomized trials of adjuvant radiotherapy inbreast cancer. Recent Results Cancer Res1988;111:109–29.

4. Cuzick J, Stewart H, Rutqvist L, et al. Cause-spe-cific mortality in long-term survivors of breastcancer who participated in trials of radiother-apy. J Clin Oncol 1994;12:447–53.

5. Recht A, Bartelink H, Fourquet A, et al. Postmas-tectomy radiotherapy: questions for the twenty-first century. J Clin Oncol 1998;16:2886–9.

6. Lipsztein R, Dalton JF, Bloomer WD. Sequelae ofbreast irradiation. JAMA 1985;253:3582–4.

7. Garfinkel L, Boring C, Heath C Jr. Changingtrends: an overview of breast cancer incidenceand mortality. Cancer 1994;74 Suppl:222–7.

8. Bland KI, Menck HR, Scott-Connor CE, et al. TheNational Cancer Data Base 10-year survey ofbreast cancer treatment at hospitals in theUnited States. Cancer 1998;83:1263–73.

9. Bjurstrum N, Bjorneld L, Duffy SW, et al. TheGothenberg Breast Screening Trial. First resultson mortality, incidence, and mode of detectionfor women 39–49. Cancer 1997;80:2091–9.

10. Garne J, Aspegren K, Balldin G, Ranstam J.Increasing incidence of and declining mortalityfrom breast cancer. Cancer 1997;79:69–74.

11. Nattinger AB, Gottleib MS, Hoffman RG, et al.Minimal increase in the use of breast-conserv-ing surgery from 1986 to 1990. Med Care,1996;34:479–89.

12. Kotwall CA, Covington DL, Rutledge R, et al.Patient, hospital, and surgeon factors associ-ated with breast conserving surgery. Astatewide analysis in North Carolina. Ann Surg1996;224:419–26.

13. Morrow M, Bucci C, Rademaker A. Medical con-traindications are not a major factor in theunder utilization of breast conserving therapy.J Am Coll Surg 1998;186:269–74.

14. Fisher B, Anderson S, Redmond CK. Reanalysisand results after 12 years of follow-up in a ran-domized clinical trial comparing total mastec-tomy with lumpectomy with or without irradi-ation in the treatment of breast cancer. N EnglJ Med 1995;333:1456–61.

232 BREAST CANCER

15. Lichter AS, Lippmann ME, Danforth DN Jr, et al.Mastectomy versus breast conserving therapyin the treatment of stage I and II carcinoma ofthe breast: a randomized trial at the NationalCancer Institute. J Clin Oncol 1992;10:976–83.

16. Jacobson JA, Danforth DN, Cowan KH, et al. Ten-year results of a comparison of conservation withmastectomy in the treatment of stage I and IIbreast cancer. N Engl J Med 1995;332:901–11.

17. Veronesi U Luini A, Galimberti V, Zurrida S. Con-servation approaches for the management ofstage I/II carcinoma of the breast: Milan CancerInstitute Trials. World J Surg 1994;18:70–5.

18. Van Dongen JA, Bartelink H, Fentiman IS, et al.Randomized clinical trial to assess the value ofbreast-conserving therapy in stage I and IIbreast cancer: EORTC 10801 trial. J Natl Can-cer Inst 1992;11:15–8.

19. Blichert-Toft M, Rose C, Anderson JA, et al. Dan-ish randomized trial comparing breast conser-vation therapy with mastectomy: six years oflife table analysis. J Natl Cancer Inst Monogr1992;11:19–25.

20. Arriagada R, Le MG, Rochard F, Contesso G. Con-servative treatment versus mastectomy in earlybreast cancer: Patterns of failure with 15 years offollow-up data. Institut Gustave-Roussy BreastCancer Group. J Clin Oncol 1996;14:1558–64.

21. Early breast cancer trialists’ Collaborative Group.Effects of radiotherapy and surgery in earlybreast cancer. N Engl J Med 1995;333:1444–55.

22. Winchester DJ, Menck HR, Winchester DP. TheNational Cancer Data Base report on theresults of a large non-randomized comparisonof breast preservation and modified radicalmastectomy. Cancer 1997;80:162–7.

23. Cowen D, Jacquemier J, Houvenaeghel G, et al.Local and distant recurrence after conservativemanagement of “very low risk” breast cancerare dependant events: a 10-year follow-up. IntJ Radiat Oncol Biol Phys 1998;41: 801–7.

24. Elkhuizen PH, Van de Vijver MJ, Hermans J, et al.Local recurrence after breast conserving ther-apy for invasive cancer; high incidence inyoung patients and association with poor sur-vival. Int J Radiat Oncol Biol Phys 1998;40:859–67.

25. Fowble BL, Schultz DJ, Overmoyer B, et al. Theinfluence of young age on outcome in earlystage breast cancer. Int J Radiat Oncol BiolPhys 1994;30:23–33.

26. Kim SH, Simkovich-Heerdt A, Tran KN, et al.Women 35 years of age have higher locoregional

relapse rates after undergoing breast conserva-tion therapy. J Am Coll Surg 1998;187:1–8.

27. Matthews RH, McNeese MD, Montague ED,Oswald MJ. Prognostic implications of age inbreast cancer patients treated with tumorec-tomy and irradiation or with mastectomy. Int JRadiat Oncol Biol Phys 14:659–63.

28. Donegan WL, Perez-Mesa CM, Watson FR. A bio-statistical study of locally recurrent breast car-cinoma. Surg Gyn Obstet 1966;122:329.

29. Fowble B. Radiotherapeutic considerations in thetreatment of primary breast cancer. J Natl Can-cer Inst Monogr 1992;11:49–58.

30. Recht A, Harris JR. Selection of patients with earlystage breast cancer for conservative surgeryand radiation. Oncology 1990;4:23–30.

31. Leopold K, Recht A, Schnitt SJ, et al. Results ofconservative surgery and radiation therapy formultiple synchronous cancers of one breast. IntJ Radiat Oncol Biol Phys 1989;16:11–16.

32. Fowble B, Yeh I, Schultz DJ, et al. The role of mas-tectomy in patients with stage I–II breast can-cer presenting with gross multifocal or multi-centric disease or diffuse calcifications. Int JRadiat Oncol Biol Phys 1993;27:567–73.

33. Kurtz J. Factors influencing the risk of local recur-rence in the breast. Eur J Cancer 1992;28:660–6.

34. Schnitt SJ, Connolly J, Khettry V, et al. Pathologicfindings on reexcision of the primary site inbreast cancer patients considered for treatmentby primary radiation therapy. Cancer 1997;59:675–81.

35. Hurd TC, Sneige N, Allen PK, et al. Impact ofextensive intraductal component on recurrencein patients with stage I or II breast cancertreated with breast conservation therapy. AnnSurg Oncol 1997;4:119–24.

36. Schnitt SJ, Abner A, Gelman R, et al. The rela-tionship between microscopic margins ofresection and the risk of local recurrence inpatients with breast cancer treated with breast-conserving surgery and radiation therapy. Can-cer 1994;74:1746–51.

37. Moran M, Haffty BG. Lobular carcinoma in situas a component of breast cancer: the long-termoutcome in patients treated with breast-con-serving therapy. Int J Radiat Oncol Biol Phys1998;40:353–8.

38. Schnitt SJ, Connolly JL, Recht A, et al. Influenceof infiltrating lobular histology on local tumorcontrol in breast cancer patients treated withconservative surgery and radiotherapy. Cancer1989;64:448–54.


39. Haffty BG, Wilson LD, Smith R, et al. Subareolarbreast cancer: long-term results with conserva-tive surgery and radiation therapy. Int J RadiatOncol Biol Phys 1995;33:53–7.

40. Bussieres E, Guyon F, Thomas L, et al. Conserva-tion treatment in subareolar breast cancers. EurJ Surg Oncol 1996;22:267–70.

41. Dale PS, Giulliano AE. Nipple-areolar preservationduring breast-conserving therapy for subareolarbreast carcinomas. Arch Surg 1996;131:430–3.

42. Pierce LJ, Haffty BG, Solin LJ, et al. The conserva-tive management of Paget’s disease of the breastwith radiotherapy. Cancer 1997;80:1065–72.

43. Nixon AJ, Schnitt SJ, Gelman R, et al. Relation-ship of tumor grade to other pathologic featuresand to treatmetn outcome of patients with earlystage breast cancer. Cancer 1996;78:1426–31

44. Chabner E, Nixon A, Gelman R, et al. Family his-tory and treatment outcome in young womenafter breast-conserving surgery and radiationtherapy for early-stage breast cancer. J ClinOncol 1998;16:2045–51.

45. Malone KE, Daling JR, Weiss NS, et al. Family his-tory and survival of young women with invasivebreast carcinoma. Cancer 1996;78:1417–25.

46. Porter D, Cohen B, Wallace M, et al. Breast can-cer incidence, penetrance and survival in prob-able carriers of BRCA1 gene mutation in fam-ilies linked to BRCA1 on chromosome17q12-21. Br J Surg 1994;81:1512–15.

47. Anscher MS, Jones P, Prosnitz LR, et al. Local fail-ure and margin status in early stage breast carci-noma treated with conservative surgery andradiation therapy. Ann Surg 1993;218:22–8.

48. Ryoo MC, Kagan AR, Wollin M, et al. Prognosticfactors for recurrence and cosmesis in 393patients after radiotherapy. Radiology 1989;172:555–9.

49. Heimann R, Powers C, Halpern HJ, et al. Breastpreservation in stage I and II carcinoma of thebreast. The University of Chicago experience.Cancer 1996;78:1722–30.

50. Spivack B, Khanna MM, Tafra L, et al. Marginstatus and local recurrence after breast-con-serving surgery. Arch Surg 1994;129: 952–6.

51. Gage I, Schnitt SJ, Nixon AJ, et al. Pathologicmargin involvement and the risk of recurrencein patients treated with breast-conservingsurgery. Cancer 1996;78:1921–8.

52. DiBiase SJ, Komarnicky LT, Schwartz GF, et al.The number of positive margins influences theoutcome of women treated with breast preser-vation of early stage breast carcinoma. Cancer1998;82:2212–20.

53. Guidi AJ, Connolly JL, Harris JR, Schnitt SJ. Therelationship between shaved margin and inkedmargin status in breast excision specimens.Cancer 1997;9:1568–73.

54. Holland R, Veling SH, Mravunac M, Hendriks JH.Histologic multifocality of Tis, T1-2 breast car-cinomas. Implications for clinical trials ofbreast conserving surgery. Cancer 1985:56:979–90.

55. Wazer DE, Schmidt-Ullrich RK, Ruthazer R, etal. Factors determining outcome for breast-conserving irradiation with margin-directeddose escalation to the tumor bed. Int J RadiatOncol Biol Phys 1998;40:851–8.

56. Fagundes MA, Fagundes HM, Brito CS, et al.Breast-conserving surgery and definitive radi-ation: a comparison between quadrantectomyand local excision with special focus on loco-regional control and cosmesis. Int J RadiatOncol Biol Phys 1993;27:553–60.

57. Veronesi U, Volterrani F, Luini A, et al. Quadran-tectomy versus lumpectomy for small breastcancer. Eur J Cancer 1990;26:671–3.

58. Ross J, Hussey D, Mayr N, Davis CS. Acute andlate reactions to radiation therapy in patientswith collagen vascular diseases. Cancer 1993;71:3744–52.

59. Antypas C, Sandilos P, Kouvaris J, et al. Fetal doseduring breast cancer radiotherapy. Int J RadiatOncol Biol Phys 1998;40:995–9.

60. Clark RM, Wilkinson RH, Miceli PN, MacDonaldWD. Breast cancer: experiences with conserva-tion therapy. Am J Clin Oncol 1987;10:461–8.

61. Romestaing P, Lehingue Y, Carrie C, et al. Role of10-Gy boost in the conservative treatment ofearly breast cancer: results of a randomizedclinical trial in Lyon, France. J Clin Oncol1997;15:963–8.

62. Solin LJ, Fowble BL, Schultz DJ, Goodman RL.The significance of pathology margins oftumor excision on the outcome of patientstreated with definitive irradiation for earlystage breast cancer. Int J Radiat Oncol BiolPhys 1991;21:279–87.

63. Smitt MC, Nowels KW, Zdeblick MJ, et al. Theimportance of the lumpectomy surgical marginstatus in long-term results of breast conserva-tion. Cancer 1995;76:259–67.

64. Fisher B, Digham J, Mamounas E, et al. Sequentialmethotrexate and fluorouracil for the treatmentof node-negative breast cancer patients withestrogen receptor-negative tumors: eight-yearresults from the National Surgical AdjuvantBreast and Bowel Project (NSABP) B-13 and

234 BREAST CANCER

first report of findings from NSABP B-19 com-paring methotrexate and fluorouracil with con-ventional cyclophosphamide, methotrexate, andfluorouracil. J Clin Oncol 1996;14:1982–92.

65. Fisher BJ, Parera FE, Cooke AL, et al. Long-termfollow-up of axillary node positive breast can-cer patients receiving adjuvant systemic ther-apy alone: patterns of recurrence. Int J RadiatOncol Biol Phys 1997;38:541–50.

66. Fisher BJ, Perera FE, Cooke AL, et al. Long-termfollow-up of axillary node positive breast can-cer patients receiving adjuvant tamoxifenalone: patterns of recurrence. Int J RadiatOncol Biol Phys 1998;42:117–23.

67. Cooke AL, Parera F, Fisher B, et al. Tamoxifenwith and without radiation after partial mastec-tomy in patients with involved nodes. Int JRadiat Oncol Biol Phys 1995;31:777–81.

68. Dalberg K, Johansson H, Johansson U, RutqvistLE. A randomzed trial of long term adjuvanttamoxifen plus postoperative radiation therapyversus radiation therapy alone for patients withearly stage breast cancer treated with breastconserving surgery, Stockholm Breast CancerStudy Group. Cancer 1998;82:2204–11.

69. Forrest P, Stewart H, Everington D, et al. Ran-domized controlled trial of conservation ther-apy for breast cancer: 6-year analysis of theScottish trial. Scottish Cancer Trials BreastGroup. Lancet 1996;348:708–13.

70. Turner BC, Haffty BG, Narayanam L, et al. Insulin-like growth factor-I receptor overexpressionmediates cellular radioresistance and localbreast cancer recurrence after lumpectomy andradiation. Cancer Res 1997;57:3079–83

71. Haffty BG, Brown F, Carter D, Flynn S. Evaluationof HER-2 neu oncoprotein expression as aprognostic indicator of local recurrence in con-servatively treated breast cancer: a case-controlstudy. Int J Radiat Oncol Biol Phys 1996;35:751–7.

72. Sjogren S, Inganas M, Lindgren A, et al. Prognos-tic and predictive value of c-erbB-2 overex-pression in primary breast cancer, alone and incombination with other prognostic markers. JClin Oncol 1998;16:462–9.

73. Gaffney DK, Brohet KM, Lewis CM, et al.Response to radiation therapy and prognosis inbreast cancer patients with BRCA1 and BRCA2mutations. Radiother Oncol 1998;47:129–36.

74. Formenti SC, Dunnington G, Uzieli B, et al. Orig-inal p53 status predicts for pathologicalresponse in locally advanced breast cancertreated preoperatively with continuous infusion

5-fluorouracil and radiation therapy. Int JRadiat Oncol Biol Phys 1997;39:1059–68.

75. Berns M, Klijn J, van Putten WL, et al. P53 pro-tein accumulation predicts poor response totamoxifen therapy of patients with recurrentbreast cancer. J Clin Oncol 1998;16:121–7.

76. Gorski DH, Mauceri HJ, Salloum RM, et al.Potentiation of the antitumor effect of ionizingradiation by brief concomitant exposures toangiostatin. Cancer Res 1998;58:5686–9.

77. Heimann R, Ferguson D, Powers C, et al. Angio-genesis as a predictor of long-term survival forpatients with node negative breast cancer. JNatl Cancer Inst 1998;88:1764.

78. Clark RM, Whelan T, Levine M, et al. Random-ized clinical trial of breast irradiation follow-ing lumpectomy and axillary dissection fornode-negative breast cancer. An update. J NatlCancer Inst 1996;88:1659–64.

79. Veronesi U. Breast cancer trials on conservativesurgery. Eur J Surg Oncol 1995;21:231–3.

80. Veronesi U, Luini A, Del Vecchio M, et al. Radio-therapy after breast preserving surgery inwomen with localized cancer of the breast. NEngl J Med 1993;328:1587–91.

81. Liljegren G, Holmberg L, Adami HO, et al. Sectorresection with or without postoperative radio-therapy for stage I breast cancer. Five yearresults of a clinical trial. J Natl Cancer Inst1994;86:717–22.

82. Nemoto T, Patel JK, Rosner D, et al. Factorsaffecting recurrence in lumpectomy withoutirradiation for breast cancer. Cancer 1991;67:2079–82.

83. Morrow M, Harris JR, Schnitt SJ. Local controlfollowing breast-conserving surgery for inva-sive breast cancer: results of clinical trials. JNatl Cancer Inst 1995;87:1669–73.

84. Liljegren G, Lindgren A, Bergh J, et al. Risk fac-tors for local recurrence after conservativetreatment in stage I breast cancer. Definition ofa subgroup not requiring radiotherapy. AnnOncol 1997;8:235–41.

85. Schnitt SJ, Hayman J, Gelman R, et al. A prospec-tive study of conservative surgery alone in thetreatment of selected patients with stage Ibreast cancer. Cancer 1996;77:1094–1100.

86. Solin LJ, Schultz DJ, Fowble BL. Ten year resultsof the treatment of early-stage breast carci-noma in elderly women using breast-conserv-ing surgery and definitive breast irradiation.Int J Radiat Oncol Biol Phys 1995;33:45–51

87. Merchant TE, McCormick B, Yahalom J, BorgenP. The influence of older age on breast cancer


236 BREAST CANCER

treatment decisions and outcome. Int J RadiatOncol Biol Phys 1996;34:565–70.

88. Holland R, Hendricks JH, Vebeek AL, et al.Extent, distribution and mammographic histo-logical correlations of breast ductal carcinomain situ. Lancet 1990;335:519–22.

89. Fisher B, Anderson S. Conservative surgery forthe management of invasive and non-invasivecarcinoma of the breast: NSABP trials.National Surgical Adjuvant Breast and BowelProject. World J Surg 1994;18:63–9.

90. Fisher B, Constantino J, Redmond C, et al.Lumpectomy compared with lumpectomy andradiation therapy for the treatment of intraductalbreast cancer. N Engl J Med 1993;328:1581–6.

91. Fisher B, Dignam J, Wolmark N, et al. Lumpec-tomy and radiation therapy for the treatment ofintraductal breast cancer: findings from theNational Surgical Adjuvant Breast and BowelProject B-17. J Clin Oncol 1998;16:441–52.

92. Fisher E, Constantino J, Fisher B. Pathologic find-ings from the National Surgical AdjuvantBreast Project (NSABP) protocol B-17. Intra-ductal carcinoma (ductal carcinoma in situ).Cancer 1995;75:1310–9.

93. Fowble B, Hanlon AL, Fein DA, et al. Results ofconservative surgery and radiation for mam-mographically detected ductal carcinoma insitu (DCIS). Int J Radiat Oncol Biol Phys1997;38:949–57.

94. Solin LJ, Kurtz J, Fourquet A, et al. Fifteen-yearresults of breast conserving surgery and defin-itive breast irradiation for the treatment of duc-tal carcinoma in situ of the breast. J Clin Oncol1996;14:754–63.

95. Vicina FA, Lacerna MD, Goldstein NS, et al.Ductal carcinoma in situ detected in the mam-mographic era: an analysis of clinical, patho-logic, and treatment related factors affectingoutcome with breast-conserving therapy. Int JRadiat Oncol Biol Phys 1997;39:627–35.

96. Silverstein MJ, Lagios MD, Craig PH, et al. Aprognostic index for ductal carcinoma in situ ofthe breast. Cancer 1996;77:2267–74.

97. Silverstein MJ, Lagios MD. Use of predictors toplan therapy for DCIS of the breast. Oncology1997;11:393–406.

98. Lagios MD, Silverstein MJ. Ductal carcinoma insitu. The success of breast conservation ther-apy: a shared experience of two single institu-tional non-randomized prospective studies.Surg Oncol Clin North Am 1997;6:385–92.

99. Winchester DP, Strom EH. Standards for diagno-sis and management of ductal carcinoma insitu of the breast. CA-A 1998;48:108–28.

100. Silverstein MJ, Lagios MD, Martino S, et al. Out-come after invasive local recurrence in patientswith ductal carcinoma in situ of the breast. JClin Oncol 1998;16:1367–73.

101. Auquier A, Rutqvist LE, Host H, et al. Post-mas-tectomy megavoltage radiotherapy: the Oslo andStockholm trials. Eur J Cancer 1992;28:433–7.

102. Gagliardi G, Lax I, Soderstrom S, et al. Predic-tion of excess risk of long-term cardiac mortal-ity after radiotherapy of stage I breast cancer.Radiother Oncol 1998;46:63–71.

103. Kuske R. Adjuvant chest wall and nodal irradia-tion: maximize cure, minimize late toxicity[editorial]. J Clin Oncol 1998;16:2579–82.

104. Arriagada R, Rutqvist LE, Mattson A, et al. Ade-quate locoregional treatment for early breastcancer may prevent secondary dissemination. JClin Oncol 1995;13:2869–78.

105. Overgaard M, Hansen PS, Overgaard J, et al.Postoperative radiotherapy in high-risk pre-menopausal women with breast cancer whoreceive adjuvant chemotherapy. N Engl J Med1997;337:949–55.

106. Ragaz J, Jackson SM, Le N, et al. Adjuvant radio-therapy and chemotherapy in node-positivepremenopausal women with breast cancer. NEngl J Med 1997;337:956–62.

107. Recht A, Bartelink H, Fourquet A, et al. Post-mastectomy radiotherapy: questions for thetwenty-first century. J Clin Oncol 1998;16:2886–9.

108. Fowble B, Gray R, Gilchrist K, et al. Identifica-tion of a subgroup of patients with breast can-cer and histologically positive axillary lymphnodes receiving adjuvant chemotherapy whomay benefit from postoperative radiotherapy. JClin Oncol 1988;6:1107–17.

109. Fowble B. Post-mastectomy radiotherapy: thenand now. Oncology 1997;11:213–31.

110. Kozeniowski S. “One to three” or “four ormore”? Selecting patients for post-mastectomyradiation therapy. Cancer 1997;80:1357–8.

111. Fisher BF, Perera F, Cooke A, et al. Extracapsularaxillary node extension in patients receivingadjuvant systemic therapy: an indication forradiotherapy? Int J Radiat Oncol Biol Phys1997;38:551–9.

112. Donegan WL, Stine SB, Samter TG. Implicationsof extracapsular nodal metastases for treatmentand prognosis of breast cancer. Cancer 1993;72:778–82.

113. Krag D, Weaver D, Ashikaga T, et al. The sentinelnode in breast cancer. N Engl J Med 1998;339:941–6.


115. Fisher B, Brown A, Mamounas E, et al. Effect ofpreoperative chemotherapy on local-regionaldisease in women with operable breast cancer:findings from National Surgical AdjuvantBreast and Bowel Project B-18. J Clin Oncol1997;15:2483–93.

116. Jacquillat C, Weil M, Baillet F, et al. Results ofneoadjuvant chemotherapy and radiation ther-apy in the breast conserving treatment of 250patients with all stages of infiltrative breastcancer. Cancer 1990;66:119–29.

117. Brenin DR, Morrow M. Breast conservationsurgery in the neo-adjuvant setting. SeminOncol 1998;25:13–18.

118. Ellis P, Smith I, Ashley S, et al. Clinical, prog-nostic and predictive factors for primarychemotherapy in operable breast cancer. J ClinOncol 1998;16:107–14.

119. Mumtaz H, Davidson T, Spittle M, et al. Breastsurgery after neoadjuvant treatment. Is it nec-essary? Eur J Surg Oncol 1996;22:335–41.

120. Bucholz TA, Austin-Seymour MM, Moe RE, etal. Effect of delay in radiation in the combinedmodality treatment of breast cancer. Int JRadiat Oncol Biol Phys 1993;26:23–5.

121. Recht A, Come SE, Henderson IC, et al. Thesequencing of chemotherapy and radiation ther-apy after conservative surgery for early-stagebreast cancer. N Engl J Med 1996;334:1356–61.

122. Pierce S, Recht A, Lingos T, et al. Long-termradiation complications following conservativesurgery (CS) and radiation therapy (RT) inpatients with early stage breast cancer. Int JRadiat Oncol Biol Phys 1992;23:915–23.

123. Fein DA, Fowble BL, Hanlon AL, et al. Does theplacement of surgical clips within the excisioncavity influence local control for patients treatedwith breast-conserving surgery and irradiation.Int J Radiat Oncol Biol Phys 1996;34:1009–17.

124. Geynes G, Gagliardi G, Lax I, et al. Evaluation ofirradiated heart volumes in stage I breast can-cer patients treated with postoperative adjuvantradiotherapy. J Clin Oncol 1997;15:1348–53.

125. Hughes-Davies L, Sacks D, Rescigno J, et al.Serum cardiac troponin T levels during treat-ment of early-stage breast cancer. J Clin Oncol1995;13:2582–4.

126. Cowen D, Gonzague-Casabianca L, Brenot-RossiI, et al. Thallium-201 perfusion scintigraphy inthe evaluation of late myocardial damage in

left-side breast cancer treated with adjuvantradiotherapy. Int J Radiat Oncol Biol Phys1998;41:809–15.

127. Rutqvist LE, Liedberg A, Hammar N, Dalberg K.Myocardial infarction among women withearly-stage breast cancer treated with conserv-ative surgery and breast irradiation. Int JRadiat Oncol Biol Phys 1998;40:359–63.

128. Nixon AJ, Manola J, Gelman R, et al. No long-term increase in cardiac related mortalityafter breast-conserving surgery and radiation therapy using modern techniques. J Clin Oncol1998;16:1374–9.

129. Hoe AL, Iven D, Royle GT, Taylor I. Incidence ofarm swelling following axillary clearance forbreast cancer. Br J Surg 1992;79:261–2.

130. Siegel BM, Mayzel KA, Love SM. Level I and IIaxillary dissection in the treatment of early-stage breast cancer. An analysis of 259 consec-utive patients. Arch Surg 1990;125:1144–7.

131. Kiel KD, Rademaker AW. Early-stage breast can-cer: arm edema after wide excision and breastirradiation. Radiology 1996;198:279–83.

132. Taylor ME, Perez CA, Halverson KJ, et al. Fac-tors influencing cosmetic results after conser-vation therapy for breast cancer. Int J RadiatOncol Biol Phys 1995;31:756–64.

133. Olivotto IA, Weir LM, Kim-Sing C, et al. Latecosmetic results of short fractionation for breastconservation. Radiother Oncol 1996;41:7–13.

134. Victor SJ, Brown DM, Horwitz EM, et al. Treat-ment outcome with radiation therapy afterbreast augmentation or reconstruction inpatients with primary breast carcinoma. Can-cer 1998;82:1303–9.

135. Ryu J, Yahalom J, Shank B, et al. Radiation ther-apy after breast augmentation or reconstructionin early or recurrent breast cancer. Cancer1990;66:844–7.

136. Hunt KK, Baldwin BJ, Strom EA, et al. Feasibil-ity of postmastectomy radiotherapy afterTRAM flap breast reconstruction. Ann SurgOncol 1997;4:377–84.

137. Karlsson P, Holmberg E, Johansson KA, et al.Soft tissue sarcoma after treatment for breastcancer. Radiother Oncol 1996;38:25–31.

138. Pendlebury SC, Bilous M, Langlands AO. Sarco-mas following radiation therapy for breast can-cer: a report of three cases and a review of theliterature. Int J Radiat Oncol Biol Phys 1995;31:405–16.

139. Boice JD, Harvey EB, Blettner M, et al. Cancer inthe contralateral breast after radiotherapy forbreast cancer. N Engl J Med 1992;326:781–5.

140. Touboul E, Buffat L, Belkacemi Y, et al. Local


recurrences and distant metastases after breast-conserving surgery and radiation therapy forearly breast cancer. Int J Radiat Oncol BiolPhys 1999;43:25–38.

141. Leborgne F, Leborgne JH, Ortega B, et al. Breastconservation treatment of early stage breastcancer: patterns of failure. Int J Radiat OncolBiol Phys 1995;31:765–75.

142. Willner J. The role of radiation therapy in themultidisciplinary management of recurrent andmetastatic breast cancer [correspondence].Cancer 1995;75:902–3.

143. Haffty BG, Reiss M, Beinfield M, et al. Ipsilat-

eral breast tumor recurrence as a predictor ofdistant disease: implications for systemic ther-apy at the time of local relapse. J Clin Oncol1996;14:52–7.

144. Haffty BG, Carter D, Flynn SD, et al. Localrecurrence versus new primary: clinical analy-sis of 82 breast relapses and potential applica-tions for genetic fingerprinting. Int J RadiatOncol Biol Phys 1993;27:575–83.

145. Mullen E, Deutsch M, Bloomer WD. Salvageradiotherapy for local failures of lumpectomyand breast irradiation. Radiother Oncol 1997;42:25–9.

238 BREAST CANCER

239

Breast cancer is not entirely a disease ofwomen. 0.7 percent of all cases of breast cancerin the United States occur in males. Menaccount for 0.9 percent of deaths from breastcancer. About 0.2 percent of cancers in malesarise in the breast. An estimated 1,300 newcases of breast cancer in men are expected in theUnited States, in 1999, and 400 men will die ofthe disease. Breast cancer causes 0.14 percent ofdeaths from cancer in men.1

The earliest record of cancer of the breast,which dates from the Edwin Smith surgicalpapyrus, describes the disease in a man. ThisEgyptian antiquity, written circa 3000 to 2500BC, indicates that no known treatment was suc-cessful for bulging tumors of the breast.2 Hollebattributes the first documented case of malebreast cancer to John of Arderne in England inthe fourteenth century;3 Meyskens attributes itto William Fabry of Germany in the sixteenthcentury.4 Clinical descriptions of male breastcancer began to appear in medical journals inFrance and England in the early nineteenth cen-tury. Considered a curiosity, male breast cancerreceived little attention until later that centurywhen collections of cases began to appear in theliterature.5 In 1883, Porier published a detaileddescription of the clinical evolution of breastcancer in males that leaves little room forimprovement. In 1927, Wainwright in Pennsyl-vania was able to report on 418 collected cases.5

He described the poor prognosis associated withhigh histologic grade, cutaneous ulceration, andaxillary node involvement. Postoperative mor-tality was 6.1 percent; only nineteen percent of111 cases with complete follow-up survived 5years. Wainwright concluded that the prognosiswas not as good in men as in women.

Reports from numerous countries documentthe pervasiveness of male breast cancer.Because of the low incidence of breast cancer inmen, information on the subject is based largelyon case reports and retrospective analyses ofdata collected in medical centers and tumor reg-istries over many years. Few individual physi-cians have personal experience with more than asmall number of cases. Clinical trials of treat-ment are nonexistent for males, and in mostrespects, advances in treatment are translationsfrom lessons learned about the disease inwomen, the prevalence of which has enabledcontrolled studies. As a consequence, guidelinesfor treatment of women are being used formales, and men are now receiving less radicaloperations and more effective systemic adjuvanttherapy. Similarly, the role of hereditary factorsand gene mutations are being explored in males.

It has recently come to be appreciated thatbreast cancer is similar in both sexes. Earlyreports emphasized differences, but accumulat-ing information makes it clear there are moresimilarities than differences; this change is

14

Carcinoma of the Breast in MenPHILIP N. REDLICH, MD, PHDWILLIAM L. DONEGAN, MD

important not only for discovering causationbut for application of treatment. While the eti-ology of breast cancer in both sexes isunknown, there is little reason to believe that itis different. Men and women are subject to sim-ilar environmental exposures. The pathologyand clinical courses are parallel, and in similarcircumstances, men and women prove equallycurable. The older age of men at diagnosis, sub-areolar origin of the tumor, and presentation inmore advanced stages with poorer overall prog-nosis can be attributed to the small size of themale breast and the scant notice it receives.Screening for early detection does not exist formen, but public and professional awareness ofbreast cancer in men, and appropriate applica-tion to men of the intensive research of the dis-ease in women, should result in progress.

EPIDEMIOLOGY

The age-specific incidence and mortality ofbreast cancer rise steadily in males beginning inthe third decade.6 The disease has been diag-nosed in teenagers, but cases usually begin toappear in the fourth decade of life, with theaverage age at diagnosis in large series cluster-ing around 65 years, 5 to 8 years older than theaverage age of women at diagnosis.7,8 Wide ageranges are reported, from 23 to 97 years.9 Inreports from various countries, the incidence inmen parallels that of women.10,11 High rates arereported in England and Wales and low rates inJapan and Finland. Black races in subSaharanAfrica have a high frequency of affected males,often attributed to a high prevalence of liverdisease which leads to alterations in estrogenmetabolism. Males account for 7 percent ofcases of breast cancer in Tanzania12 and 9 per-cent of cases in Nigeria.13 The lower averageage at diagnosis in African countries also sug-gests an earlier onset.

Factors identified with high risk for men arefragmentary and sometimes controversial butsuggest genetic, hormonal, and environmentalinfluences. In many respects, they reflect the risk

patterns known for women (Table 14–1). Casecontrol studies associate high risk variously withhigh socioeconomic status, higher levels of edu-cation, Ashkenazi Jewish descent, childlessness,obesity, limited exercise, tallness, and consump-tion of red meat.14,15 Linkage between malebreast cancer and exposure to low frequencymagnetic fields has not been confirmed.16

From 11 to 27 percent of affected malesreport a family history of breast cancer.17,18–20

Families with high rates of breast cancer some-times include affected males; multiple males maybe affected and males in more than one genera-tion of such families have developed breast can-cer.21 Female descendents of males with breastcancer are at increased risk, indicating transmis-sion through the male line. In males, inheritanceof breast cancer risk has been linked to germlinemutations in the BRCA2 gene on chromosome13q12-13. Between 35 and 45 percent of familialbreast cancer can be accounted for by BRCA2mutations, often including families in whichboth males and females are affected. For femaleswho are members of high-risk families, muta-tions in BRCA1 or BRCA2 carry an estimated 56to 85 percent lifetime risk of breast cancer. Lim-ited figures indicate that from 4 to 43 percent ofmales with breast cancer carry various mutationson chromosome 13q. Family history of breastcancer is usually present in reported series ofmale breast cancer patients with BRCA2 muta-tions but the frequency of a positive family his-tory ranges up to 85 percent.18,22,23 Mutations inthe androgen-receptor (AR) gene associatedwith androgen insensitivity syndrome are alsolinked to male breast cancer.24,25

Men with Klinefelter’s syndrome (obesity,hypogonadism, aspermatogenesis, increasedurinary gonadotropins, and gynecomastia),identified by an XXY karyotype are estimatedto have a 20- to 50-fold increase in risk forbreast cancer and a 3 percent lifetime risk.26–29

Nevertheless, cases of Klinefelter’s syndromeare not regularly found in reported series ofmales with breast cancer; the frequency varieswidely, from 0 to 7.5 percent.

240 BREAST CANCER

Endocrine abnormalities are not often foundin males with breast cancer, but available infor-mation suggests some role for excess estrogen ora deficiency of androgen. High levels of endoge-nous estrogens may result from obesity and livercirrhosis, which are often associated with malebreast cancer in Denmark and in African coun-tries.30,31 Testicular function declines with agingas the incidence of breast cancer rises. Breastcancer has been reported in three orchiec-tomized male transsexuals treated with estrogento enhance breast development.32,33 Crichlowcites four cases of breast cancer in men treatedwith estrogens for prostate cancer; more fre-quent than primary breast carcinoma, however,among men with prostatic carcinoma is metasta-tic involvement of the breast.34 In one series, nobreast cancers were seen in over 4,000 malestreated with estrogens for prostate cancer, butdurations of exposure may have been relativelyshort. Androgen deficiency is suggested by thefrequent histories of orchitis, inguinal hernior-rhaphy, mumps infections in adulthood, orchiec-tomy, and testicular injury among men withbreast cancer.35 Impaired testicular function mayresult from occupational exposure to high envi-ronmental temperatures and chemicals, whichis reported by many affected men.

There have been a number of reports associ-ating male breast cancer with chronic hyper-prolactinemia. Such cases have included bilat-eral breast involvement36a and a history ofprolactinoma and head injury.36b The preciserole of prolactin and any associated endocrinedisturbances is undetermined.

Case reports document primary breast can-cer in men after exposure of the breast to ioniz-ing radiation, in one case to treat pubertalgynecomastia.37,38 Radiation is known to be car-cinogenic for the breasts of women, particularlywith exposure early in life. Women exposed toatomic radiation or to multiple fluoroscopies inthe course of treatment for tuberculosis, or whohave been irradiated for mastitis or treated withradiation for Hodgkin’s disease, are known to beat increased risk. Reid and colleagues found a

prior history of breast irradiation in 3.1 percentof 229 men with breast cancer.9

Ductal and lobular development of the malebreast from genetic, environmental, or endoge-nous causes may place it at increased risk forcarcinogenesis. Up to 40 percent of breast can-cers in males are associated with gynecomastia.This relationship is inconclusive in view of thehigh frequency of gynecomastia in adult males.Noteworthy, however, is the parallel increase ofbreast cancer and gynecomastia in men withaging and the derivation of cancers from ductaland lobular elements when present. Ductal hyper-plasia is often seen in association with ductalcarcinoma in males, and in situ and invasive lob-ular carcinoma has been seen with Klinefelter’ssyndrome39 and after chronic cimetidine stimu-lation of the male breast.40 The presence ofgynecomastia and the influences that produce itare often indistinguishable.

Of importance in the epidemiology of breastcancer in men is freedom of men from theunique reproductive functions of women thatare so prominent in risk for breast cancer. The

Carcinoma of the Breast in Men 241

Table 14–1. ASSOCIATIONS WITH INCREASED RISKFOR BREAST CANCER IN MEN

GeneticFirst-degree relatives with breast cancerAshkenazi Jewish descentBRCA2 gene mutationsKlinefelter’s syndromeAndrogen insensitivity

Environmental exposuresIonizing radiationEstrogens

Occupational exposuresSoap and perfume workersBlast furnace workers and steelworkers

Reduced testicular functionMumps orchitisInguinal herniorrhaphyUndescended testesGynecomastia

HyperprolactinemiaHead traumaHyperprolactinemia

OtherHigh body weight early in lifeHigh socioeconomic statusHigher educationChildlessness

absence of these promoters is potentially usefulin providing a less cluttered view of the disease.The fact remains that in the majority of men orwomen, no special risk feature is evident otherthan age. Avoiding potential mammary carcino-gens and aspiring to a low-risk profile are somelessons in prevention derived from studies ofbreast cancer in males. Fortunately, ionizingradiation is no longer used to treat pubertalgynecomastia, acne, and other benign condi-tions of youth. Hormonal stimulation of themale breast and obesity are avoidable. Identifi-cation of individuals with an inherited high riskfor breast cancer through genetic testing canpermit more informed decisions about prophy-lactic mastectomy for men.41

PATHOLOGY

The same histologic types of breast canceroccur in men and women but the frequencies ofthese types vary (Table 14–2). Noninvasive duc-tal carcinoma has been described either in a

pure form or mixed with an invasive compo-nent.42,43 Ductal carcinoma in situ (DCIS) com-prises approximately 5 percent of all cases ofmale breast carcinoma but ranges as high as 17percent in reported series.43 The median age ofoccurrence of DCIS is usually the late 50s tomid-60s but has been reported in men under theage of 40 years. The most frequent histologicpattern is the papillary subtype, with the major-ity of cases being of low or intermediate grade.In a recent review of Paget’s disease, this histo-logic type is characterized as presenting in thefifth to sixth decade of life and being associatedwith a palpable mass in 50 percent of cases.44

Virtually all known histologic types of inva-sive breast cancer have been identified in men.Invasive ductal carcinomas predominate,accounting for up to approximately 90 percentof cases. Also, special histologic types havebeen noted. Both invasive lobular carcinomaand lobular carcinoma in situ (LCIS) have beenreported in men but are much less commonthan in women. Sarcomas comprise a minorityof reported invasive breast cancers; there havebeen a variety of types noted (see Table 14–2).Metastatic cancer to the breast must beincluded in the differential diagnosis of breastmasses. Lung carcinoma has been reported tometastasize to the male breast.45

CLINICAL PRESENTATION AND EVALUATION

The clinical features of male breast cancer havebeen well described in the literature9,20,46–54 andrecently summarized.55 Signs and symptoms ofmale breast cancer are shown in Table 14–3.The mean age of patients presenting with thisdisease as noted above, is usually in the late 50sto mid-60s, with a range from the mid-20s tothe early 90s. The most common presentingcomplaints are related to a breast mass, usuallyoccurring in > 70 percent of cases, and axillaryadenopathy, occurring in 30 to 50 percent ofcases. For pure DCIS, a subareolar mass andnipple discharge were the two most common

242 BREAST CANCER

Table 14–2. HISTOLOGIC TYPES OF BREAST CANCER IN MEN

Noninvasive CarcinomaDuctal carcinoma in situLobular carcinoma in situPaget’s diseasePapillary carcinoma in situ

Invasive CarcinomaArgyrophilic neuroendocrine carcinomaColloid carcinomaInflammatory carcinomaIntracystic papillary carcinomaInvasive ductal carcinomaInvasive lobular carcinomaInvasive papillary carcinomaMedullary carcinomaMucinous carcinomaOncocytic carcinomaSecretory carcinoma

SarcomaPhyllodes tumorFibrosarcomaLeiomyosarcomaLymphosarcomaMyxoliposarcomaOsteosarcomaSpindle cell sarcoma

Adapted from Donegan WL, Redlich PN. Breast cancer in men.Surg Clin North Am 1996;76:343–63.

symptoms in a recent series, occurring in 58and 35 percent of patients, respectively.43 Invirtually all series, there is a report of signifi-cant delay in diagnosis of breast cancer in men.In early series, the mean duration of symptomswas > 14 months. In recent series, the meanduration is declining to a range of 3 to 6months.48,50,52,54 There is a history of trauma inmany series, ranging from 5 to 10 percent ofcases. Many series report the presence ofgynecomastia associated with this disease in 7to 23 percent of cases.9,20,50 The mass is cen-trally located in the majority of cases and hasan average diameter of 2.5 to 3.0 cm, with arange of 0.5 to 12 cm. Bilaterality of the diseaseis present in usually < 1 percent of cases,although in one series,50 7 percent of patientswere found to have bilateral disease. Clinicallysuspicious axillary adenopathy is often foundin these patients, ranging as high as 55 percent.The accuracy of the clinical exam is question-able, however, and pathologically provenmetastases are usually more frequent. In someseries, histologically proven axillary metastasesoccur as often as 70 percent of the time butmore frequently are in the 40 to 60 percentrange. The stage of disease at presentation issomewhat variable between reported series andmay not be entirely comparable from series toseries due to the large time spans involved andmodification of staging systems over time.Stratification of patients by TNM stage at thetime of presentation is presented in Table 14–4.

Evaluation of breast lesions includes the useof mammography, ultrasonography, fine needleaspiration cytology (FNAC), needle core biopsy,and open biopsy. Characteristics of male breastcancer on mammography include a mass eccen-tric to the nipple, spiculated margins, andmicrocalcifications (Figure 14–1).56–58 Malig-nancy must be differentiated from gynecomas-tia, which often presents as an area of increaseddensity positioned symmetrically in the retroare-ola region, but may obscure tumors.56 Secondaryradiologic signs of malignancy include architec-tural distortion, nipple and skin changes, andenlarged axillary nodes.56 The ultrasound fea-


Table 14–3. SIGNS AND SYMPTOMS OF MALE BREAST CANCER

FrequentBreast massAxillary adenopathyNipple retractionNipple dischargeRetraction of skinUlceration of nipple or skin

Less FrequentFixation to muscleBreast painInflammatory skin changesSkin discoloration

Table 14–4. STAGE OF DISEASE AT PRESENTATION

Stage Frequency (%)

0 0–17 I 10–40II 20–40III 15–40IV 10–15

Figure 14–1. Mammography of a male patient with breastcancer. This 83 year-old man presented with a 4-month his-tory of a right-breast mass. He had a 4.5 cm tumor, T2N1M0(stage IIB), treated by modified radical mastectomy and adju-vant tamoxifen. He is free of disease 3 years later. Shown arethe medial-lateral-oblique views of both breasts, the tumorbeing in the right breast on the left side of the figure.

tures of male breast cancer compared to otherbenign entities, including gynecomastia, lipoma,and fat necrosis have been reported.58 Malebreast cancer appears as a hypoechoic lesionwith irregular margins with architectural distor-tion of surrounding normal breast tissue andsubcutaneous fat. Ultrasound should be regardedas complimentary to mammography in the eval-uation of the male breast.

The first diagnostic step in the evaluation ofa male breast mass is often FNAC. Cytologicfeatures of male breast cancer are similar tothose seen in the female and allow this modalityto be a reliable means of assessment.59 Difficul-ties encountered using FNAC include epithelialhyperplasia associated with gynecomastia andthe differentiation between primary andmetastatic lesions of the breast.59 Combinedphysical examination and FNAC for the evalua-tion of palpable breast masses in males has beenstudied.60 This combination was found to bediagnostically accurate and resulted in a reduc-tion of patient charges compared to routine openbiopsy. In another series reviewing the diagnos-tic evaluation of over 700 male patients, the roleof palpation, mammography, cytology, andultrasound was evaluated. The combined palpa-tion and mammography demonstrated a veryhigh sensitivity for an accurate diagnosis.61

Accurate diagnosis by cytology requires anexperienced cytologist, the absence of whichmandates either needle core biopsy or ulti-mately open biopsy of lesions. Open biopsyshould be performed in all lesions where uncer-tainty exists regarding the diagnosis, both toconfirm the diagnosis and obtain tissue forestrogen and progesterone receptor measure-ments. Receptor status is important as a prog-nostic indicator for survival and as an indicatorfor response to hormonal manipulation.54,62

PROGNOSTIC FACTORS

Overall survival for men with breast cancer inlarge series ranges from 53 to 70 percent at 5years and 38 to 53 percent at 10 years.52,63,64

Observed survival of men is regularly inferior tothat of women. The Winchesters and colleaguesreported on 4,755 cases of male breast cancerobtained through the National Cancer Data Baseand compared them with 624,174 cases of breastcancer in women.7 The mean age of men, 64.7years, was older than that of women, which was60.9 years. Similar distributions of tumor gradeswere found. Men presented in more advancedstages than women, and 5-year survival was sig-nificantly lower. However, when adjusted for ageand comorbidity, survival was equivalent.

The survival of men with breast cancercompared to women with breast cancer hasbeen addressed in many series. An overallworse prognosis for men has been identified bya number of authors.51,54,65 Other authors sug-gest that the prognosis in male breast cancer isno worse than that for women with comparabledisease.7,47,50,66 Guinee and colleagues sug-gested that the prognosis is the same for maleand female patients when stratified on the basisof histologically positive nodes.49 There hasbeen a similar prognosis for male and femalepatients when analyzed by disease-specific sur-vival, tumor size, and axillary node involve-ment reported by other authors as well.50,52,66

There are a number of influences unrelated tobreast cancer itself that contribute to the unfa-vorable comparison of men and women. Amongthem are the older age at diagnosis and theirshorter life expectancy after 65 years of age dueto comorbid disease; men have higher rates ofdeath from heart disease, second cancers, andstroke.55 These confounding variables bias com-parisons of observed survival and disease-freesurvival in favor of women.7 More valid compar-isons require the use of survival adjusted for nat-ural mortality (adjusted survival) or of disease-specific survival (DSS).7 Adjusted five-andten-year survivals reported by Joshi and col-leagues67 were 76 and 42 percent, respectivelyand DSS of 74 percent at 5 years and 51 percentat 10 years were reported by Cutuli and col-leagues.52 Further detrimental to the survival ofmen is the high frequency of locally and region-

244 BREAST CANCER

ally advanced disease and of disseminated dis-ease, features in keeping with delay in diagnosisand not necessarily with inherently aggressivecancers.7,67 Skin involvement is often presentand involved axillary nodes are found in 45 to65 percent of men with axillary dissections.When factors of stage and comorbidity aretaken into account, however, investigators findlittle or no difference between the prognosis ofmales and females with breast cancer.44,50,52,66,67

The patient’s TNM stage at diagnosis isimportant prognostically for men, and outcomeby stage is largely not influenced by variationsin local treatment (Figure 14–2). The currentTNM staging system, derived from studyingbreast cancer in women, may not be entirelyappropriate for men. Male cancers average 2.0to 2.9 cm in diameter, but the diminutivebreasts of men allow even small tumors to read-ily reach underlying muscle and overlying skinor nipple.8,67 Pivot and colleagues found skin ormuscle involvement in 45 percent of 85 cases;skin involvement was directly correlated withtumor size.68 Forty-five percent of tumors < 2cm in diameter had produced nodal metastases,and all tumors > 5 cm in diameter had producednodal metastases. There was a similar prognosisfor T3 and T4 tumors. These authors proposed areduced T1, T2, and T4 classification for men.

Among traditional prognosticators, the pres-ence of axillary metastases and primary tumorsize are the most important features in undis-seminated cases.51,69,70,71 There is a direct corre-lation between the size of the primary tumor andthe involvement of axillary lymph nodes thatlinks these two clinical features.68,71 Guinee andassociates found a progressive drop in 5-yearsurvival from 94 percent for cases with tumors 0to 10 mm in diameter to 39 percent in cases withtumors > 51 mm in diameter.49 Tumor size isimportant prognostically, independent of axillarynode status. In node-negative cases, the relativerisks of death associated with T0-T1, T2, and T3-T4 cases were 1.0, 2.0, and 3.2, respectively.52

The presence of axillary metastases is thesingle-most important prognostic indicator of

survival for male breast cancer. Crichlowreported 5-year survivals of 79 and 28 percent for143 patients without and with pathologic axil-lary metastases, respectively.29 In a review of397 nondisseminated cases, the 5-year DSS of77 and 51 percent for cases with histologicallyuninvolved and involved nodes, respectively,were reported.52 As in women, the absolutenumber of involved nodes is inversely related tosurvival. In 335 collected cases, Guinee andcolleagues demonstrated that the 5-year sur-vival for those with negative nodes was 90 per-cent, for one to three positive nodes was 73 per-cent, and for four or more positive nodes 55percent.49 Others have found the same relation-ship53,54 (Figure 14–3). Unequal numbers ofinvolved nodes contribute to varied prognosesreported for node-positive cases (see Figure14–3). Skin and nipple involvement are identi-fied with adverse survival (Figure 14–4).67 Skinulceration becomes insignificant, however,when tumor size is taken into consideration,and fixation to skin and chest wall are notimportant prognosticially when size and nodalstatus are taken into account.49 There has beena statistically significant difference in survivalbased on histologic grade reported in one series,a worse survival associated with grade III ver-sus grade II disease.53


Figure 14–2. Survival of men with breast cancer stratified byTNM stage. Differences between the following stages reachedstatistical significance: 0 vs III, 0 vs IV, I vs II, I vs III, I vs IV, IIvs III, II vs IV, and III vs IV. The number of cases is shown inparentheses. Reprinted with permission from Donegan WL,Redlich PN, Lang PJ, Gall MT. Carcinoma of the breast inmales: a multi-institutional survey. Cancer 1998;83:498–509.

The influence of steroid receptors on theprognosis of men is controversial. Estrogenreceptor (ER) positivity, which is regularlymore frequent in men than in women, is a weakbut favorable prognostic sign for women. Inmen, ER positivity has been associated withboth increased and decreased survival.54,63,72

High tumor grade and aneuploidy are bothassociated with shortened survival.63,73 Pichand colleagues found median survival signifi-cantly less for grade III than for grade IItumors.74 Visfelt and Scheike graded 150 malebreast carcinomas according to the degree oftubule formation, mitoses, and atypia and

found 5-year survivals of 60, 40, and 5 percentin men with tumor grades of I, II, and III,respectively.75 For diploid tumors, Pich and col-leagues found a median survival of 77 monthsand only 38 months for aneuploid tumors73

Mutation of p53 cellular protein shortensmedian survival and disease-free survival.73,76

The mean S-phase fraction of male breast can-cers (7.2%) approximates that of females.77

Winchester and colleagues found high S-phasefraction (SPF) to be a significant indicator ofpoor disease-free survival.70 In a study of 27male breast cancers, Pich and associates foundthat strong staining for argyrophilic nuclearorganizer regions and for proliferating cellnuclear antigen were correlated with inferiorsurvival.74 The frequency of tumor markerswith and without established prognostic signif-icance are shown in Table 14–5.

TREATMENT OF LOCALIZED DISEASE

The treatment of male breast cancer localizedto the breast and axillary nodes is mastectomywith axillary lymph node dissection.55 In sev-eral recent series, modified radical mastectomywas the most common procedure performed,with from 34 to 76 percent of patients treated inthis fashion.8,20,50–52,54,65 In one multi-institu-tional survey, 82 percent of patients diagnosedsince 1986 were treated by modified radicalmastectomy.54 Of 242 patients treated in theDepartment of Veterans’ Affairs, 51 percentunderwent modified radical mastectomy.8

Other surgical procedures reported in thesepatients include radical mastectomy, simplemastectomy, and lumpectomies.

The use of radical mastectomy has decreasedmarkedly in recent decades. No significant dif-ference in outcome for patients who underwentradical mastectomy compared to modified radi-cal mastectomy was found in a number ofseries.20,78 The effect of the extent of mastec-tomy on the local regional recurrence rate isunclear, but a trend toward a lower local recur-rence rate was identified for patients undergo-

246 BREAST CANCER

Figure 14–3. Survival by men with breast cancer stratified bynumber of axillary nodes with metastases. Differencesbetween 0 vs 4+ positive nodes and 1–3 vs 4+ positive nodesreached statistical significance.The number of cases is shownin parentheses. Reprinted with permission from Donegan WL,Redlich PN, Lang PJ, Gall MT. Carcinoma of the breast inmales: a multi-institutional survey. Cancer 1998;83:498–509.

Figure 14–4. Direct involvement of the nipple in a man withbreast carcinoma. The nipple is fixed to the underlying tumor,endurated, and retracted. The site of skin biopsy is marked bya stitch.

ing mastectomy versus lumpectomy.52 Axillarydissection is considered part of the local–regional treatment of breast cancer. Cutuli andcolleagues reported a statistically significantdifference in the regional nodal recurrence rateof patients undergoing axillary dissection com-pared to those without dissection—1.2 percentof 320 patients with axillary dissection had aregional recurrence, compared to 13 percent of77 patients without axillary dissection.52

Postoperative radiation therapy is frequentlyused as adjuvant therapy for male breast cancer.Its use, however, varies widely.55 Some seriessuggest a decrease in the local–regional recur-rence rate with the use of postoperative radia-tion therapy,52,69 whereas no efficacy was notedin other series.20,78 Chest wall radiation offersno survival benefit.20,55,69 Overall, postopera-tive radiotherapy may reduce the local recur-rence rate and should be considered as part ofthe overall treatment plan for cases at high riskfor local or regional recurrence.52,79

Systemic adjuvant therapy, either chemo-therapy, hormonal therapy, or both, is used formale patients based on the experience in femalepatients. The modality most often used for post-operative adjuvant hormonal therapy is tamox-

ifen.46,50,80,81 Orchiectomy has been reported in afew series but is usually limited to only 3 per-cent of cases.9,20,50 Tamoxifen is generally welltolerated by men, but side effects have beenreported that, on occasion, may lead to termina-tion of treatment.80,81 Combination chemotherapyadministered in an adjuvant setting has beenused and reported by a number of authors.9,52,54,55

Treatment with cyclophosphamide, methotrex-ate, and 5-fluorouracil (CMF) has been reportedfrom the National Cancer Institute in a series of24 patients with stage II disease. The projected5-year survival rate was > 80 percent, represent-ing an improvement over survival rates reportedin other series.82 In another report, 5-fluo-rouracil, doxorubicin, and cyclophosphamide(FAC) or CMF was administered in the adjuvantsetting, with the projected 5-year survival > 85percent.83 Adverse effects of these chemother-apy regimens have been reported, limiting theability of patients to tolerate all the plannedtreatments. In a recent series, Donegan and col-leagues evaluated the effect of systemic adju-vant therapy, either chemotherapy, hormonaltherapy, or combinations of both, on survival.54

No improvement in overall survival was demon-strated; however, further analysis revealed


Table 14–5. TUMOR MARKERS IN MALE BREAST CANCERS

Marker Frequency (%) Prognosis Reference

ER+ 64–93 Controversial Rayson,94b Willsher,95 Joshi,67 Bruce,63 Cutuli52

PR+ 73–96 Rayson,94b Willsher,95 Joshi,67 Cutuli52

pS2 50 Bruce63

Cathepsin D 46 Bruce63

AR+ 95 Bruce63

HER2/neu+ 21–45 Unfavorable Rayson,94b Bines,96 Willsher,95 Joshi,67 Bruce63

p53+ 21–58 Unfavorable Rayson,94b Bines,96 Willsher,95 Dawson,97 Joshi67

Bruce,63 Anelli,76 Pich73

MIB-1+ 38–40 Unfavorable Rayson,94b Willsher95

bcl-2+ 93 Rayson94b

EGFR+ 20 Willsher95

Mean SPF 7.2 (mean) Highly unfavorable Jonasson,77 Winchester70

Grade III 33–73 Unfavorable Willsher,95 Jonasson,77 Bruce,63 McLachlan53

Aneuploidy 57 Unfavorable Jonasson,77 Pich73

Ki-ras 12 Dawson97

AgNOR+ Highly unfavorable Pich74

pcna Highly unfavorable Pich74

ER = estrogen receptor protein; PR = progesterone receptor protein; AR = androgen receptor; EGFR = epidermal growth factor receptor;pS2 = estrogen-dependent protein; HER2/neu = transmembrane oncoprotein; Grade III = histologic grade; AgNOR = argyrophilic nucleolar organizer regions; pcna = proliferating cell nuclear antigen.

improved DSS in a subset of patients. Diseasespecific survival was improved in patients whowere axillary lymph node positive with estrogenreceptor positive tumors, although the numberof patients with information regarding such sys-temic therapy was small.54

TREATMENT OF METASTATIC DISEASE

Many men with breast cancer present withmetastatic disease, ranging from 11 to 16 per-cent in several recent series.20,46,47,50,54 The pat-tern of spread in male patients is similar to thatseen in female patients including local regionalrecurrences and metastases to bone, lung, liver,skin, and other areas.9,20,50,51,53 The first-line pal-liative therapy used by most authors is hormonaltherapy, most often with tamoxifen.9,50,84,85 Reidand colleagues reported the use of hormonaltherapy in 73 percent of patients treated formetastatic disease.9 There has been a responserate of 25 to 58 percent to tamoxifen therapyreported in various series, the mean duration ofresponse being 9 to 12 months, with fewreported side effects. The response to tamox-ifen seems to correlate with estrogen receptorstatus in that patients with receptor negativetumors show no response.85,86 Historically,other hormonal treatments have beenemployed, including orchiectomy, adrenalec-tomy, and hypophysectomy. Response rates fororchiectomy are in the 30 to 60 percentrange.47,87–89 Response to secondary endocrineprocedures have been reported in > 50 percentof cases.87 Other hormonal manipulations,including androgen therapy and gonadotropin-releasing hormone agonist analogue therapyhave been reported.55 Systemic chemotherapymay be considered as a second-line palliativetreatment, with an overall response rate of 30 to40 percent.28,62,87,90–92 Combination chemother-apy, such as CMF or doxorubicin-containingregimens, have been reported in various seriesand may be useful in patients who have failedprior therapies.62,87,91,93

CHANGES OVER TIME

During the last century, clinical acumen, classi-fication systems, techniques of pathologic exam-ination and methods for reporting the results oftreating breast cancer have undergone extensivechanges. True comparability of reportingbetween periods is unlikely, particularly whendata are few and incomplete, as is the case formale breast cancer. Wainwright’s review of col-lected cases diagnosed before 1927 provides onebaseline against which to evaluate progress. Bymodern standards, it provides clear signs ofdelayed diagnosis and advanced disease.5 Theaverage symptomatic interval prior to diagnosiswas 2.4 years. The modal age of cases was 60 to64 years, with a mean of 52.6 years. Skin ulcer-ation was observed in 38 percent of cases, and68.9 percent of patients had involved axillarylymph nodes. Overall 5-year survival was only19 percent.20 While the age of men at the time ofpresentation has not declined convincingly,recent reports document shorter median sympto-matic intervals of 3 to 8 months, suggesting anearlier diagnosis.17,94a Average tumor sizes stub-bornly remain at 2.0 to 2.5 cm, but skin ulcera-tion is now reported in only 12 to 13 percent ofcases.34,49 The rate of dissemination at diagnosisdoes not exceed that for females (in most reports7 to 14 percent) and 2.9 to 5.6 percent of casesare now being diagnosed in situ—both favorablesigns even in the absence of a concerted publicscreening effort.49,55 The frequency of metastasisto axillary nodes generally remains high (45 to51 percent), but in some reports the rate is as lowas 33 percent.34,52,55

Favorable changes in managment of men havebeen less radical surgery and more frequent use ofsystemic adjuvant chemohormonal therapy. Menare now treated with modified radical mastec-tomy more often than with radical mastectomy,with no apparent detrimental effect and withimproved cosmetic outcome. Chemotherapy,tamoxifen, or both, targeted at micrometastases,are an increasingly frequent component of pri-mary treatment, with expectations that the bene-

248 BREAST CANCER

fits will parallel those for women. The indica-tions are that these changes are improving theoutlook for cure or extended survival of men.Current 5-year survivals of 47 to 51 percent over-all, and as high as 76 percent in undisseminatedcases, compare favorably with results reportedearly in the century. Some investigators have sug-gested that changes in therapy have resulted in lit-tle improvement in symptomatic interval, stage atdiagnosis, or survival for men with breast cancer.In a report from the Mayo Clinic comparing theresults from the period 1933 to 1958 with theperiod 1959 to 1983, fewer radical operations anda higher frequency of adjuvant systemic therapywere found in the more recent period, but not animprovement in median survival (5.5 years vs 6.3years, respectively) or in 5-year disease-free sur-vival of males with breast cancer (52 vs 47 per-cent).20 In a more recent 10-year span from 1986to 1995 compared with the period 1953 to 1985,investigators at the Medical College of Wiscon-sin noted a change of surgical treatment to mod-ified radical mastectomy and increased use ofsystemic therapy and found improvement in5-year observed survival of males treated inWisconsin hospitals, from 42 to 59 percent(p = .055).54 Stage for stage, the prognosis formen is comparable to, or somewhat inferior to,that for women. Truly accurate comparisons,however, are challenged by the need to controlfor multiple tumor-related and nontumor-relatedinfluences on outcome.

SUMMARY

Breast cancer in men is infrequent and occurs inan older population than does the disease inwomen. The disease in men is remarkably simi-lar to the disease in women. Men present withadvanced stages of disease more often thanwomen, but a trend toward earlier diagnosis isevident. There is an increased awareness of theimportance of family history and an associationwith BRCA2 mutations. Analysis of treatmenttrends demonstrates the use of less radicalsurgery and increased use of systemic adjuvant

therapy, especially in view of the high incidenceof estrogen-receptor positivity of breast cancerin men compared to women. Important prog-nostic variables include stage, tumor size, andaxillary node involvement. Also, survivaldecreases with increasing number of axillarynodes containing metastatic disease. Survivaltrends suggest improvement in the last one totwo decades, although the length of thisimprovement is limited by the older male popu-lation with this disease. Though the survival ofmen is reported to be similar to women com-pared stage for stage, comorbidities in men,such as heart disease, strokes, and secondmalignancies, contribute to an overall survivalthat may be inferior to that found in women.

REFERENCES

1. Landis SH, Murray T, Bolden S, Wingo PA. Can-cer Statistics, 1999. CA Cancer J Clin 1999;49:8–31.

2. Breasted JH. The Edwin Smith surgical papyrus.Vol. I. Chicago: University of Chicago Press;1930. p. 403–6.

3. Holleb AI, Freeman HP, Farrow JH. Cancer of themale breast. I and II. N Y State J Med 1968;68:544,656.

4. Meyskens FL Jr, Tormey DC, Neifeld JP. Malebreast cancer: a review. Cancer Treat Rev1976;3:83–93.

5. Wainwright JM. Carcinoma of the male breast.Arch Surg 1927;14:836–59.

6. Cutler SJ, Young JL Jr (eds). Third National Can-cer Survey: Incidence Data. National CancerInstitute Monograph 41. March 1975. DHEWPublication No (NIH) 75–787. US Departmentof Health, Education, and Welfare. PublicHealth Service. National Institutes of Health.National Cancer Institute, Bethesda, Maryland20014. US Government Printing Office. Wash-ington C.C. 20402. p. 106–107.

7. Winchester DJ, Bilimoria MM, Scott-Conarr CEH,Menck HR, Winchester DP. Male breast cancer:presentation, treatment, and survival (In press).

8. Schultz MZ, Coplin M, Radford D, Virgo KS,Johnson FE. Outcome of male breast cancer(MBC) in the Department of Veterans Affairs(DVA). Proc Annu Meet Am Soc Clin Oncol1996;15:A257.


9. Reid C, Pintilie M, Goncalves S, et al. A review of229 male breast cancer patients presenting tothe Princess Margaret Comprehensive CancerCentre. Proc Annu Meet Am Soc Clin Oncol1997;16:A476.

10. Ewertz M, Homnberg L, Kajjalaninen S, et al.Incidence of male breast cancer in Scandinavia1943–1982. Int J Cancer 1989;43:27–31.

11. Schottenfeld D, Lilienfeld AM, Diamond H. Someobervations on the epidemiology of breast can-cer among males. Am J Pub Health1963;52:890–7.

12. Amir H, Makwaya CK, Moshiro C, Kwesigabo G.Carcinoma of the male breast: a sexually trans-mitted disease? East Afr Med J 1996;73:187–90.

13. Hassan I, Mabogunje O. Cancer of the male breastin Zaria, Nigeria. East Afr Med J 1995;72:457–8.

14. Hsing AW, McLaughlin JK, Cocco P, et al. Riskfactors for male breast cancer (United States).Cancer Causes Control 1998;9:269–75.

15. D’Avanzo B, LaVecchia C. Risk factors for malebreast cancer. Br J Cancer 1995;71:1359–62.

16. Stenlund C, Floderus B. Occupational exposure tomagnetic fields in relation to male breast cancerand testicular cancer: a Swedish case-controlstudy. Cancer Causes Control 1997;8:184–91.

17. Donegan WL, Perez-Mesa CM. Carcinoma of themale breast. Arch Surg 1973;106;273–9.

18. Olsson H, Andersson H, Johansson O, et al. Popu-lation-based cohort investigations of the riskfor malignant tumors in first-degree relativesand wives of men with breast cancer. Cancer1993;71:1273–8.

19. Friedman LS, Gayther SA, Kurosaki T, et al. Muta-tion analysis of BRCA1 and BRCA2 in a malebreast cancer population. Am J Hum Genet1997;60:313–9.

20. Gough DB, Donohue JH, Evans MM, et al. A 50-year experience of male breast cancer: is out-come changing? Surg Oncol 1993:2:325–33.

21. Everson RB, Fraumeni JF Jr, Wilson RE. Familialmale breast cancer. Lancet 1976 Jan 3; 9–14.

22. Couch FJ, Farid LM, DeShano ML, et al. BRCA2germline mutations in male breast cancer casesand breast cancer families. Nat Genet 1996;13:123–5.

23. Haraldsson K, Loman N, Zhang QX, et al. BRCA2germ-line mutations are frequent in malebreast cancer patients without a family historyof the disease. Cancer Res 1998;58:1367–71.

24. Hiort O, Naber SP, Lehners A, et al. The role ofandrogen receptor gene mutations in male

breast carcinoma. J Clin Endocrinol Metab1996;81:3404–7.

25. Poujol N, Lobaccaro JM, Chiche L, et al. Func-tional and structural analysis of R607Q andR608K androgen receptor substitutions associ-ated with male breast cancer. Mol CellEndocrinol 1997;130:43–51.

26. Klinefelter HF Jr, Reifenstein EC Jr, Albright F.Syndrome characterized by gynecomastia,aspermatogenesis without A-leydigism andincreased excretion of follicle-stimulating hor-mone. J Clin Endocrinol 1942;2(11):615–27.

27. Hultborn R, Hanson C, Kopf I, et al. Prevalence ofKlinefelter’s syndrome in male breast cancerpatients. Anticancer Res 1997;17(6D):4293–7.

28. Volm MD, Gradishar WJ. How to diagnose andmanage male breast cancer. Contemp Oncol1994;4:17–28.

29. Crichlow RW. Breast cancer in males. Breast, Dis-eases of the Breast 1976;Oct/Dec:2(4):12–16.

30. Casagrande JT, Hanisch R, Pike MC, et al. A case-control study of male breast cancer. CancerRes 1988;48:1326–30.

31. Srensen HT, Friis S, Olsen JH, et al. Risk of breastcancer in men with liver cirrhosis. Am J Gas-troenterol 1998;93:231–3.

32. Symmers WC. Carcinoma of the breast in trans-sexual individuals after surgical and hormonalinterference with the primary and secondarysex characteristics. Br Med J 1968;2:83–5.

33. Pritchard TJ, Pankowsky DA, Crowe JP, et al.Breast cancer in a male-to-female transsexual.JAMA 1988;259:2278–80.

34. Crichlow RW, Kaplan EL, Keamey WH. Malemammary cancer: an analysis of 32 cases. AnnSurg 1972:175;489–94.

35. Mabuchi K, Bross DS, Kessler II. Risk factors formale breast cancer. J Natl Cancer Inst 1985;74:371–5.

36a. Karamanakos P, Apostolopoulos V, Fafouliotis S,et al. Synchronous bilateral primary malebreast carcinoma with hyperprolactinemia.Acta Oncol 1996:35:757–9.

36b. Volm MD, Talamonti MS, Thangavelu M, GradisharWK. Pituitary adenoma and bilateral malebreast cancer: an unusual association. J SurgOncol 1997:64:74–8.

37. Thompson DK, Li FP, Cassady Jr. Breast cancer ina man 30 years after radiation for metastaticosteogenic sarcoma. Cancer 1979;44:2362–5.

38. Lowell DM, Marineau RG, Luria SB. Carcinomaof the male breast following radiation. Cancer1968;22:581–6.

250 BREAST CANCER

39. Sanchez AG, Villanueva AG, Redondo C. Lobularcarcinoma of the breast in a patient with Kline-felter’s syndrome. Cancer 1986;57:1181–3.

40. San Miguel P, Sancho M, Enriquez JL, et al. Lob-ular carcinoma of the male breast associatedwith the use of cimetidine. Virchows Arch1997;430:261–3.

44. Daltry IR, Eeles RA, Kissin MW. Bilateral pro-phylactic mastectomy: not just a woman’sproblem! Breast 1998;7:236–7.

42. Cutuli B, Dilhuydy JM, DeLafontan B, et al. Duc-tal carcinoma of the male breast: analysis of 31cases. Eur J Cancer 1997;33:35–8.

43. Hittmair AP, Lininger RA, Tavassoli FA. Ductalcarcinoma in situ (DCIS) in the male breast.Cancer 1998;83:2139–49.

44. Desai DC, Brennan EJ Jr, Carp NZ. Paget’s diseaseof the male breast. Am Surg 1996;62:1068–72.

45. Muttarak M, Nimmonrat A, Chaiwun B. Metasta-tic carcinoma of the male and female breast.Australas Radiol 1998;42:16–9.

46. Ribeiro, G. Male breast carcinoma—A review of301 cases from the Christies Hospital & HoltRadium Institute, Manchester. Br J Cancer1985;51:115–9.

47. van Geel AN, van Slooten EA, Mavrunac M, HartAAM. A retrospective study of male breastcancer in Holland. Br J Surg 1985;72:724–7.

48. Hultborn R, Friberg S, Hultborn KA. Male breastcarcinoma. I. A study of the total materialreported to the Swedish Cancer Registry1958–1967 with respect to clinical and histo-pathologic parameters. Acta Oncol 1987;26:241–56.

49. Guinee VF, Olsson H, Moller T, et al. The progno-sis of breast cancer in males. Cancer 1993;71:154–61.

50. Borgen PI, Wong GY, Vlamis V, et al. Current man-agement of male breast cancer: a review of 104cases. Ann Surg 1992;215:451–9.

51. Salvadori B, Saccozzi R, Manzari A, et al. Progno-sis of breast cancer in males: an analysis of 170cases. Eur J Cancer 1994;30A:930–5.

52. Cutuli B, Lacroze M, Dilhuydy JM, et al. Malebreast cancer: results of the treatments andprognostic factors in 397 cases. Eur J Cancer1995;31A:1960–4.

53. McLachlan SA, Erlichman C, Liu FF, et al. Malebreast cancer: an 11 year review of 66 patients.Breast Cancer Res Treat 1996;40:225–30.

54. Donegan WL, Redlich PN, Lang PJ, Gall MT. Car-cinoma of the breast in males: a multi-institu-tional survey. Cancer 1998;83:498–509.

55. Donegan WL, Redlich PN. Breast cancer in men.Surg Clin North Am 1996;76:343–63.

56. Dershaw DD, Borgen PI, Deutch BM, Liberman L.Mammogrphic findings in men with breastcancer. AJR 1993;160:267–70.

57. Cooper RA, Gunter BA, Ramanurthy L. Mam-mography in men. Radiology 1994;191:651–6.

58. Stewart RAL, Howlett DC, Hern FJ. Pictorialreview: the imaging reatures of male breastdisease. Clin Radiol 1997;52:739–44.

59. Sneige N, Holder PD, Katz RL, et al. Fine-needleaspiration cytology of the male breast in a can-cer center. Diagn Cytopathol 1993;9:691–7.

60. Vetto J, Schimdt W, Pommier R, et al. Accurate andcost effective evaluation of breast masses inmales. Am J Surg 1998;175:383–7.

61. Ambrogetti D, Ciatto S, Catarzi S, Muraca MG.The combined diagnosis of male breastlesions: a review of a series of 748 consecutivecases. Radiol Med 1996;91:356–9.

62. Bezwoda WR, Hesdorffer C, Dansey R, et al.Breast cancer in men: clinical features, hor-mone receptor status and response to therapy.Cancer 1987;60:1337–40.

63. Bruce DM, Heys SD, Payne S, et al. Male breastcancer: clinico-pathological features, immuno-cytochemical characteristics and prognosis.Eur J Surg Oncol 1996;22:42–6.

64. Caton J, Rearden T, Ellis R. Male breast cancer: theDepartment of Defense (DOD) experience. ProcAnnu Meet Am Soc Clin Oncol 1995;14:A140.

65. Ciatto S, Iossa A, Bonardi R, Pacini P. Male breastcarcinoma: review of multicenter series of 150cases. Tumori 1990;76:555–8.

66. Borgen PI, Senie RT, McKinnon WM, Rosen, PP.Carcinoma of the male breast: analysis of prog-nosis compared with matched female patients.Ann Surg Oncol 1997;4:385–8.

67. Joshi MG, Lee AK, Loda M, et al. Male breast car-cinoma: an evaluation of prognostic factorscontributing to a poorer outcome. Cancer1996;77:490–8.

68. Pivot X, Llombart-Cussac A, Rhor-Albarddo A, etal. Clinical staging for male breast cancer: anadaptation of the international classification(TNM). Proc Annu Meet Am Soc Clin Oncol1997;16:A655.

69. Lartigau E, El-Jabbour JN, Dubray B, Dische S.Male breast carcinoma: a single centre report ofclinical parameters. Clin Oncol 1994;6:162–6.

70. Winchester DJ, Goldschmidt RA, Khan SJ, et al.Flow cytometric and molecular prognosticmarkers in male breast carcinoma patients[abstract]. Presented at the 46th Annual Cancer


Symposium of the Society of Surgical Oncol-ogy; Los Angeles; 1993; March 18–21.

71. Yap HY, Tashima CK, Blumenschein GR, EcklesNE. Male breast cancer: a natural history study.Cancer 1979;44:748–54.

72. Winchester DJ. Male breast cancer. Semin SurgOncol 1996;12:364–9.

73. Pich A, Margaria E, Chiusa L, et al. DNA ploidyand p53 expression correlate with survival andcell proliferative activity in male breast carci-noma. Hum Pathol 1996;27:676–82.

74. Pich A, Margaria E, Chiusa L. Proliferative activityis a significant prognostic factor in male breastcarcinoma. Am J Pathol 1994;145:481–9.

75. Visfelt J, Scheike O. Male breast cancer. I. Histo-logic typing and grading of 187 Danish cases.Cancer 1973;32:985–90.

76. Anelli A, Anelli TF, Youngson B, et al. Mutationsof the p53 gene in male breast cancer. Cancer1995;75:2233–8.

77. Jonasson JG, Agnarsson BA, Thorlacius S, et al.Male breast cancer in Iceland. Int J Cancer1996;65:446–9.

78. Hultborn R, Friberg S, Hultborn DA, et al. Malebreast carcinoma. II. A study of the total mate-rial reported to the Swedish Cancer Registry1958–1967 with respect to treatment, prognos-tic factors and survival. Acta Oncol 1987;26:327–41.

79. Schuchardt U, Seegenschmiedt MH, Kirschner MJ,et al. Adjuvant radiotherapy for breast carci-noma in men: a 20-year clinical experience.Am J Clin Oncol 1996;19:330–6.

80. Anelli TFM, Anelli A, Tran KN et al. Tamoxifenadministration is associated with a high rate oftreatment-limiting symptoms in male breastcancer patients. Cancer 1994;74:74–7.

81. Ribeiro G, Swindell R. Adjuvant tamoxifen formale breast cancer. Br J Cancer 1992;65:252–4.

82. Bagley CS, Wesley MN, Young RC, Lippman ME.Adjuvant chemotherapy in males with cancerof the breast. Am J Clin Oncol 1987;10:55–60.

83. Patel, HZ, Buzdar AU, Hortobagyi GN. Role ofadjuvant chemotherapy in male breast cancer.Cancer 1989;64:1583–5.

84. Becher R, Höffken K, Pape H, Schmidt CG.Tamoxifen treatment before orchiectomy inadvanced breast cancer in men. N Engl J Med1981;305:169–70.

85. Ribeiro GG. Tamoxifen in the treatment of malebreast carcinoma. Clin Radiol 1983;34:625–8.

86. Patterson JS, Batersby LA, Back BK. Use oftamoxifen in advanced male breast cancer.Cancer Treat Rep 1980;64:801–4.

87. Jaiyesimi IA, Buzdar AU, Sahin AA, et al. Carci-noma of the male breast. Ann Intern Med1992;117;771–7.

88. Kantarjian H, Yap H-Y, Hortobagyi G, et al. Hor-monal therapy for metastatic male breast can-cer. Arch Intern Med 1983;143:237–40.

89. Patel JK, Nemoto T, Dao TL. Metastatic breastcancer in males: assessment of endocrine ther-apy. Cancer 1984;53:1583–5.

90. Kraybill WG, Kaufman R, Kinne D. Treatment ofadvanced male breast cancer. Cancer 1981;47:2185–9.

91. Sandler B, Carman C, Perry RR. Cancer of themale breast. Am Surg 1994;60:816–20.

92. Yap HY, Tashima CK, Blumenschein GR, et al.Chemotherapy for advanced male breast can-cer. JAMA 1980;243:1739–41.

93. Lopez M, DiLauro L, Papaido P, et al. Chemother-apy in metastatic male breast cancer. Oncology1985;452:205–9.

94a. Stierer M, Rosen H, Weitensfelder W, et al. Malebreast cancer: Austrian experience. World JSurg 1995;19:687–92.

94b. Rayson D, Erlichman C, Wold LE, et al. Molecu-lar markers in male breast cancer. Proc AnnuMeet Am Soc Clin Oncol 1997;16:A477.

95. Willsher PC, Leach IH, Ellis IO, et al. Male breastcancer: pathological and immunohistochemi-cal features. Anticancer Res 1997;17:2335–8.

96. Bines J, Goss B, Hussong J, et al. c-erbB2 and p53overexpression as predictors of survival inpatients with male breast cancer. Proc AnnuMeet Am Soc Clin Oncol 1997;16:A558.

97. Dawson PJ, Schroer KR, Wolman SR. Ras and p53genes in male breast cancer. Mod Pathol 1996;9:367–70.

252 BREAST CANCER

253

Ninety percent of women will live to the cli-macteric age, compared to only 30 percent 200years ago. Attrition and aging of ovarian folli-cles results in termination of the maturation ofgranulosa cells, which are responsible for estro-gen production. Sources of estrogen in the pre-menopausal woman are several fold, includingdirect production of estradiol by the ovaries aswell as the extraglandular aromatization in adi-pose cells of androstenedione created in theadrenal glands and ovary. The hallmark ofmenopause is a drop in ovarian production ofestriol and testosterone. Peripheral aromatiza-tion of other steroids not produced by theovaries is an additional source of estrogen in allwomen. However, this source is not sufficientin most women to prevent the symptoms char-acteristic of estrogen deprivation.

Given the current population, 30 millionwomen in the United States will spend approx-imately 40 percent of their lifetime in the post-menopausal period. These women have a life-time risk of one in eight of developing breastcancer. Thus, a considerable number of Ameri-can women are likely to have a history of breastcancer treatment and at the same time be poten-tial candidates for hormone-replacement ther-apy. In the last decade, indications forchemotherapy as adjuvant treatment to surgeryhave widened and now encompass many morepremenopausal women.1 Adjuvant therapy forbreast cancer includes the use of alkylating

agents and other drugs that cause amenorrheain 84 percent of women aged 35 to 44 years.Other studies indicate that this treatment causespermanent ovarian failure in 86 percent ofwomen > 40 years of age.2 As a result, a largernumber of women will potentially be renderedmenopausal in the fourth, and fifth decades oftheir lives, which has serious consequences interms of the risk of cardiovascular disease andosteoporosis.

The major concern of many physicians inprescribing estrogen-replacement therapy(ERT) for breast cancer survivors is the theorythat metastatic quiescent tumor foci might beactivated and the “fire” of breast cancer ignitedby the “fuel” estrogen. Other fears are thatestrogen might cause a second primary in thealready environmentally/genetically primedcontralateral breast or might change breast den-sity and mask new mammographic findings.These concerns are in part based upon epi-demiologic studies demonstrating a relation-ship between duration of postmenopausal estro-gen replacement and breast cancer.3,4 Also,surgical oophorectomy is beneficial in a subsetof premenopausal breast cancer patients, andestrogen withdrawal has also been observed topromote regression of metastatic breast cancerlesions.5 Despite very limited clinical data tosupport these concerns, it remains standardpractice to prohibit breast cancer survivorsfrom receiving ERT.

15

Estrogen Replacement Therapy for Breast Cancer SurvivorsWENDY R. BREWSTER, MDPHILIP J. D ISAIA, MD

The well-substantiated benefits of estrogenreplacement therapy must be balanced againsttheoretical concerns. Arguments in support ofthe safety of ERT are based on several naturalexperiments and observations, discussed indetail below.

BENEFITS OF ESTROGEN REPLACEMENT THERAPY

Over the last two decades, overwhelming evi-dence has been accrued demonstrating thatpostmenopausal estrogen replacement protectsagainst ischemic heart disease, osteoporosis,deterioration in cognitive function, colorectalcancer, and provides relief from vasomotorsymptoms and urogenital atrophy. Multimodal-ity screening has resulted in an increase in theincidence of breast cancer diagnoses; thisincrease, however, reflects more frequent detec-tion of early-stage breast cancer. Becausebreast cancer survival is inextricably linked toearly diagnosis, there are now more breast can-cer survivors than ever. Morbidity and mortal-ity associated with estrogen deprivation presentserious health concerns. The risk/benefit ratioof estrogen replacement therapy (ERT) is anappropriate consideration for all patients.

Coronary Artery Disease

Cardiovascular disease is the leading cause ofmortality among women in the United States.The number of deaths from diseases of the cir-culatory system in women in the United Statesis greater than the number who die from can-cers of the breast, reproductive tract, and mater-nal morbidity combined. It is only during thereproductive years that more women die frommalignancy than from cardiovascular disease.This is reversed past 60 years of age.

The endocrine influences of factors thoughtto be contributors to the risk of cardiovasculardisease have been studied extensively. The liter-ature is vast and has been well summarized inseveral recent reviews. Unopposed estrogen

raises the serum level of high-density lipoproteincholesterol, especially the HDL2 subfraction,and lowers the serum level of low-densitylipoprotein cholesterol.6 Other less well-studiedfactors that may influence cardiovascular healthduring treatment with estrogen, with or withoutprogestin, include beneficial effects on the circu-lation, blood pressure, coagulation, and fibrinol-ysis.7,8 Estrogen also has vasodilating propertiesmediated by the generation of prostacyclin inthe cell membrane.

Many epidemiological studies have foundthat postmenopausal women who use estrogenare at a much lower risk for coronary diseasethan are nonusers. Observational studies sug-gest a 50 percent reduction in the risk of coro-nary heart disease among healthy post-menopausal women taking oral estrogen.9

In 1981, Henderson and colleaguesrecruited over 8,000 women from a retirementcommunity in Laguna Hills, California calledLeisure World. This is a stable community andvery few individuals were lost to follow-up. Ofthis cohort, 57 percent reported estrogen use,14 percent were current users at the time of thequestionnaire, and 43 percent reported previoususe. The incidence of mortality from acutemyocardial infarction was statistically loweramong current users and those who had usedestrogen in the past compared to nonusers. Therelative risk was 0.59, with the 95 percent con-fidence interval (CI) of 0.42 to 0.82.

Hunt10 reported on a cohort of 4,544 womenwho had taken hormone replacement therapycontinuously for at least 1 year at the time ofrecruitment. When compared with the generalfemale population, mortality rates for ischemicheart disease among the cohort were signifi-cantly lower, with a relative risk of 0.41 and a95 percent CI of 0.2 to 0.61. Bush11 evaluated acohort of 2,270 women, 593 of whom wereestrogen users. The age-adjusted relative risk ofdeath from cardiovascular disease was 0.34,with the 95 percent CI of 0.12-0.81.

Stampfer12 evaluated postmenopausalestrogen therapy and cardiovascular disease in

254 BREAST CANCER

the Nurses’ Health Study, with a 10-year fol-low-up. Women currently using post-menopausal hormone therapy accounted for21.8 percent of the total follow-up time of337,854 person-years. There was a reduction inthe age-adjusted relative risk of fatal cardio-vascular disease among current hormoneusers. In the same study, the age-adjusted riskof major coronary artery disease among cur-rent estrogen users was about half that ofwomen who had never used estrogen, with arelative risk of 0.51 p < .0001. For formerusers, the age-adjusted relative risk (RR) was0.91. When this was adjusted for other risk fac-tors, the relative risk was 0.83. The relative riskof fatal cardiovascular disease was decreasedin both current and former users.

The above studies were all based on post-menopausal estrogen use only. Given the factthat current medical recommendations call forthe addition of a progestin to estrogen therapyin nonhysterectomized women, there is thevalid concern that progestin therapy may negatethe benefits gained by estrogen (Table 15–1).

The investigators in the PostmenopausalEstrogen/Progestin Interventions (PEPI) trialexamined this issue.13 They found, as had beenconfirmed in numerous previous studies, thatunopposed estrogen decreased the risk factorsfor cardiovascular disease. However, estrogengiven with medroxyprogesterone acetate ormicronized progesterone hormone-replace-ment therapy (HRT) was associated with lowerfibrinogen levels and improved lipoproteinprofiles. No adverse effects on the rate of car-

diovascular incidents were observed for HRTover ERT.

Grodstein14 evaluated the effect of com-bined estrogen and progestin use and the risk ofcardiovascular disease in the Nurses’ HealthStudy. Among the 59,337 women enrolled,there were 770 casualties of myocardial infarc-tion or deaths from coronary artery disease.There was a marked decrease in the risk ofmajor coronary artery disease among womenwho took estrogen with progestin compared tothat for women who did not use hormones. Themultivariate-adjusted relative risk was 0.39,with the 95 percent CI of 0.19 to 0.78.

Osteoporosis

Postmenopausal women are at risk for loss ofcancellous bone in the vertebrae and other longbones, which places them at increased risk forfracture. Bone mineral density decreasesrapidly within 5 years of menopause due toestrogen deficiency. This ultimately results inmicroarchitectural deterioration and a progres-sive increased fracture risk. Postmenopausaluntreated women may lose 35 percent of theircortical bone and up to 50 percent of their tra-becular bone. It is estimated that 1.2 millionmajor fractures per year in the United States inwomen are related to osteoporosis. Fifteen per-cent of postmenopausal women will sufferwrist fractures, and an even larger number willincur spinal compression fractures. Compres-sion fractures of the vertebral bones may resultin loss of stature, pulmonary restriction, and

Estrogen Replacement Therapy for Breast Cancer Survivors 255

Table 15–1. EPIDEMIOLOGIC STUDIES OF THE CARDIOVASCULAR BENEFITS OF POSTMENOPAUSAL ESTROGEN AND PROGESTERONE USE

Study Design Number Results

Falkeborn et al.41 Prospective 227 MI cases RR = 0.74 ever estrogen only23,174 women RR = 0.50 ever combined therapy

Psaty et al42 Case-control 502 MI cases RR = 0.69 estrogen alone1,193 controls RR = 0.68 current combined therapy

Grodstein et al14 Prospective 770 MI cases RR = 0.60 current estrogen alone59,337 women RR = 0.39 current combined therapy

MI= myocardial infarction; RR = relative risk.

decreased ambulation. An estimated 40 percentof the women who will live to the age of 80years will develop spinal fractures and 33 per-cent will experience a hip fracture.

Of concern is the morbidity and mortalityassociated with hip fractures in older women.Within this group, 12 to 20 percent will diewithin 6 months of the fracture, and half of thesurvivors require long-term nursing care.Osteoporotic fractures in the United Statesresulted in health care costs of $7 billion in1986. This is estimated to increase to as muchas $62 billion by the year 2020.

Alzheimer’s Disease

As the population ages, Alzheimer’s disease hasemerged as a major health problem. After theage of 65 years, the prevalence of dementia andAlzheimer’s disease doubles every 5 years; 30to 50 percent of women older than 83 yearsmay suffer from dementia of some sort.

Laboratory studies suggest that estrogenmay affect Alzheimer’s disease through severalmechanisms. Estrogen has been shown toimprove regional cerebral blood flow and toincrease glucose utilization. It can also stimu-late neurite growth and synapse formation invitro. Under some circumstances, estrogen maymodify neural sensitivity to neurotrophin andplay a role in the reparative neuronal responseto injury. One key histologic feature ofAlzheimer’s disease is the deposition of beta-amyloid protein in cores of neuritic plaques.Estrogen may promote the breakdown of theamyloid precursor protein to fragments lesslikely to accumulate as beta amyloid. Acetyl-choline is a key neurotransmitter in learningand memory. Estrogen affects several neuro-transmitter systems, including the cholinergicsystem. Finally, estrogen may modify inflam-matory responses postulated to participate inneuritic plaque formation.15

Tang and colleagues examined the effect ofa history of estrogen use on the development ofAlzheimer’s disease in 1,200 women.16 These

subjects were initially free of Alzheimer’s dis-ease, Parkinson’s disease, and stroke and werepart of a longitudinal study of aging and healthin a New York community. Overall, 158 (12.5percent) reported taking estrogen after theonset of menopause. The age of onset ofAlzheimer’s disease was significantly later inwomen who had taken estrogen than in thosewho did not, 78 years versus 73 years. Evenafter adjustment for differences in educationand ethnic origin, the relative risk ofAlzheimer’s disease was significantly reducedin estrogen users over nonusers: 0.4, with a 95percent CI of 0.22 to 0.85.

Even among postmenopausal women whoare not demented, ERT may help maintain cog-nitive function.17 Estrogen appears to have aspecific effect on verbal memory skills inhealthy postmenopausal women.18,19

The emotional, physical, social, and finan-cial costs of Alzeimer’s disease to patients,families, caregivers, and society are tremen-dous. The estimated total cost of the disease in1991 was estimated to be $173,932 per case.The estimated prevalence cost for both men andwomen for that year was $67.3 billion.20 Theeconomic cost of care alone is greater than thecost of care for heart disease and cancer com-bined. If the use of estrogen could delay theonset of Alzheimer’s disease by several years,there would be a substantial saving in bothemotional and financial costs.

Colorectal Cancer

There have been > 20 retrospective studies ofthe risk of colon cancer and ERT, with morethan 70 percent of these reports illustrating astatistically significant reduction in incidencewith users versus nonusers. One proposedmechanism affecting this protection is thatestrogen reduces the concentration of bileacids, and may limit carcinogenic action to thecolon mucosa. It has been demonstrated thatbile acid concentrations are higher in coloncancer cases than in control subjects, and it is

256 BREAST CANCER

known that estrogen decreases bile acid synthe-sis and secretion.21 Estrogen receptors are pre-sent in both normal and cancerous colonmucosal cells, and there is laboratory evidenceto suggest that estrogen may inhibit the growthof colon cancer cells.22

Calle and colleagues23 investigated the rela-tionship between postmenopausal estrogen useand fatal colorectal cancer in a large prospectivestudy of adults in the United States. Eight hun-dred and seventy-nine colon cancer case patientswere compared to 421,476 noncase subjects.Ever use of ERT was associated with a signifi-cantly decreased risk of fatal colon cancer (RR =0.71; 95% CI = 0.61 to 0.83). Reduction in riskwas strongest among current users (RR = 0.55;95% CI = 0.40 to 0.76) compared to formerestrogen users. There was a significant trend ofdecreasing risk with increasing years of estrogenuse among all users (p = .0001). Those womenwho used estrogen for ² 1 year had a RR = 0.81,whereas users of ³ 11 years had a RR of 0.54(95 percent CI = 0.39 to 0.76). These associa-tions were not altered in multivariate analysescontrolling for age, race, parental history ofcolon cancer, body mass index, exercise, parity,type of menopause, age of menopause, oral con-traceptive pill use, aspirin use, and smoking.

Vasomotor Instability

The menopausal state most commonly producesvasomotor instability and genital organ atrophy.Vasomotor symptoms affect 70 percent of post-menopausal women but only about 30 percentseek medical assistance. For 25 percent ofmenopausal women, these symptoms may per-sist for > 5 years and may be lifelong in others.Vasomotor instability is more commonly termed“hot flushes” or “hot flashes.” The frequency,severity, or diurnal variation with which hotflushes occur can result in significant disrup-tions of sleep and daytime function. Menopausalsymptoms are the most common side effectassociated with the use of adjuvant chemother-apy for breast cancer, with approximately two-

thirds of women experiencing symptoms classi-fied as moderate to severe.24 This effect may becompounded by tamoxifen therapy, which alsoleads to vasomotor instability.

Urogenital Atrophy

Because the vagina and urethra share a com-mon embryologic origin, it is believed thatestrogen deficiency causes atrophy of bothstructures. Atrophy of the vaginal epitheliummay cause vaginal itching, dryness, and dys-pareunia, with resulting inflammation. Oneeffect of estrogen deficiency is to causechanges in the vaginal pH, which predisposewomen to urinary tract infections that causeurgency, incontinence, frequency, nocturia, anddysuria. The loss of estrogen on periurethral tis-sues will contribute to pelvic laxity and stressincontinence. Recurrent urinary tract infectionscan be prevented with systemic estrogen ther-apy, and low-dose topical estrogen is effectivein managing atrophic vaginitis. Estrogen pro-vides relief of these symptoms and may protectagainst recurrent urinary tract infections.

EXPOSURE TO EXOGENOUS ORENDOGENOUS ESTROGEN DURINGBREAST CANCER DEVELOPMENT

The decision whether or not to take hormonereplacement remains difficult for the post-menopausal woman because of conflicting risksand benefits and is even more difficult for thebreast cancer survivor for whom there is evenless data. One can therefore analyze situationsin which women are inadvertently exposed toexogenous or endogenous estrogen at a timewhen they may have been harboring subclinicalbreast cancer. Does such exposure adverselyaffect survival outcome for these patients?

Such situations include those in which thediagnosis of breast cancer is made in post-menopausal women receiving ERT at the time ofdiagnosis or in whom the diagnosis is made inpregnancy or during lactation, or in those women


with a history of oral contraceptive pill usearound the time of diagnosis of breast cancer.

Breast Cancer in Women on Estrogen Replacement Therapy

Bergkvist and co-workers25 compared 261women who developed breast cancer while onERT to 6,617 breast cancer patients who had norecorded treatment with estrogen. The relativesurvival rate over an 8-year period was higher inthe breast cancer patients who had previouslyreceived ERT. This corresponded to a 32 percentreduction in excess mortality. Gambrell,26 in aprospective study, also evaluated the effect onsurvival in breast cancer patients diagnosed whileon ERT. Mortality was 22 percent among thosediagnosed with breast cancer while on ERT com-pared to 46 percent among those who had neverreceived hormone replacement. Henderson andcolleagues27 observed a 19 percent reduction inbreast cancer mortality among 4,988 previousERT users, compared to 3,865 nonusers whosubsequently developed this disease.

Breast Cancer Associated with Pregnancy

Pregnancy coincident with, or subsequent to, thedetection of breast cancer provides another excel-lent opportunity to evaluate the outcome of breastcancer patients inadvertently exposed to high lev-els of estrogen at times when they were harboringoccult disease. During pregnancy, the serum lev-els of estriol increase 50-fold. Only 0.5 to 4 per-cent of all breast cancers are diagnosed duringpregnancy. Because the average breast cancerremains occult in the breast some 5 to 8 yearsprior to diagnosis, some authors include in thiscategory women in whom a diagnosis of breastcancer has been made within 12 months of deliv-ery. The outcome in women with subclinicalbreast cancers exposed to elevated levels of pro-gesterone and estrogen under these circumstancescould provide insight into the influence of thesehormones on the malignant disease process.

The physiologic changes and engorgementthat occur in the breast during pregnancy oftenhinder early detection of breast cancer. Thisresults in a diagnosis at more advanced stagesin pregnant and lactating women. Comparisonsto nonpregnant women matched for similar agestage of breast cancer and reproductive capac-ity do not suggest a worse prognosis for thepregnant patients with breast cancer.28,29 vonSchoultz30 performed a comparison of womendiagnosed with breast cancer 5 years beforepregnancy to women without a pregnancy dur-ing the same time period. There was no survivaldisadvantage to the women who were pregnant5 years prior to the diagnosis of breast cancer.These and other studies have discouraged thepractice of prohibiting breast cancer survivorsfrom becoming pregnant on clinical grounds.Subsequent pregnancies do not negativelyaffect survival outcomes.

Anderson and colleagues31 reported theirexperience at the Memorial Sloan KetteringCancer Center with breast cancer in women< 30 years of age. Two hundred and twenty-seven cases were identified, of whom 22 hadpregnancy-associated breast cancer. Theauthors confirmed that pregnancy-associatedbreast cancers were usually larger and presentin more advanced stages at the time of diagno-sis, compared to a similar group who were notpregnant. However, the survival probability forwomen with early stage disease was indepen-dent of pregnancy status.

The experience of women who have com-pleted term pregnancies after treatment ofantecedent breast cancer is another situationthat deserves analysis. There are inherent biasesassociated with evaluation of this particulargroup of subjects. This cohort is representativeof the young women who did well after primarybreast cancer therapy; since pregnancy data isnot uniformly coded in cancer registry data-bases, the true denominator of postbreast can-cer pregnancies is unknown. Clark32 reported a71 percent 5-year survival in a series of 136women with pregnancies after breast cancer

258 BREAST CANCER

(stages I to III). Equivalent survival outcomeswere reported by von Schoultz30 for breast can-cer patients with no subsequent pregnancy,compared to those who became pregnant within5 years of their diagnosis.

Breast Cancer in Oral Contraceptive Pill Users

Given the long natural history of this neoplasm,it is certain that a large number of patients sub-sequently diagnosed with breast cancer haveused oral contraceptive pills (OCP) during thegenesis and progression of their malignant dis-ease process; they are another group thatdeserves examination.

Rosner33 evaluated 347 women < 50 diag-nosed with breast cancer, of whom 112 wereOCP users. The distribution of tumor size,estrogen-receptor status, and family and repro-ductive history was the same between the twocohorts. There was no difference in disease-freesurvival or survival between the two groups.Women who used OCP within a year of diag-nosis of their breast cancer had a similar sur-vival to those who had discontinued use > 1year before. There was no difference in survivalamong those who used OCP ³ 10 years prior totheir diagnosis of breast cancer.

Schonborn and colleagues34 evaluated theinfluence of a positive history of OCP use onsurvival. Four hundred and seventy-one breastcancer patients were investigated. Two hundredand ninety-seven patients (63 percent) had usedOCP during any period of their life, and 92 (20percent) still used them at the time of diagnosis.Sixty months after diagnosis, the OCP usershad a significantly increased overall survival (p = .037). Survival rates amounted to 79.5 per-cent and 70.3 percent for OCP users andnonusers, respectively.

Sauerbrei35 investigated the relationshipbetween OCP use and standard prognostic fac-tors, and the effect of OCP use on disease-freesurvival and overall survival, in 422 pre-menopausal node-positive patients from two tri-

als of the German Breast Cancer Study Group.One hundred and thirty-seven OCP users (32.5percent) were younger than those who did notuse OCP (mean age 41.5 years versus 45 years).Noteworthy was the fact that the percentage ofpatients with smaller tumors was higher in thegroup of OCP users. No significant effect ofOCP use on either disease-free or overall sur-vival could be demonstrated in univariate andmultivariate analyses after adjustment for tumorsize and other prognostic factors.

STUDIES ON ESTROGEN-REPLACEMENT THERAPY

IN BREAST CANCER SURVIVORS

DiSaia36 reported on 71 breast cancer survivorswho received ERT. There was no exclusion basedon time interval from diagnosis, stage, age,receptor status, or lymph node status. Womenreceived combination therapy with progestin onlyif they had not previously undergone hysterec-tomy. Later, the author reported a comparison of41 of these ERT survivors to 82 non-ERT breastcancer subjects, matched for both age and stageof disease.37 Survival analysis did not indicate asignificant difference between the two groups.An updated series of 145 patients who receivedERT for at least 3 months after diagnosis hasidentified 13 recurrences. The duration of estro-gen use prior to the diagnosis of recurrent breastcancer ranged from 4 months to 11.5 years (Fig-ures 15–1 and 15–2).

Other authors have reported their experi-ence of ERT in breast cancer survivors. Eden38

reported six recurrences among 90 womenreceiving ERT. These ERT users were matchedtwo to one with control subjects with no historyof hormone use after diagnosis of breast cancer.The recurrence rate was 7 percent in the ERTusers and 30 percent in the non-ERT users.Bluming39 reported on 155 breast cancerpatients who received ERT for between 1 and56 months, among whom 7 recurrences wereidentified. The only published prospective ran-domized trial is being undertaken by Vas-


Table 15–2. ESTROGEN-REPLACEMENT THERAPY IN BREAST CANCER SURVIVORS

Study No. of Patients Stage of Disease Duration of ERT Recurrences

Stoll43 Unknown Early stage 3–6 mo NoneWile et al44 25 All stages 24–82 mo 3Powles et al45 35 All stages 1–44 mo 8Eden et al38 90 Local 4–144 mo 6Vassilopoulou-Sellin et al40 49 0–III 24–142 mo 0

(ER negative only) (oral or vaginal estrogen only)Bluming et al39 155 Local 1–56 mo 7Brewster et al* 145 All stages 3–144 mo 13

*In press.ERT=estrogen-replacement therapy; ER=estrogen replacement.

silopoulou-Sellin.40 Subjects are randomized toeither a placebo or ERT without a progesta-tional agent. Ninety women have been random-ized and 49 have received ERT for a minimumof 2 years. No breast cancer recurrences havebeen observed in the ERT arm. The singlerecurrence was in the placebo arm.

The series discussed above are representativeof the reported experience of ERT in 499 femalebreast cancer survivors (Table 15–2). This groupof women is very heterogeneous with respect tobreast cancer stage, the interval between diagno-sis of breast cancer and initiation of ERT, thehormonal combinations prescribed, estrogen-receptor status, and finally in the duration of useof estrogen. Despite these limitations, it remainsobvious that the use of estrogen is not associatedwith a rash of occurrences. Overall, the data donot suggest that ERT has an adverse effect onbreast cancer outcome.

CONCLUSIONThe fear that administration of estrogen towomen with a history of breast cancer willresult in the activation of quiescent metastaticfoci, as well as the climate of medical litigation,are the basis of much of the reluctance of physi-cians to prescribe this agent. The standard ofcare no longer supports prophylactic oophorec-tomy in young women who do not becomeamenorrheic after cytotoxic therapy. In addition,many women continue to menstruate regularlyafter treatment and may even complete pregnan-cies. If castration and pregnancy termination arenot routinely recommended, why then shouldthe replacement of estrogen at a much lowerdose than is physiologic be prohibited?

Fifty-year-old women have a 13 percent life-time probability of developing breast cancer anda 3 percent probability of dying from it; theyhave a 46 percent chance of developing coronary

260 BREAST CANCER

Figure 15–2. Comparison by known stage of subjects whoreceived estrogen replacement therapy.

0 50 100Months

150

Figure 15–1. Subjects with recurrent breast cancer. The inter-val between initiation of use of estrogen replacement therapyand diagnosis of recurrent breast cancer.

0102030405060708090

100

Recurrence

0I/IIIII/IVUnknown

Stage

24

13

89

6

No Recurrence

Num

ber o

f Pat

ient

s

00

13

0

heart disease and a 31 percent probability ofdying from it. While breast cancer claims 43,000lives per year in the United States, coronaryheart disease will kill approximately 233,000women annually. Nearly 65,000 women die eachyear from the complications of hip fracture.

No guarantee can be made that ERT will beaccompanied with freedom from recurrentbreast cancer, because some women will haverecurrent disease coincident with renewed hor-monal exposure. However, should one discussthe risks and benefits of estrogen replacementtherapy with patients to help them make aninformed decision?

REFERENCES

1. Early Breast Cancer Trialists’ CollaborativeGroup. Systemic treatment of early breast can-cer by hormonal, cytotoxic, or immune ther-apy: 133 randomized trials involving 31,000recurrences and 24,000 deaths among 75,000women. Lancet 1992;339:1–15, 71–85.

2. Mehta RR, Beattie CW, Das Gupta T. Endocrineprofile in breast cancer patients receivingchemotherapy. Breast Cancer Res Treat 1991;20:125–32

3. Colditz GA, Hankinson SE, Hunter DJ, et al. Theuse of estrogens and progestins and the risk ofbreast cancer in postmenopausal women. N EnglJ Med 1995;332:1589–93.

4. Stanford JL, Weiss NS, Voigt LF, et al. Combinedestrogen and progestin hormone replacementtherapy in relation to risk of breast cancer inmiddle age women. JAMA 1995;274:137–41.

5. Dhodapkar MV, Ingle JN, Ahmann DL. Estrogenreplacement therapy withdrawal and regressionof metastatic breast cancer. Cancer 1995;75:43–6.

6. Sacks FM, Walsh BW. The effects of reproductivehormones on serum lipoproteins: unresolvedissues in biology and clinical practice. Ann NYAcad Sci 1990;592:272–85.

7. Gebera OC, Mittleman MA, Walsh BW, et al. Fib-rinolytic potential is significantly increased byoestrogen treatment in postmenopausal womenwith mild dyslipidemia. Heart 1998;80:235–9.

8. Koh KK, Mincemoyer R, Bui MN, et al. Effects ofhormone-replacement therapy of fibrinolysisin postmenopausal women. N Engl J Med1997;336:683–90.

9. Grady D, Rubin SM, Petitti DB, et al. Hormonetherapy to prevent disease and prolong life inpostmenopausal women. Ann Intern Med1992;117:1016–37.

10. Hunt K, Vessey M, McPherson K, Coleman M.Long-term surveillance of mortality and can-cer incidence in women receiving hormonereplacement therapy. Br J Obstet Gynaecol.1987;94:620–35.

11. Bush TL, Barrett-Connor E, Cowan LD, et al. Car-diovascular mortality and noncontraceptive useof estrogen in women. Circulation 1987;75:1102–9.

12. Stampfer MJ, Colditz GA, Willet WC, et al. Post-menopausal estrogen therapy and cardiovasculardisease. Ten-year follow-up from the Nurses’Health Study. N Engl J Med 1991;325:756–62.

13. The Writing Group for the PEPI Trial. Effects ofestrogen or estrogen/progestin regimens onheart disease risk factors in postmenopausalwomen. The Postmenopausal Estrogen/Prog-estin Interventions (PEPI) Trial. JAMA 1995;274:199–208.

14. Grodstein F, Stampfer MJ, Colditz GA, et al. Post-menopausal hormone therapy and mortality. NEngl J Med 1997;336:1769–75.

15. Paganini-Hill A, Henderson VW. Estrogen replace-ment therapy and risk of Alzheimer’s disease.Arch Intern Med 1996;156:2213–17.

16. Tang MX, Jacobs D, Stern Y, et al. Effect ofoestrogen during menopause on risk and age atonset of Alzheimer’s disease. Lancet 1996;348:429–32.

17. Jacobs DM, Tang MX, Stern Y, et al. Cognitivefunction in nondemented women who tookestrogen after menopause. Neurology 1998;50:368–73.

18. Kampen DL, Sherwin BB. Estrogen use and ver-bal memory in healthy postmenopausalwomen. Obstet Gynecol 1994;83:979–83.

19. Rice MM, Graves AB, McCurry SM, Larson EB.Estrogen replacement therapy and cognitivefunction in postmenopausal women withoutdementia. Am J Med 1997;103:26S–35S.

20. Ernst RL, Hay JW. The US economic and socialcosts of Alzheimer’s disease revisited. Am JPublic Health 1994;84:1261–4.

21. Davis RA, Fern F Jr. Effects of ethinyl estradioland phenobarbital on bile acid synthesis andbiliary bile acid and cholesterol excretion. Gas-troenterology 1976;70:1130–5.

22. Lointier P, Wildrick DM, Bowman BM. Theeffects of steroid hormones on a human cancercell line in vitro. Anticancer Res 1992;12:1327–34.


23. Calle EE, Miracle-McMahill HL, Thun MJ, HeathCW Jr. Estrogen replacement therapy and therisk of fatal colorectal cancer in a prospectivecohort of postmenopausal women. J Natl Can-cer Inst 1995;87:517–23.

24. Canney PA, Hatton MQF. The prevalence ofmenopausal symptoms in patients treated forbreast cancer. Clin Oncol 1994;6:297–9.

25. Bergkvist L, Adami HO, Persson I, et al. Progno-sis after breast cancer diagnosis in womenexposed to estrogen and estrogen progesteronereplacement therapy. Am J Epidemiol 1992;130:221–8.

26. Gambrell DR. Proposal to decrease the risk andimprove the prognosis in breast cancer. Am JObstet Gynecol 1984;150:119–28.

27. Henderson BE, Paganini-Hill A, Ross RK.Decreased mortality in users of estrogen replace-ment therapy. Arch Intern Med 1991;151:75–8.

28. Holleb AI, Farrow JH. The relation of carcinomaof the breast and pregnancy in 283 patients.Surg Gynecol Obstet 1962;115:65.

29. Nugent P, O’Connell TX. Breast cancer and preg-nancy. Arch Surg 1985;120:1221–4.

30. von Schoultz E, Johansson H, Wilking N,Rutqvist LE. Influence of prior and subsequentpregnancy on breast cancer prognosis. J ClinOncol 1995;13:430–4.

31. Anderson BO, Petrek JZ, Byrd DR, et al. Preg-nancy influences breast cancer stage at diagno-sis in women 30 years of age and younger. AnnSurg Oncol 1996;3:204.

32. Clark RM, Chua T. Breast cancer and pregnancy:the ultimate challenge. Clin Oncol (R CollRadiol) 1989;1:11–8.

33. Rosner D, Lane W. Oral contraceptive use has noadverse effect on the prognosis of breast can-cer. Cancer 1986;57:591–6.

34. Schonborn I, Nischan P, Ebeling K. Oral contra-ceptive use and the prognosis of breast cancer.Breast Cancer Res Treat 1994;30:283–92.

35. Sauerbrei W, Blettner M, Schmoor C, et al. The

effect of oral contraceptive use on the progno-sis of node positive breast cancer patients. Ger-man Breast Cancer Study Group. Eur J Cancer1998 Aug;34(9):1348–51.

36. DiSaia PJ, Odicino F, Grosen EA, et al. Hormonereplacement therapy in breast cancer [letter].Lancet 1993;342:1232.

37. DiSaia PJ, Grosen EA, Kurosaki T, et al. Hormonereplacement therapy in breast cancer survivors.A cohort study. Am J Obstet Gynecol 1996;174:1494–8.

38. Eden JA, Bush T, Nand S, Wren BG. A case con-trol study of combined continuous estrogen-progestin replacement therapy among womenwith a personal history of breast cancer.Menopause 1995;2:67–72.

39. Bluming AZ, Waisman JR, Dosik GM. Hormonereplacement therapy (HRT) in women with pre-viously treated primary breast cancer. Update III.Proc Am Soc Clin Oncol 1997;16A:463,131a.

40. Vassilopoulou-Sellin R, Therriault R, Klein MY.Estrogen replacement therapy in women withprior diagnosis and treatment for breast cancer.Gynecol Oncol 1997;65:89.

41. Falkeborn M, Persson I, Adami HO, et al. The riskof acute myocardial infarction after oestrogenand oestrogen-progestogen replacement. Br JOb Gyn 1992;99:821–8.

42. Psaty BM, Heckbert SR, Atkins D, et al. The riskof myocardial infarction associated with thecombined use of estrogens and progestins inpostmenopausal women. Arch Intern Med1994;154:1333–9.

43. Stoll BA. Hormone replacement therapy inwomen treated for breast cancer. Eur J CancerClin Oncol 1989;25:1909–13.

44. Wile AG, Opfell RW, Margileth DA. Hormonereplacement therapy in previously treated breastcancer patients. Am J Surg 1993;165:372–5.

45. Powles TP, Casey S, O’Brien M, Hickish T. Hor-mone replacement after breast cancer. Lancet1993;342:60.

262 BREAST CANCER

263

In recent years, extensive clinical trials haveestablished the roles for conservative surgery,radiation, and adjuvant chemo/hormonal ther-apy in the primary therapy of breast cancer. Itmay seem self-evident that repeated postopera-tive contact between cancer patients and theirphysicians, that is, follow-up, is a good thing.Follow-up practice patterns vary greatly, withsome oncologists frequently following theirbreast cancer patients with various intensiveinvestigation and others only doing sporadicfollow-up. The possible beneficial effects offollow-up include:

1. Management of postsurgical complications.This is essential and need not be elaboratedupon here.

2. Early detection of recurrence or of new primaries.

3. Reassuring patients. This can be a double-edged sword, as some patients are reassuredby the process while others are made anx-ious by an impending visit to the physician.

4. Measurement of quality control of out-comes. Participation in clinical trials orAmerican College of Surgeons-sponsoredaudits can be helpful.

This chapter evaluated criteria for follow-upand provides background on the principal ofscreening and economic analysis.

DEFINITIONS

It will be useful to review some of the termsused in analyzing the surveillance data. Thedecision matrix is a term most commonlyapplied to the simple decision of whether thedisease is present (D+) or absent (D-) whenthe test is abnormal, that is, positive (T+) ornormal, that is negative (T–). When these twobinary results are plotted on a 2 ´ 2 table, fourpossible combinations form the ratios shownbelow.1–3

The true positive (TP) ratio represents theproportion of positive tests in all patients whohave the disease. The ratio therefore expressesthe sensitivity of the test and can be express-ed as:

sensitivity =diseased subjects with positive test

´ 100all diseased subjects tested

More simply stated, sensitivity is deter-mined by the false-negative (FN) ratio, which isthe proportion of negative tests in all patientswith the disease. The TP and the FN ratio thenrepresent the sensitivity of the test.

The false-positive (FP) ratio is the propor-tion of positive tests in all patients who donot have the disease. The true-negative (TN)ratio is the proportion of negative tests in allpatients who do not have the disease. The FP

16

Surveillance of the Breast Cancer PatientJANARDAN D. KHANDEKAR, MD

and TN ratios then express specificity, defin-ed as:

sensitivity =nondiseased subjects with negative test

´100all nondiseased subjects tested

The sensitivity of a test under considerationis usually determined by evaluating its efficacyagainst a known standard. Depending on the sen-sitivity of that standard, the sensitivity can bespuriously high or low. For example, the sensi-tivity of a bone scan is usually evaluated in rela-tion to radiography, a technique which itself isnot very sensitive. Therefore, the sensitivity ofthe bone scan is very high, in the vicinity ofabout 99 percent. However, if in the future evenmore sensitive tests for detecting bone metasta-sis become available, such as use of either betterimaging techniques and/or polymerase chainreaction-based molecular assays, the sensitivityof the bone scan will decrease. The specificity ofthe test is also dependent upon the FP ratio. Gen-erally, the FP rate increases with more data, andthe initial specificity of a given diagnostic testdecreases with time. These factors then governthe sensitivity and specificity of various diag-nostic tests. When sensitivity and specificity of atest are determined, it is possible to calculate thepredictive value (PV).

The positive PV is defined as the likeli-hood that a subject yielding a positive testactually has the disease. Conversely, a nega-tive PV indicates the likelihood that a subjectwith a negative test does not have the disease.This likelihood is related to the actual preva-lence of disease in the total population. Moresimply stated, PV for positive tests can bedefined as the percentage of time that a posi-tive test will detect the diseased individual.The PV of a positive test can be calculated asfollows:4

number of diseased subjects (or proportion)

PV =with positive test

´ 100total number (or proportion) of subjects

with positive test

The PV of a negative test can be calculatedas follows:

number (or proportion) of

PV =nondiseased persons with positive test

´ 100total number (or proportion) of

persons with positive test

The usefulness of a diagnostic test is there-fore directly proportional to the prevalence ofthe target disease in the population. The FP rateof a test is usually constant and is often relatedto the test and not the disease itself. Therefore,when the prevalence of the target disease is low,the PV of the positive test is also low since FPis constant. On the other hand, if there is ahigher prevalence of the disease, the PV of apositive test will also be high.

The evaluation of various diagnostic testsare also affected by various biases that signifi-cantly affect the interpretation of the data.These are outlined below.

Lead-Time Bias

The apparent increased duration of survival sim-ply reflects a longer time that the recurrence wasclinically known but there is no true gain inlongevity. In other words, there is an illusion ofan increased survival because of the longer dura-tion of observation but there is no impact onmortality rate. It is because of this considerationthat reduction of mortality rate has become the“gold standard” in evaluating the impact of adiagnostic or therapeutic intervention.

Length-Time Bias

An event such as cancer detected in an asymp-tomatic phase often has an indolent course andis therefore detected at the time of evaluationrather than between the visits. Cancers that areaggressive will often present with symptoms inthe intervening visits, creating the illusion thatmore intensive surveillance would haveresulted in detecting of the disease earlier and abetter outcome.

264 BREAST CANCER

EVALUATION OF VARIOUS TESTSUSED IN BREAST CANCER

The American Society of Clinical Oncology(ASCO) established an expert panel to evaluatethe use of various tests in breast cancer.5 Thepanel modified the scale developed by theCanadian Taskforce on Periodic Health Exami-nation to evaluate various tests (Table 16–1).6

There are only two prospective randomized tri-als that have evaluated the impact of a multi-tude of surveillance tests on the overall survivaland quality of life in breast cancer patients.They fulfill the criteria of providing Level 1,that is, highest level of evidence. However,beginning with studies published in 1979 byWinchester and colleagues,7 a large databasehas been developed that has retrospectivelyanalyzed the value of various diagnostic tests inthe follow-up of breast cancer patients. Thesestudies have tried to answer some of the fol-lowing questions:

1. Do the available tests diagnose early asymp-tomatic recurrence in breast cancer? If so,which tests are useful?

2. Does early detection and recurrence resultin better therapy and thus improve qualityand quantity of survival? (This is the mostimportant question).

3. What is the cost-benefit analysis for the pos-sible improved quality or quantity of life?

Winchester and colleagues analyzed 87patients with recurrent breast cancer for pat-

terns of recurrence and methods of detection.7

In 79 of 87 patients, recurrence was detectedby symptoms such as pain or shortness ofbreath, while physical examination detected anadditional 5 patients with recurrence. Reviewof the literature indicates that only 12 to 22percent of recurrences occur in truly asympto-matic women.8–14 In the prospective intergroupfor cancer care evaluation trial (GIVIO), 31percent of the recurrences were detected inasymptomatic patients in the intensely investi-gated group compared to 21 percent of recur-rences in the control group.11 However, therewas no effect on survival in patients detected inthe asymptomatic phase. It has been arguedthat patients with local recurrence only thosewho undergo aggressive therapy have a 50 per-cent 5-year survival, and may therefore have animproved outcome. However, it can be arguedthat patients with local recurrence may havemore biologically inert disease, as these wereretrospective studies. Dewar and Kerr in anEnglish study reported that of 546 breast can-cer patients followed with 6,863 clinic visits,only 1 percent of the visits were associatedwith recurrences that were curable.14 Theseauthors have therefore questioned even routinephysical examination. In their studies, recur-rences were found five times more often dur-ing spontaneous visits than during routine vis-its, illustrating the lead-time bias. Dewar andKerr suggest that negative physical examina-tions may give false assurance to patients,leading to an even further delay in diagnosis ofa recurrence.

Surveillance of the Breast Cancer Patient 265

Table 16–1. MODIFIED CANADIAN CRITERIA FOR EVALUATING DIAGNOSTIC TEST EVIDENCE

Level of Evidence Type of Evidence for Recommendation

Level 1 Meta-analysis or large high-powered concurrently controlled studies with a primary objective to evaluate(highest level) the utility of a given testLevel 2 Prospective clinical trials designed to test given hypothesisLevel 3 Large size retrospective trialsLevel 4 Similar to Level 3, but even less reliable. Comparative and correlative descriptive and case studies can

be included.Level 5 Case reports and clinical examples

Adapted from Khandekar JD. Preoperative and postoperative follow up of cancer. In: Winchester DJ, Scott Jones R, Murphy GP, editors. Surgicaloncology for the general surgeon. Philadelphia, PA: Lippincott, Williams and Wilkins; 1999.p. 43–54.

Scanlon and colleagues carefully analyzedthe information from the last examination torecurrence in 93 patients.8 They reported that43 percent of recurrences were detected within3 months of the last examination, 64 percentwithin 6 months, and 94 percent within 1 year.They therefore recommended that examina-tions be conducted every 3 months for the first2 to 3 years after primary therapy for breastcancer, and then at a reduced interval.

The vast majority of recurrences aredetected by history and physical examination.Although the impact of such detection on over-all survival is unknown, it may have the psy-chological benefit of reassuring the patient ofhaving had a contact with their physician. Sincethe cost of such surveillance is approximately$150 per annum per patient and allows evalua-tion of other parameters such as the effects ofprimary therapy, physical and psychosocialrehabilitation, and detection of contralateralprimary, the current author agrees with ASCO’srecommendation that the patients should beseen at 3-month intervals for the first 2 yearsthen every 6 months for the next 3 years.5

Chemistry

An abnormal chemical evaluation has been thefirst evidence of recurrent breast cancer inapproximately 1 to 12 percent of patients.7,14 Anabnormal blood count is rarely seen as a firstindicator of recurrent breast cancer. Hannisdaland colleagues reviewed their experience with430 patients.15 In 8 of 430 patients, an elevatederythrocyte sedimentation rate, gamma glu-tamyl transferase, and alkaline phosphatase val-ues heralded recurrences. The sensitivity andspecificity of these tests for relapse was 55 and91 percent, respectively.15 The ASCO panel rec-ommends no laboratory tests in the follow-up.5

However, the chemotherapeutic drugs used asadjuvant therapy can be leukemogenic, andperiodic blood tests may have to be undertakento detect these changes. Further, many patientswith breast cancer have other concurrent dis-

eases and/or can develop new problems. There-fore, an occasional blood test may be warranted.

Tumor Markers: CA-15-3 and CEA

Hayes and colleagues defined the marker thatrepresents qualitative or quantitative alteration ordeviation from normal of a molecule, substance,or process that can be detected by some type ofassay.16 This includes measurement of a gene,RNA, a product such as protein, carbohydrate, alipid, or a process such as vascular density.Markers used in follow-up are considered below.

CA-15-3

The CA-15-316 marker measures the serum levelof a mucin-like membrane glycoprotein that isshed from the tumor cells into the blood stream.Recently, the Food and Drug Administrationapproved the Truquant assay which uses a mon-oclonal antibody, CA-27-29, to measure CA-15-3-like antigen. This marker is elevated only inpatients with advanced disease.16 The level ofCA-15-3 is highest in patients with liver or bonemetastasis. The ASCO panel evaluated 12 stud-ies reporting on the value of CA-15-3 in detect-ing asymptomatic recurrent breast cancer. Ofthese studies, only seven could be properly ana-lyzed.5 Of 1,672 patients followed in these trials,352 developed recurrence. About two-thirds ofthese were detected by an elevated CA-15-3before other parameters revealed recurrence.The mean lead-time from marker elevation toclinical diagnosis was 5 to 7 months. However,the sensitivity of the test was only 57 to 79 per-cent. It is also not known whether such earlydetection leads to improved survival. If andwhen new therapies for the treatment of metasta-tic breast cancer are developed, the role of CA-15-3 may need re-evaluation.

CEA

The value of CEA5 in detecting recurrent breastcancer is even less than that of CA-15-3. At

266 BREAST CANCER

present, there are no indications for using CEAin routine surveillance of breast cancer patients.

Mammography

Patients with breast cancer are at higher risk fordeveloping contralateral breast cancer. Further,patients with breast cancer who have undergoneconservative therapy with lumpectomy or radi-ation are at risk for ipsilateral recurrence. Cur-rently, it is recommended that patients whohave undergone unilateral mastectomy shouldundergo contralateral mammography on ayearly basis. Patients who have been treatedwith conservative techniques and develop localrecurrence can be salvaged by appropriatetreatment such as mastectomy. In the GIVIOtrial,11 patients randomized to intensive follow-up had an 11.4 percent incidence of contralat-eral breast cancer, versus 6.6 percent in thegroup with only routine follow-up. Local ther-apy for ipsilateral breast cancer recurrence ordetection of a new primary in the contralateralbreast will improve quality and quantity of life.

Routine Chest Radiography

Several investigators have studied the value ofannual chest radiography in patients with breastcancer. In general, only 0.2 to 4 percent of radi-ographs were abnormal in truly asymptomaticpatients. As pointed out by Loprinzi,17 the onlyvalue of routine annual chest radiography is todetect lymphangitic pulmonary metastasisbefore it causes significant pulmonary symp-toms and thus impairs quality of life. Althoughthe ASCO panel does not recommend annualchest radiography,6 the current author agreeswith Loprinzi17 that it may be beneficial up to 3years following primary therapy. It is unlikelythat this will improve survival, but it can behelpful in preserving quality of life. Patientswith aggressive breast cancer tend to recur inthe first 2 years following primary therapy,meaning that the annual radiograph can beceased after 3 years of follow-up.3

Bone Scans

In 1979, it was proposed on the basis of Bayesdecision analysis that since there is no evidencethat early detection and therapy for metastaticbreast cancer alters the clinical course of thepatient, bone scans should be performed only insymptomatic patients. Since then, several stud-ies have confirmed this recommendation. TheEastern Cooperative Group confirmed the find-ing in the mid-1980s.18 The National SurgicalAdjuvant Breast and Bowel Project (NSABP)followed 1,989 patients on the B-09 arm oftheir study.19 Of these patients, 779 had treat-ment failure, of whom roughly one-fifth hadrecurrences limited to bone. Only 52 (0.6%)patients had screening scans that were useful indetecting lesions in asymptomatic patients. TheNSABP changed their recommendation aboutsurveillance of breast cancer patients in 1994based on this study.

In the GIVIO trial, in which patients wererandomized between intensive follow-up versusobservation only, compliance was > 80 percentin both groups.11 At a median follow-up of 71months, there was no difference in overall sur-vival in the two groups. The study also showedno impact on the quality of life because ofintensive intervention. Another Italian trial, thatof Del Turco and colleagues,12 evaluated 1,243consecutive patients with intensive interventionversus minimum follow-up. In this study, therewas increased detection of isolated intratho-racic and bone metastasis in the intensive fol-low-up group compared to clinical follow-upgroup. No difference was observed, however,for other sites, and 5-year overall mortality was18.6 versus 19.5 percent (statistically insignifi-cant difference) between the two groups.

The bone scan has a false-positive rate ofapproximately 15 percent.2 If the prevalence ofmetastasis is low, the predictive value of a pos-itive test will be low and the patient may besubjected to additional unnecessary interven-tions. On the basis of Baye decision analysis, itwas pointed out that routine bone scans in sur-


veillance of a breast cancer patient will lead toa low predictive value of a positive test andunnecessary and expensive interventions.2 It istherefore recommended that bone scans be per-formed only in breast cancer patients who aresymptomatic with bone pain or with significantelevations in their alkaline phosphatase.

Imaging Studies of the Liver

There have been no prospective studies com-paring sensitivity and specificity of CT andMRI evaluations in the postoperative surveil-lance of breast cancer patients. However, onthe basis of evaluation in preoperative breastand lung cancers, it is safe to conclude that inthe absence of symptoms and abnormal liverfunction tests, these imaging techniques willbe of little or no value. The Italian GIVIOinvestigators performed liver echography intheir intervention group. In that study, 6.5 per-cent of patients had their first recurrence diag-nosed by liver echography, versus 6.1 percentin the control group, who had echographybecause of abnormal examination and/orhepatic function tests.11 Based on these stud-ies, it can be concluded that routine imagingtechniques for detecting liver metastasis arenot warranted.

Table 16–2 summarizes recommendationsfor follow-up of breast cancer patients whohave completed their primary therapy, and,when appropriate, adjuvant chemotherapy.

IMPLICATIONS FOR HEALTH-CARE COSTS

In response to spiraling health-care costs, man-aged care was aggressively introduced.20 Man-aged care is under intense pressure these days,and its future is uncertain. However, the silverlining of managed care has been that it hasforced clinicians to critically evaluate their prac-tices. This has resulted in developing pathwaysand guidelines for various illnesses. Cohen andcolleagues argue that physicians have a respon-sibility to assure the delivery of appropriatehealth care without sacrificing the quality ofcare.21 The upper limit for an acceptable cost-effectiveness ratio remains controversial. More-over, it is important to bear in mind that not onlyquantity of life but the impact of an interventionon the quality of life should be measured.22

Measurement of quality of life is still somewhatsubjective, but as new tools are developed, itmust be incorporated into cost analyses.

In two prospective clinical trials conductedin Italy,11,12 no cost-effective analysis wasavailable. Schapira attributed savings of $636million in 1990 costs and projected a $1 billionsaving in the year 2,00023 when a minimal fol-low-up schema, as described here, isemployed. However, it is unclear whether hisanalysis includes additional expenses incurredas the result of FP tests, which can lead toadditional interventions. The negative psycho-logical impact of an FP test cannot be quanti-

268 BREAST CANCER

Table 16–2. PROPOSED FOLLOW-UP SCHEDULE

Procedure Low-Risk* High-Risk †

History and physical examination 3 mo ´ 2 yr and then 3 mo ´ 2 yr6 mo ´ 3 yr then yearly 6 mo ´ 3 yr then yearly

Complete blood count and chemistry Every 6 mo ´ 2 yr 3 mo ´ 2 yrand then yearly 6 mo ´ 3 yr then yearly

Markers: CEA, CA-15-3 — —Mammogram Yearly YearlyChest radiograph Yearly YearlyScans (bone, liver) — —

*Patients with negative nodes and/or positive receptors.†Patients with positive nodes and/or received chemotherapy.Adapted from Khandekar JD. Preoperative and postoperative follow up of cancer. In: Winchester DP, Scott Jones R, Murphy GP, editors.Surgical oncology for the general surgeon. Lippincott Williams and Wilkins; 1999. p.

fied at this time, but needs to be evaluated infuture analyses.

In summary, the minimal surveillanceschema proposed has considerable implicationsfor health-care costs. Reallocation of health-care resources to areas that will lead toimproved survival and quality of life is animportant aspiration for health-care policy.

LEGAL IMPLICATIONS

In our litiginous society, there is apprehensionabout the legal consequences of a delayed diag-nosis, even for the metastatic disease. It isimportant to educate the public as well as thelegal profession on the differentiation betweendiagnosis of a primary breast cancer and that ofa metastatic disease. If in the future better andmore effective treatments become available thatchange the natural history of the disease, detec-tion at an earlier time may become important.The guidelines developed by associations suchas ASCO5 and the Society of Surgical Oncol-ogy21 are helpful for physicians protectingagainst these legal threats.

CONCLUSION

It is only in recent years, primarily because ofeconomic pressure, that guidelines for surveil-lance have been developed and adapted. In1999, about $1.3 trillion will be spent on healthcare.24 Several studies as well as the analysispresented here indicate that a minimal surveil-lance approach is clearly warranted.17,25 Onecan argue that there need not be any follow-upfor breast cancer patients after primary inter-vention. Although routine history and physicalexaminations may not have a direct benefit interms of survival, they have an immense psy-chological effect. Most patients need reassur-ance, which leads to self-confidence. Visits alsoprovide time to educate patients and discusspsychosocial and physical rehabilitation—it isimperative that patients do not feel abandonedby their physicians. These visits also allow diag-

nosis of local and regional recurrences as wellas new contralateral breast cancers, which canbe cured. Although controversial, the currentauthor continues to believe that an annual radi-ograph for the first 3 years after primary treat-ment and occasional blood tests are indicated.On the other hand, more expensive tests such asbone scans, ultrasound, and CT scans of thechest, brain, and liver, and measurement oftumor markers, are not indicated.

At this time, there is no evidence thatmetastatic disease, when detected early, can becured by present techniques. However, the rec-ommendation of minimal follow-up may needto be altered if early intervention in recurrentbreast cancer will lead to improved survival.Patients who are in clinical trials should have amore intensive follow-up to differentiate thedisease-free interval and overall survival in thecontrol and experimental groups. However, theclinical trial should keep these diagnostic teststo a minimum so that managed care and otherhealthcare providers do not object to extraexpenses. To improve therapy, it is importantthat an increasing number of patients beenrolled in clinical trials. Therefore, true collab-oration needs to be developed between variouscooperative groups and healthcare providers. Itis important to improve the health-care ofpatients while giving appropriate considerationto the problem of escalating health-care costs.

REFERENCES

1. McNeil BJ, Keller E, Adelstein SJ. Primer on cer-tain elements of medical decision-making. NEngl J Med 1975;293:211.

2. Khandekar JD. Role of routine bone scans in oper-able breast cancer: an opposing viewpoint.Cancer Treat Rep 1979;63:1241.

3. Khandekar JD. Preoperative and postoperative fol-low-up of cancer. In: Winchester DP, Scott JonesR, Murphy GP, editors. Surgical oncology forthe general surgeon. Philadelphia, PA: Lippin-cott, Williams & Wilkins; 1999. p. 43–54.

4. Vecchio TJ. Predictive value of a single diagnostictest in unselected populations. N Engl J Med1966;274:1171.


5. ASCO Special Article. Recommended breast can-cer surveillance guidelines. J Clin Oncol 1997;15:2149.

6. Canadian Medical Association. The Canadian taskforce on the periodic health examination. CanMed Assoc J 1979;121:1193.

7. Winchester DP, Sener SF, Khandekar JD, et al.Symptomatology as an indicator of recurrent ormetastatic breast cancer. Cancer 1979;43:956.

8. Scanlon EF, Oviedo MA, Cunnigham MP, et al. Pre-operative and follow-up procedures on patientswith breast cancer. Cancer 1980;46:977.

9. Gerber FH, Goodreau JJ, Kirchner PT, Fouty WF.Efficacy of preoperative and postoperativebone scanning in the management of breastcarcinoma. N Engl J Med 1977;297:300.

10. Burke W, Daly M, Garber J, et al. Recommenda-tions for follow-up care of individuals with aninherited predisposition to cancer. II. BRCA1and BRCA2. JAMA 1997;277:997.

11. The GIVIO Investigators. Impact of follow-up test-ing on survival and health-related quality of lifein breast cancer patients. A multicenter random-ized controlled trial. JAMA 1994;271:1587.

12. Del Turco MR, Palli D, Cariddi A, et al. Intensivediagnostic follow-up after treatment of primarybreast cancer. A randomized trial. JAMA 1994;271:1593.

13. Tomin R, Donegan WL. Screening for recurrentbreast cancer: its effectiveness and prognosticvalue. J Clin Oncol 1987;5:62.

14. Dewar JA, Kerr GR. Value of routine follow-up ofwomen treated for early carcinoma of thebreast. BMJ 1985;291:1464.

15. Hannisdal E, Gundersen S, Kvaloy S, et al. Follow-up of breast cancer patients stage I-II: a baselinestrategy. Eur J Cancer 1993;29A:992–7.

16. Eskelinen M, Hippelainen M, Carlsson L, et al. Adecision support system for predicting a recur-rence of breast cancer; a prospective study ofserum tumor markers TAG 12, CA 15-3 andMCA. Anticancer Res 1992;12:1439–42.

17. Loprinzi CL. It is now the age to define the appro-priate follow-up of primary breast cancerpatients. J Clin Oncol 1994;12:881.

18. Pandya KJ, McFadden ET, Kalish LA, et al. A ret-rospective study of earliest indicators of recur-rence in patients on Eastern CooperativeOncology Group adjuvant chemotherapy trialsfor breast cancer. A preliminary report. Cancer1985;55:202.

19. Wickerham L, Fisher B, Cronin W. The efficacy ofbone scanning in the follow-up of patients withoperable breast cancer. Breast Cancer ResTreat 1984;4:303.

20. Bodenheimer T. The American health care system:physicians and the changing medical market-place. N Engl J Med 1999;340:584–8.

21. Cohen AM, Bland KI, Gardner B, Winchester DP.Society of Surgical Oncology and PracticeGuidelines. Oncology 1997;11:869.

22. Cella D, Fairclough DL, Bonomi PB, et al. Qual-ity of life (QOL) in advanced non-small celllung cancer (NSCLC) results from EasternCooperative Oncology Group (ECOG) studyE5592. Proc ASCO 1997;16(4):2a.

23. Schapira DV, Urbrn N. A minimalist policy forbreast cancer surveillance. JAMA 1991;265:380.

24. Modern Healthcare. Economic data: some thingsnever change. Modern Healthcare 1999;29:16.

25. Schapira DV. Breast cancer surveillance: a cost-effective strategy. Breast Cancer Res Treat1993;25:107.

270 BREAST CANCER

271

Despite the improvements in prognosis achievedfor many patients with breast cancer, approxi-mately 46,000 women die of this disease eachyear. The increase in incidence of breast cancerseen through the early 1990s has been success-fully offset by two factors: widespread applica-tion of screening mammography, permittingmore frequent early diagnosis and, more recently,the decrease in recurrence and mortality ratesachieved through the now standard application ofeffective systemic adjuvant therapy.1 Unfortu-nately, similar gains have not been achieved forwomen who present with metastatic breast canceror for those with distant disease relapse after ini-tial treatment. For these women, palliation ofsymptoms and some prolongation of survival ispossible but there is no known curative treatment.In fact, the death rate for this disease hasremained stubbornly constant over decades.2

Metastatic breast cancer is the second most com-mon cause of cancer death among women.

While only 6 percent of patients present ini-tially with metastatic breast cancer,2 metastaseswill eventually develop in at least 30 percent ofpatients with node-negative primary breast can-cer and 50 percent of those with positive nodesat diagnosis. The event rate for recurrence is rel-atively constant over the first 10 years forwomen who receive adjuvant therapy, that is,each year healthy survivors face the same riskfor recurrence as they did in the preceding year.1

The most common sites of metastatic involve-ment are bone, lungs, and liver.3 While both infil-trating ductal carcinoma and infiltrating lobularcarcinoma will relapse at the same rate over timebased on their size and degree of nodal involve-ment, a predilection for certain sites of involve-ment can be related to histology. Infiltrating lob-ular carcinomas are more likely to recur in bonemarrow, peritoneum, pelvic organs, andmeninges than are infiltrating ductal carcinomas.On the other hand, lung metastases are morecommon with infiltrating ductal cancers.4–6

When metastatic breast carcinoma is firstdiagnosed, a brief staging workup is indicated todetermine the extent of disease and thus treat-ment priorities. In the absence of symptoms,chest radiography, abdominal CT scan, andbone scan are needed. Any abnormalities onbone scan should be further pursued with boneradiographs, to confirm metastatic disease anddetermine whether it is lytic or blastic in nature.The finding of lytic disease, particularly inweight-bearing bones, has specific palliativeimplications. In addition to the above studies,any symptoms should be fully investigated.

Prior to beginning treatment for metastaticdisease, biopsy to confirm clinical suspicionsshould be considered mandatory in all but themost unusual of circumstances. The diagnosisof incurable metastatic disease has obvious andprofound prognostic implications, and often

17

Treatment of Metastatic Breast CancerDOUGLAS E. MERKEL, MD

commits a patient to lifelong systemic treat-ment. The possibility of a benign lesion musttherefore be excluded. In a disease such asbreast cancer, which may recur many yearsafter initial treatment, the possibility of a sec-ond primary carcinoma must also be consid-ered and excluded. In addition to confirmingthe diagnosis of metastatic cancer, the tumorshould be analyzed for estrogen receptor, prog-esterone receptor, and HER2/c-erbB2 protein,7

as these results will largely determine theoptions for systemic therapy.

The prognosis for metastatic breast cancer isrelated to a number of variables, perhaps mostimportantly to the disease-free interval, or theduration of time between initial diagnosis andrecurrence. This duration provides some mea-sure of the growth rate of the cancer; longer sur-vivals are reported when the disease-free inter-val exceeds several years.8 The extent ofmetastatic involvement, or the number ofinvolved sites, also has an impact on survival, asdoes location.9 There is a particularly goodprognosis observed for patients with a singlemetastatic focus amenable to surgery or radio-therapy.10 Survival in excess of 2 years is alsocommon when the disease is limited to bone butis not expected when there is visceral involve-ment.11 Finally, improved survival is reportedfor estrogen receptor-positive metastatic breastcancer,12 although other markers such as ploidy,S-phase fraction, and HER2 status are not infor-mative.13 The relationship of estrogen receptorstatus and prognosis may not be independent ofother factors, however, as these metastases arealso more likely to be found in bone and soft tis-sue, and to occur at a longer disease-free inter-val, than those lacking estrogen receptors.14

As many patients will present initiallywith only a single site of metastasis, or adominant lesion, treatment considerations forspecific sites of metastases will be surveyedbelow. Systemic measures designed to palli-ate symptoms and offer some hope of delay-ing the progression of metastatic disease willthen be discussed.

BRAIN METASTASES

Brain metastases are diagnosed in 16 to 25percent of all women with breast cancer, but thebrain is seldom the first site of relapse.15 Thus,while brain imaging with gadolinium-enhancedMRI is not part of the routine initial workup fornewly diagnosed metastatic breast cancer, anypatient with new neurologic complaints shouldbe promptly evaluated. Most commonly, a pro-gressively worsening headache develops overdays to weeks. Other common clinical featuresof brain metastases include behavioral or cogni-tive changes, focal weakness, ataxia, speech dis-orders, and seizures.16

Papilledema is present in only 15 percent ofpatients, and the screening neurologic exam maybe negative. Any of the above symptoms are thusindications for scheduling a gadolinium-enhanced MRI. Computed tomography (CT)scan is less sensitive and more likely to result inequivocal or false-positive findings. These scanscannot detect meningeal involvement and shouldonly be obtained where MRI is unavailable.17

Corticosteroids, usually dexamethasone at adosage of 4 mg every 6 hours, can produceimmediate but shortlived improvement in neu-rologic symptoms and are indicated as initialtreatment in all patients with strongly suspectedor newly diagnosed brain metastases.18 Anti-convulsants, however, should be reserved forthe 20 to 30 percent of patients who suffer focalor generalized seizures.19

Treatment of brain metastases may consistof either surgical extirpation, whole brainradiotherapy, or stereotactic “gamma-knife”radiosurgery. As retrospective analyses and asingle randomized trial have demonstratedimproved neurologic control and survival forpatients undergoing surgery for brain metas-tases, resection must be the first considerationfor all appropriate patients.20 The most appro-priate candidates for resection have single,accessible lesions, particularly those that arerelatively bulky and thus unlikely to respondcompletely to radiotherapy. The surgical candi-

272 BREAST CANCER

date should also be one whose other sites ofmetastatic disease are responding, or are likelyto respond to systemic therapy; for those whoseexpected survival is limited, surgical interven-tion has little or no advantage over radiotherapy.

Radiation therapy is indicated as initial pal-liative treatment for all other patients, for exam-ple, those with multiple lesions or poorly con-trolled systemic disease. Median survival forpatients treated in this fashion is 3 to 6 monthsbut a majority receive symptomatic benefit.21

Those who survive for ³ 1 year or longer afterwhole brain radiotherapy are at risk for a varietyof complications ranging from subtle cognitivedeficits to leukoencephalopathy manifesting asprogressive dementia with ataxia. This is pri-marily a concern in good-risk patients and hasmade the use of adjuvant whole brain radiother-apy following surgical removal of a solitarybrain metastasis controversial.

Stereotactic radiosurgery is a new techniquethat delivers a single, large, tightly focuseddose of radiation to a metastatic site, using mul-tiple beams. This technique is highly effectivefor tumors < 3 cm, can be performed on an out-patient basis, and appears to result in far lessrisk of long-term damage to surrounding nor-mal tissue.22 While the current treatment ofchoice for recurrent disease after whole brainradiotherapy, and for patients with surgicallyinaccessible lesions, this technique may in timereplace primary surgery for some patients.

Perhaps surprisingly, brain metastases frombreast cancer have been reported to respondfavorably to systemically administered chemo-therapy23 or tamoxifen.24 While currently no tri-als have established this as frontline therapy forbrain metastases, this approach can certainly betried in patients who relapse following whole-brain radiotherapy, or those who decline toundergo it.

LEPTOMENINGEAL CARCINOMATOSIS

Involvement of the leptomeninges occurs in upto 5 percent of patients with breast cancer, usu-

ally in the setting of disseminated, progressivedisease.25 As mentioned above, this complica-tion is more commonly observed in patientswith infiltrating lobular cancer. The majority ofpatients will present with neurologic signsreferable to some combination of cerebrum,cranial nerves, and spinal cord, although thepatient may complain only of a single symp-tom.26 The single-most common complaint isweakness of the legs, perhaps accompanied bypain or paresthesias. Cranial nerve involvementcan produce diplopia, facial numbness or weak-ness, and hearing loss. Involvement of the cere-bral cortex is heralded by headache, impairedmemory, lethargy, and nausea.

Definitive diagnosis of leptomeningeal car-cinomatosis is difficult, as initial cytologicexamination of the cerebrospinal fluid (CSF) isfalsely negative in up to 46 percent ofpatients.27 Elevated CSF protein levels andmonocytosis may be observed, and repeatedsampling may yield positive cytology. Gadolin-ium-enhanced MRI of any areas of clinicalinvolvement should be obtained, both to ruleout parenchymal brain metastases or epiduralcord compression and to detect enhancing,nodular meningeal enhancement—this may beseen along the convexity of the cerebrum, alongthe brain stem, or involving spinal nerve rootsin up to 70 percent of patients.28

Treatment of leptomeningeal carcinomatosisis difficult, as it often arises in the midst of pro-gressive systemic breast cancer and there hasbeen no optimal approach established. Radia-tion therapy is usually administered to areas ofbulky or symptomatic disease, although studiesto establish this practice are lacking. Radiationtherapy to the entire neuraxis is to be avoided asit can result in severe and prolonged myelosup-pression, thus preventing the subsequent admin-istration of systemic chemotherapy.

As the entire neuraxis is potentially at riskfor leptomeningeal spread, direct CSF installa-tion of chemotherapy is also indicated. Becauseof improved distribution of drug throughout theCSF, intraventricular administration via an

Treatment of Metastatic Breast Cancer 273

Ommaya reservoir is preferred over lumbarpuncture. Methotrexate 12 mg two or threetimes weekly has been used most often, withimprovement reported in 60 to 80 percent ofpatients.26,29 The most common complication istransient aseptic meningitis, manifesting asheadache, fever, and stiff neck. Particularly withsimultaneous cranial radiotherapy, a necrotizingleukoencephalopathy with impaired mentationand focal defects may develop.29 Leakage ofmethotrexate outside of the CSF may result inmucositis or myelosuppression but may becounteracted by concurrent administration oforal or intravenous folinic acid. The median sur-vival for patients who develop carcinomatosismeningitis is 3 to 6 months, although respon-ders may live in excess of 1 year.

MALIGNANT EFFUSIONS

Breast cancer is the most common cause ofmalignant pleural effusions in women. They aremore commonly seen ipsilateral to the primarytumor, suggesting that the effusion sometimesarises via direct extension through the chest wallor through involvement of internal mammarylymph nodes. While 80 percent of malignantpleural effusions arise in the presence of othersites of metastatic involvement,30 they are usu-ally symptomatic and require specific treatment.

In a previously untreated patient with newlymetastatic breast cancer, an attempt may bemade to relieve the malignant pleural effusionwith therapeutic thoracentesis and initiation ofsystemic chemotherapy or hormonal therapy. Inthis setting, a positive response to systemictherapy is likely and may be sufficiently rapidto prevent reaccumulation of fluid. In patientswith previously treated metastatic disease,however, the likelihood of objective response toany systemic therapy is certainly < 50 percent;definitive treatment with chest tube drainageand sclerosis is recommended. Failure to ade-quately manage a malignant pleural effusioncan result in a trapped lung, with permanentdyspnea, cough, and pain.

The purpose of chest tube placement andsuction drainage is to empty the pleural spaceto permit approximation of the visceral andparietal pleura. When chest tube output is min-imal, any of a variety of topical irritants isinstilled and the patient repositioned every 15minutes for 2 hours to distribute the irritantthroughout the pleural space. The goal is to cre-ate adhesions between the irritated visceral andparietal pleura to prevent subsequent massivereaccumulation of fluid with atelectasis. Therehave been a variety of agents employed, includ-ing talc slurry, tetracycline, bleomycin, andother chemotherapeutic agents. In a random-ized trial comparing the first three agents, aninsufficient number of patients was accrued; inthe absence of a direct comparison, talc appearsto have the highest success rate.31

Pericardial effusions are not uncommon andmay eventually occur in up to 25 percent of allwomen with metastatic breast cancer.32 The pre-senting complaint is typically exertional dysp-nea. Chest radiography and resting arterial oxy-gen saturation may both be normal, requiringthis diagnosis to be specifically considered inthe dyspneic patient.33 As pericardial effusionsoccur not infrequently in conjunction withmalignant pleural effusions but go unrecognizedon chest radiography, pericardial effusionsshould also be considered whenever pleuraleffusions are diagnosed. Physical exam mayshow tachycardia, an absent precordial cardiacimpulse, a pericardial friction rub, atrial fibrilla-tion, and pulsus paradoxus. Electrocardiogramwill show decreased voltages in the precordialleads. Definitive diagnosis of pericardial effu-sion requires echocardiography, which may alsodemonstrate cardiac tamponade with diastoliccollapse of the right atrium and ventricle.34

Patients with symptomatic or hemodynami-cally significant pericardial effusions shouldundergo immediate drainage. Immediate catheterdrainage can prevent cardiovascular collapse inpatients with tamponade but does not providedefinitive treatment. The creation of a subxiphoidpericardial window is a relatively simple surgical

274 BREAST CANCER

solution with a high success rate.35 Open thora-cotomy with pericardial stripping has a muchhigher morbidity and is required only for rarepatients with constrictive pericarditis.

Malignant ascites can develop as a manifes-tation of peritoneal metastases, occurring morefrequently in patients with infiltrating lobularcarcinoma. Symptoms include bloating, disten-sion, early satiety, and shortness of breath. Bothultrasound and CT scan can demonstrateascites, with the latter also revealing peritonealstudding or omental thickening in somepatients. While the most satisfactory control ofmalignant ascites is achieved with effectivesystemic therapy, this is often not possiblewhere ascites occurs as a late complication ofadvanced disease. Therapeutic paracentesismay provide transient relief of symptoms.Repeated drainage of several liters of asciticfluid may result in hypotension or hypoalbu-

minemia, however. Diuretics are seldom help-ful in managing malignant ascites.

BONE METASTASES

Bone is the most frequent site of metastaticspread, with autopsy series revealing skeletalinvolvement in 85 percent of all breast cancerpatients. The axial skeleton is most commonlyaffected, for example, the pelvis, spine, ribs,skull, and proximal long bones (Figure 17–1).36

Constant, dull, progressive pain is the usual pre-sentation, and any such complaint should beinvestigated radiographically, particularly if thepain is unrelieved by rest. Plain radiographsmost often show lytic lesions, although 15 per-cent of breast carcinomas are associated withblastic lesions and both patterns may be evi-dent.37 Technetium bone scans are more sensi-tive but less specific than plain films, as bone


Figure 17–1. A bone scan with multiple focal areas of increase radionuclide uptake (e.g., spine, skull)responding over time to hormonal therapy.

loss of up to one-third may occur before becom-ing visible. The first discovery of a bone metas-tasis in a patient should prompt a bone scan todetermine the extent of disease. Unfortunately,available serum markers of osteoblastic activitysuch as alkaline phosphatase or tumor markerssuch as CA15-3 antigen are insufficiently sensi-tive to rule out bone metastases in a patient withbony pain.37

External beam radiation therapy is the main-stay of palliating the pain of bone metastasesand will provide at least some relief in 90 per-cent of patients.38 Relief may be experiencedearly on as a result of decreasing periosteousinflammation, with maximal palliation within 3weeks. The main side effect of radiation ther-apy to the axial skeleton is myelosuppression,which may be cumulative and prolonged whenlarge fields or multiple sites are treated. As thismay preclude effective dosing of chemother-apy, radiation should be reserved for sites ofsevere or dominant symptoms, to prevent frac-ture of long bones, or in patients unsuitable for,or unlikely to respond to, chemotherapy.

Surgical stabilization is required in patientswith impending fractures of the femur or foroccasional patients with extensive and painfulhumeral lesions. The proximal femur is at par-ticular risk of pathologic fracture due to the highmechanical stresses of ambulation; an aggres-sive approach to prevention is appropriate dueto the catastrophic effects of this complication.Prediction of an impending fracture is notentirely accurate, but the usual criteria for pro-phylactic stabilization are cortical destruction of> 50 percent or proximal lesions > 1 inch.39

Lytic lesions are more prone to fracture than isblastic disease. Depending on location, a varietyof orthopedic approaches may be required; theuse of methylmethacrylate cement permits earlyreambulation.40 Following fixation, adjuvantradiotherapy to the involved bone is usuallyindicated to prevent progressive destruction ofbone, with resulting destabilization.

Recently, medical therapy has been able tomore directly address the pathophysiology of

bone metastases. Bisphosphonates are a classof drugs related to pyrophosphate that bind tohydroxyapatite crystals, stabilizing bone andinhibiting reabsorption. The bisphosphonatepamidronate, administered intravenously, hasbeen shown to produce sclerosis or stabilizationof lytic metastases in 50 percent of breast can-cer patients.41 This results in a decrease in therate of pathologic fractures, in the need forradiotherapy, and in the use of analgesics,42

suggesting an important palliative role forpamidronate. Similar trials using orallyabsorbed bisphosphonates such as alendronatehave also been completed.

Strontium 89 is an emitter of b-radiationthat is taken up by sites of active bone destruc-tion, and can be administered intravenously.Improvement in bone pain is reported by 80percent of patients,43 and toxicity is largely lim-ited to myelosuppression. Indications includewidespread, painful bony metastases in patientswho will not be candidates for chemotherapyand recurrent pain in sites already treated byexternal beam radiotherapy.

LOCAL RECURRENCE

Recurrence of cancer within a breast afterlumpectomy and radiation therapy has differentimplications from a recurrence involving theskin or chest wall following mastectomy.44

Treatment involves mastectomy and often acourse of systemic “pseudoadjuvant” chemo-therapy or change in hormonal adjuvant ther-apy. No randomized trials have addressed thisissue, however.

Initial recurrence in the skin overlying amastectomy site is associated with synchronouspresentation of distant metastatic disease inone-third of patients. Discovery of local recur-rence should therefore prompt restaging with abone scan and CT scan of chest and liver.45 Ifdistant metastasis are not discovered, the skinrecurrence, usually in the form of one or sev-eral dermal or subdermal nodules, shouldreceive local treatment. Complete excision of

276 BREAST CANCER

isolated, small nodules should be attempted.Wide excision with partial or full-thicknesschest wall resection are seldom indicated, how-ever, due to morbidity and a 50 percent failurerate.46 Rather, radiation therapy delivered to theentire chest wall at a dose of 45 to 50 cGy, witha boost to the site of recurrence, should be con-sidered standard therapy. This will yield 5-yearlocal control rates of 85 percent if the tumor isfirst excised or 63 percent if radiation is givenwithout excision.47

Nearly all patients with isolated local recur-rence will subsequently develop distant metas-tases. Only 30 percent remain free from distantmetastases after 5 years, with the disease-freeinterval between mastectomy and skin recur-rence the most important predictive factor.Only 20 percent of those suffering local recur-rence within 2 years will be free of distant dis-ease 3 years later, compared to 36 percent ofthose who first recurred > 2 years after mastec-tomy.48 This suggests that most patients couldpotentially benefit from systemic therapy fol-lowing local recurrence. The benefits of postre-currence hormonal therapy with tamoxifenhave been established in a randomized trial forpatients with estrogen receptor-positivetumors.49 Unfortunately, similar benefits havenot been established for chemotherapy inreceptor-negative patients.

SYSTEMIC THERAPY

When metastatic breast cancer presents clini-cally as an isolated, symptomatic site, specificpalliative measures, discussed above, are indi-cated. Many patients, however, present withvisceral or multi-site disease, or will be foundto have additional metastases upon restaging.While widely metastatic breast cancer is incur-able, lessening of the symptomatic burden andprolongation of survival are possible for mostwomen through judicious use of systemic hor-mone therapy and chemotherapy. It is importantto understand that the goal of systemic therapyis ultimately palliation, so that every decision

must involve weighing the potential improve-ment in quality of life against the expected tox-icities of treatment.

HORMONAL THERAPY

In this context, hormonal therapy is almostalways preferred as initial therapy for womenwith estrogen receptor-positive metastatic breastcancer. Whenever possible, and particularly forwomen who relapse after adjuvant hormonaltherapy, this decision should be based on recep-tor assays performed on a biopsy of the metasta-tic disease. While in the absence of interveningtherapy the receptor status of metastatic diseaseis predicted by that of the primary tumor, estro-gen receptor becomes negative in one-third ofpatients and progesterone receptor becomes neg-ative in one-half of patients who receive tamox-ifen in the interval before relapse.50

There will be a complete or partial responseobtained by initial hormonal therapy in overthree-quarters of women with estrogen receptorand progesterone receptor-positive metastaticdisease and no prior therapy.51 If the estrogenreceptor assay is negative, the response ratedrops to less than half, and to one-third if theprogesterone receptor assay is negative. If nei-ther receptor is detected, response is seen in < 10percent of patients; in this circumstance,chemotherapy is often a preferable option. Hor-monal therapy is also contraindicated in womenwith lymphangitic carcinomatosis or extensivemetastases to the liver, due to the need for a rapidresponse. Finally, if biopsy of a metastatic site isnot possible, the decision to employ hormonetherapy can be based on those clinical criteria(eg, a long disease-free interval, disease limitedto bone or soft tissue, and elderly patients) asso-ciated with the receptor-positive phenotype.

Estrogen Antagonists

The likelihood of response to initial hormonaltherapy is similar for several classes of drugs,and so initial treatment can often be selected


based on their side-effect profiles. In previ-ously untreated patients, or for those severalyears removed from adjuvant hormonal ther-apy, competitive inhibitors of estrogen bindingare usually the first choice.

The oldest and most widely prescribed estro-gen antagonist is tamoxifen,52 but toremifenehas also been approved for this indication.Raloxifene is currently marketed for preventionof osteoporosis and may also have some effi-cacy against metastatic breast cancer, althoughfurther study is clearly required.53 Tamoxifenappears to be effective for both pre- and post-menopausal women with advanced, receptor-positive disease.54 Common side effects oftamoxifen include hot flashes (particularly inperimenopausal women), disruption of men-strual cycles, and vaginal dryness or dis-charge.55 In addition, weight gain and mild fluidretention are frequent, with nocturnal legcramps not uncommonly reported. Patients withbone metastasis may suffer a syndrome of“tumor flare,” typically 7 to 10 days after initia-tion of tamoxifen. This is seen in 1 to 3 percentof patients and consists of increased pain at sitesof metastases; it may lead to hypercalcemia. Asthis is predictive of subsequent response totamoxifen, therapy should be continued, withsupportive measures as needed. Approximately1 percent of healthy patients on tamoxifen willdevelop deep-vein thrombosis, although womenwith metastatic breast cancer also have anincreased incidence of thromboembolic dis-ease.56 Other, rare complications such ascataract formation or an increased incidence ofendometrial cancer are seldom of concern towomen with metastatic disease.

The average duration of response to initialhormone therapy is approximately 1 year.Women whose disease stabilizes on tamoxifenappear to do as well as those achieving objec-tive remissions. While responses lasting foryears are not uncommon (particularly if therehas been a long disease-free interval), eventu-ally most tumors will develop resistance totamoxifen, leading to clinical progression. This

may occur due to outgrowth of receptor-negative clones within a heterogenous popula-tion or to acquired, specific resistance to estro-gen antagonists. Once a responding tumorprogresses on tamoxifen, other agents in thisclass have little activity. Indeed, a fraction ofsuch patients will briefly improve when tamox-ifen is withdrawn, suggesting that changes inthe receptor or the cellular estrogen-responsemachinery has led to the drug behaving as anestrogen agonist.57

Aromatase Inhibitors

About half of women who initially respond totamoxifen will also respond to second-line hor-monal therapy. Randomized trials have sug-gested somewhat greater efficacy, lesser sideeffects, and perhaps slight improvement in sur-vival when specific aromatase inhibitors arecompared to oral progestins in this setting.58,59

The new generation of aromatase inhibitors—anastrozole, letrozole, and others not yetapproved for use—have replaced the older drugaminoglutethimide due to much improvedsafety profiles. Anastrozole and letrazole workby binding competitively to the porphyrinnucleus of the aromatase enzyme, which isresponsible for estrogen production fromandrostenedione. This extraovarian pathway isimportant only in postmenopausal women,therefore anastrozole and letrazole should beused only after menopause. The most commonside effects seen with these drugs are headacheand mild nausea. Prior to the development ofthese agents, aminoglutethimide had beenemployed as an aromatase inhibitor, but hasnow fallen into disuse because of its high fre-quency of unacceptable side effects, includingrash, lethargy, and ataxia.

Progestins

Before the development of the newer aromataseinhibitors, second-line hormonal therapy formost women consisted of progestins, usuallyoral megestrol acetate or parenteral medrox-

278 BREAST CANCER

yprogesterone acetate.60,61 Up to half of womenreceiving these drugs will respond withimprovement or stabilization of their disease.Unlike other hormonal agents, there is evidenceof a dose-response effect with progestins,although their mechanism of action is unknown.These drugs also produce an increased sense ofwell-being, improved appetite, and suppress hotflashes. Unfortunately, the side effects ofchronic weight gain, fluid retention, and dysp-nea make them unacceptable to many.

Ovarian Ablation

For premenopausal women with receptor-posi-tive disease, medical or surgical castration is alsoan effective approach to hormonal therapy. Theendocrinologic effect of castration is achieved bytwo analogs of gonadotropin-releasing hormone,goserelin and leuprolide, which suppress follicle-stimulating hormone and luteinizing hormone,and thus estrogen production by the ovary.62,63

Either agent will achieve the same benefit asoophorectomy, that is, a 45 percent likelihood ofdisease regression or stabilization, but requireparenteral administration on a monthly or tri-monthly basis. Side effects are limited to pain atthe injection site and menopausal symptomssuch as hot flashes, mood swings, and dry skin.Once disease progresses after either medical orsurgical castration, the alternate approach haslittle chance of benefit. Obviously, castration byeither technique can only be of benefit to pre-menopausal patients, where the ovary is the pri-mary site of estrogen production.

There have been a number of studies thathave attempted to combine hormonal agents formore effective control of metastatic disease.64 Ingeneral, a small increase in response rate is seenwith combinations, but time-to-progression isnot improved over the use of the same agentsemployed sequentially and there is clearly nosurvival advantage. Additional toxicity is oftenreported when hormonal agents are used incombination; since the goal of all such therapyis palliative, this approach is not recommended.

CHEMOTHERAPY

Systemic chemotherapy is often indicated tocontrol disseminated breast cancer and relievesymptoms. While prolonged remissions may beachieved, there is no evidence that metastaticbreast cancer can be cured by chemotherapy.Thus, the ultimate goals are again palliative,and the toxicity of chemotherapy must be care-fully weighed against a realistic appraisal ofbenefits. With this caveat, chemotherapy iscommonly indicated as frontline therapy formetastases to liver or lung, those arising fromestrogen receptor-negative tumors, and thosethat fail to respond to initial or subsequent hor-monal treatments.

There are a wide variety of chemothera-peutic agents that show some activity againstmetastatic breast cancer. Response rates areaffected by site of disease, with soft tissuemetastases typically most responsive, andliver metastases least responsive, to manyagents. Prior treatment history has a majoreffect on the likelihood of response, due to thephenomenum of pleiotropic drug resistance,which occurs when cancer cells undergoingtreatment become resistant not only to thatparticular agent but also to unrelated classesof cytotoxic drugs. Attempts to overcome drugresistance have included the use of chemother-apeutic agents in combination and at increaseddose intensity. These approaches have resultedin higher response rates, but the average dura-tion of response to initial chemotherapyremains < 1 year. High dose chemotherapywith bone marrow or stem cell rescue has notdemonstrated any survival advantage whencompared to conventional regimens in strictlyrandomized prospective trials and shouldremain investigational. The use of alternativeagents after progression of disease is markedby lower response rates and shorter durationsof response, so much so that patients rarelybenefit from more than three sequentialchemotherapy regimens. The major classes ofuseful cytotoxic agents are reviewed below.


Anthracyclines

Doxorubicin has long been considered thebenchmark drug for treatment of metastaticbreast cancer, with a single-agent response rateof 40 to 50 percent.65 Increases in the dose ofdoxorubicin, sometimes given with granulocytecolony-stimulating factor support, can yieldresponse rates as high as 80 percent, but at theprice of increasing toxicity.66 As this higherresponse rate does not result in any noticeableimprovement in survival, dose-intense schedulescannot be recommended at present. Commontoxicities include moderate nausea, mucositis,neutropenia, and a cumulative dose-related riskof congestive cardiomyopathy. Despite thesetoxicities, doxorubicin, often given in combina-tion, has become the standard frontlinechemotherapy for metastatic breast cancer. Sev-eral randomized trials have established that dox-orubicin-containing combinations are superiorto similar regimens lacking an anthracycline.

Mitoxantrone is a potentially less toxic deriv-ative of doxorubicin that is also widely used forpalliative treatment of metastatic disease. It isclearly less emetogenic and appears to have lesscardiotoxicity than the parent compound,although cardiotoxicity is cumulative and addi-tive to that induced by prior doxorubicin expo-sure.67 In direct comparison to doxorubicin,either alone or in combination, response rateswere lower for mitoxantrone, although overallsurvival was not compromised.68 The most effec-tive use of mitoxantrone may be in combinationwith 5-fluorouracil and folinic acid, which isassociated with a response rate of up to 65 per-cent and quite manageable toxicity.69 Anotherapproach to lessening the toxicity of doxorubicinis to encapsulate the drug in lipid liposomes.70

This strategy permits more selective tissueuptake, resulting in less nausea, neutropenia, andcardiotoxicity, although various cutaneous reac-tions are seen with the currently available formu-lation. Studies using liposomal doxorubicin arestill in progress and the drug is not currentlyapproved for treatment of breast cancer.

Taxanes

The complex, semisynthetic paclitaxel and thesynthetic docetaxol have recently establishedthemselves to be of equal or greater single-agent efficacy than doxorubicin and maintainsignificant activity in patients previouslytreated with doxorubicin.71,72 Paclitaxel admin-istered by 24-hour infusion has achievedresponse rates as high as 60 percent, but theoptimum dose and schedule for this drug havenot been established. Toxicities include neu-tropenia, a delayed arthralgias/myalgia syn-drome occurring 48 hours after administration,and a peripheral neuropathy with higher cumu-lative doses. Bradycardia is observed but is sel-dom clinically significant. Frequent type Ihypersensitivity reactions require premedica-tion with steroids and antihistamines. Thesetoxicities are lessened by administration on aweekly rather than triweekly schedule.

Docetaxol has recently been introduced fortreatment of metastatic breast cancer, where ithas demonstrated a higher response rate andmore durable remissions than doxorubicin.72

Toxicity consists of neutropenia and a capil-lary leak syndrome, resulting in peripheraledema and pleural or pericardial effusions,which are preventable with a 3-day course ofcorticosteroids.73 These taxanes have beencombined by a number of investigators, butthus far both activity and toxicity appears tobe additive rather than synergistic. An excep-tion is the combination of doxorubicin andpaclitaxel, which appears to produce a veryhigh response rate but results in cardiotoxicityat lower than expected cumulative doxoru-bicin doses.74

Alkylating Agents

Cyclophosphamide has reasonable single-agentactivity but is usually used in combination withother agents such as doxorubicin or methotrexateand fluorouracil. Toxicity is limited at conven-tional doses to neutropenia, moderate nausea,

280 BREAST CANCER

mucositis, and occasional hemorrhagic cystitis.Ifosfamide, an analog of cyclophosphamide,appears to have similar efficacy, and is notentirely cross-resistant to cyclophosphamide. Inaddition to neutropenia, toxicity includes fre-quent hemorrhagic cystitis (requiring the use ofa urothelial protective agent), interstitial nephri-tis, and temporary encephalopathy. These toxic-ities have limited the use of ifosfamide.

Antimetabolites

Five-fluorouracil, a pyrimidine analog thatbinds to thymidylate synthase, is widelyemployed in the treatment of breast cancer.Intermittent bolus administration, as is found inmany classic cytotoxic combinations, is prob-ably the least effective schedule for this cell-cycle active agent, given its short half-life.Continuous infusion of 5-fluorouracil will yieldresponses in even heavily pretreated patients,with manageable toxicity consisting largely ofpalmar/planter dermatitis.75 Alternatively, theintracellular binding of fluorouracil tothymidylate synthase can be stabilized by coad-ministration of folinic acid. Again, higherresponse rates are seen, although neutropenia,mucositis, and diarrhea become significant.76

Methotrexate, a folic acid analog, has a lowsingle-agent response rate in metastatic breastcancer but is occasionally useful, primarily toprovide biochemical synergy with fluorouracil.

Vinca Alkaloids

Whereas vincristine has little activity againstbreast cancer, vinblastine yields response ratesas high as 37 percent when given by 120-hourinfusion.77 Toxicity is limited to myelosuppres-sion but the schedule is inconvenient. Vinblas-tine was given as a bolus in many earlier com-binations but probably added little. Vinorelbine,a newer vinea derivative, yields a single-agentresponse rate of 35 to 40 percent when given asa weekly bolus.78 Toxicity consists of neutrope-nia, peripheral neuropathy, myalgias, and short-lived pain at sites of metastatic disease.

Other Agents

Cisplatin is an active single agent in previouslyuntreated metastatic breast cancer, but dosageis inconvenient due to the need for prolongedhydration to prevent nephrotoxicity; also, thedrug has a poor response rate in previouslytreated patients. Gemcitabine may be non-crossresistant with anthracyclines and taxanes,with toxicity limited to fatigue and mildmyelosuppression, suggesting that this drugmay find a role in previously treated patients.79

Further study is required, however, and neitherthis drug nor cisplatin has been approved fortreatment of breast cancer. Finally, capecitabine,an oral drug, has recently been approved foruse in previously treated patients with metasta-tic breast cancer. Once absorbed, this drug isconverted to fluorouracil, explaining its similartoxicity spectrum.

The higher response rates seen with combi-nation chemotherapies often justify their initialuse over single agents. After progression onfrontline treatment, however, the toxicity ofmultiagent therapy may make adequate dosingimpossible and obviate any advantage seen withthis approach. Thus, the sequential use of singleagents, especially after initial treatment, mayprovide a higher quality of life, equal palliativebenefit, and no compromise of overall survival.

In patients who achieve a complete responseor whose disease stabilizes after a partialresponse, the question of duration of chemo-therapy arises. Several small studies have sug-gested that the time-to-treatment failure isextended by several months and that quality oflife is improved by continuing with maintenancetherapy once a response is achieved, rather thanwithholding further therapy until relapse.80

IMMUNOTHERAPY

Earlier studies in which the nonspecificimmunostimulants bacille Calmette-Guerin orlevamisole were added to chemotherapy showedno advantage to this procedure. In the last 2


years, however, studies using specificimmunotherapy with the humanized murinemonoclonal antibody trastuzumab have yieldedpromising results. This antibody recognizes andbinds to a transmembrane tyrosine kinase codedfor by the c-erbB2 or HER2 gene, which isamplified and/or overexpressed in up to one-third of all breast cancer specimens (Figure17–2). When given to a heavily pretreated groupof patients whose tumors overexpressed thisgene product, trastuzumab produced a 16 per-cent objective response rate.81 Toxicity was min-imal, consisting of fever and chills after the firstweekly infusion, and mild pain, asthenia, nausea,diarrhea, and dyspnea. In addition, 5 percent ofpatients had evidence of cardiac dysfunction.

Trastuzumab was also studied in a placebo-controlled study involving women simultane-ously receiving chemotherapy with either dox-orubicin-cyclophosphamide or paclitaxel.82

When added to doxorubicin-cyclophosphamide,the response rate increased from 43 to 52 percent,with a 3.6-month prolongation of responses.When trastuzumab was added to paclitaxel inpatients who had prior exposure to doxorubicin,response rate increased from 16 to 42 percent,and duration of response from 4.4 to 11 months.Unfortunately, cardiotoxicity was seen in 27 per-cent of women receiving the doxorubicin combi-

nation plus antibody treatment and in 12 percentreceiving paclitaxel plus antibody. The mecha-nism of this synergy, manifest both in increasedresponse rates and cardiotoxicity, is not yetunderstood, and a wide variety of trials havebegun to define the role of trastuzumab in thetreatment of metastatic breast cancer.

CONCLUSION

Metastatic breast cancer is responsible for over40,000 deaths of American women each year,with most of these women having lived withmetastatic disease for 2 or more years prior totheir death. During this time, many symptomscan be palliated or avoided by addressing bothlocal problematic sites with surgery or radia-tion therapy and the overall course of the dis-ease with hormonal therapy or chemotherapy.At all times, the impact of a therapeutic inter-vention on quality of life must be weighed, asmany options, particularly second- or third-linetherapies, offer little chance of prolonging life.

Improvements in breast cancer prevention,early detection, and postsurgical adjuvant ther-apy are likely to reduce the overall mortality ofbreast cancer by reducing the number of womenwho suffer metastatic recurrence. Those womenwho nonetheless are forced to contend withmetastases will have an increasing number ofoptions in coming years. Newer chemothera-peutic drugs, hormonal agents, and immuno-logic approaches offer the hope of more selec-tive, less toxic, and ultimately more effectivetreatments for metastatic breast cancer.

REFERENCES

1. Early Breast Cancer Trialists’ CollaborativeGroup. Tamoxifen for early breast cancer: anoverview of the randomised trials. Lancet1998;351:1451–67.

2. Landis S, Murray T, Bolden S, et al. Cancer sta-tistics, 1999. CA Cancer J Clin 1999;49:8–31.

3. Valagussa P, Bonadonna G, Veronesi U. Patternsof relapse and survival following radical mas-tectomy. Cancer 1978;41:1170.

4. Borst MJ, Ingold JA. Metastatic patterns of inva-

282 BREAST CANCER

Figure 17–2. A breast cancer showing overexpression ofmembrane associated HER-2 protein. Metastatic breast can-cer with this phenotype demonstrates a 16% response rateto trastuzumab when employed as a single agent.

sive lobular versus invasive ductal carcinomaof the breast. Surgery 1993;114:637.

5. Harris M, Howell A, Chrissohou M, et al. A com-parison of the metastatic pattern of infiltratinglobular carcinoma and infiltrating duct carci-noma of the breast. Br J Cancer 1984;50:23.

6. Lamovec J. Metastatic pattern of infiltrating lobu-lar carcinoma of the breast: an autopsy study. JSurg Oncol 1991;48:28.

7. Van De Vijver M, Mooi W, Wisman P, et al.Immunohistochemical detection of the neuprotein in tissue sections of human breasttumors with amplified neu DNA. Oncogene1988;2:175–8.

8. Rouesse J, Friedman S, Guash-Jordan I, et al. Sur-vival effect of systemic therapy on patientsdeveloping metastatic breast carcinoma. BreastCancer Res Treat 1990;15:13.

9. Perez JE, Machiavelli M, Leone BA, et al. Bone-only versus visceral-only metastatic pattern inbreast cancer: analysis of 150 patients. Am JClin Oncol 1990;13:294.

10. Blumenschein GR, Pinnamaneni K, Buzdar AU.Combined regional and systemic therapy inbreast cancer patients with an isolated metasta-sis with or without prior chemotherapy. In:Jones SE, Salmon SE, editors. Adjuvant ther-apy of cancer. IV. Orlando: Grune and Stratton;1984. p. 311.

11. Vogel C, Azevedo S, Hilsenbeck S, et al. Survivalafter first recurrence of breast cancer. Cancer1992;70:129–35.

12. Clark G, Sledge G, Osborne C, et al. Survivalfrom first recurrence: relative importance ofprognostic factors in 1,015 breast cancerpatients. J Clin Oncol 1987;5:55–61.

13. Blanco G, Holli K, Heikkinen O, et al. Prognosticfactors in recurrent breast cancer: relationshipsto site of recurrence, disease-free interval,female sex steroid receptors, ploidy and histo-logical malignancy grading. Br J Cancer1990;62:142–6.

14. Qazi R, Chuang JL, Drobyski W. Estrogen recep-tors and the pattern of relapse in breast cancer.Arch Intern Med 1984;144:2365–7.

15. DiStefano A, Yap HY, Hortobagyi GN, et al. Thenatural history of breast cancer patients withbrain metastases. Cancer 1979;44:1913.

16. Posner JB. Neurologic complications of systemiccancer. Med Clin North Am 1979;63:783.

17. Davis PC, Hudgins PA, Peterman SB, et al. Diag-nosis of cerebral metastases: double-dosedelayed CT vs contrast-enhanced MR imaging.AJNR Am J Neuroradiol 1991;12:293.

18. Galicich JH, French LA, Melby J. Use of dexa-methasone in treatment of cerebral edema asso-ciated with brain tumors. Lancet 1961;1:46.

19. Cohen N, Strauss G, Lew R, et al. Should prophy-lactic anticonvulsants be administered topatients with newly diagnosed cerebral metas-tases? a retrospective analysis. J Clin Oncol1988;6:1621.

20. Patchell R, Tibbs P, Walsh J, et al. A randomizedtrial of surgery in the treatment of singlemetastases to the brain. N Engl J Med 1990;322:494–500.

21. Borgelt B, Gelber R, Kramer S, et al. The palliationof brain metastases: final results of the first oftwo studies by the Radiation Therapy OncologyGroup. Int J Radiat Oncol Biol Phys 1980;6:1.

22. Flickinger JC, Kondziolka D, Lansford LD, et al. Amulti-institutional experience with stereotacticradiosurgery for solitary brain metastases. Int JRadiat Oncol Biol Phys 1994;28:979.

23. Rosner D, Taukuma N, Lane W. Chemotherapyinduces regression of brain metastases inbreast carcinoma. Cancer 1986;58:832.

24. Colomer R, Cosos D, Del Campo JM, et al. Brainmetastases from breast cancer may respond toendocrine therapy. Breast Cancer Res Treat1988;12:83.

25. Tsukada Y, Fouad A, Pickren JW, et al. Centralnervous system metastasis from breast carci-noma: autopsy study. Cancer 1983;52:2349.

26. Ongerboer de Visser BW, Somers R, Nooyen WH,et al. Intraventricular methotrexate therapy ofleptomeningeal metastasis from breast carci-noma. Neurology 1983;33:1565.

27. Wasserstrom WR, Glass JP, Posner JB. Diagnosisand treatment of leptomeningeal metastasesfrom solid tumors: experience with 90 patients.Cancer 1982;49:759.

28. Sze G, Soletsky S, Bronen R, et al. MR imagingof the cranial meninges with emphasis on con-trast enhancement and meningeal carcinomato-sis. AJNR Am J Neuroradiol 1989;10:965.

29. Yap HY, Yap BS, Rasmussen S, et al. Treatmentfor meningeal carcinomatosis in breast cancer.Cancer 1982;49:219.

30. Raju RN, Kardinal CG. Pleural effusion in breastcarcinoma: analysis of 122 cases. Cancer 1981;48:2524.

31. Hausheer FH, Yarbro JW. Diagnosis and treatmentof malignant pleural effusion. Semin Oncol1985;12:54.

32. Hagemeister FB, Buzdar AU, Luna MA, et al.Causes of death in breast cancer. Cancer 1980;46:162.


33. Cham WC, Freiman AH, Carstens PHB, et al.Radiation therapy of cardiac and pericardialmetastases. Ther Radiol 1975;114:701.

34. Gillam LD, Guyer DE, Gibson TC, et al. Hydro-dynamic compression of the right atrium: anew echocardiographic sign of cardiac tam-ponade. Circulation 1983;68:294.

35. Vaitkus PT, Herrmann HC, LeWinter MM. Treat-ment of malignant pericardial effusion. JAMA1994;272:59.

36. Tubiana-Hulin M. Incidence, prevalence and dis-tribution of bone metastases. Bone 1991;12Suppl 1:S9

37. O’Brien DP, Horgan PG, Gough DB, et al. CA 15-3: a reliable indicator of metastatic bone dis-ease in breast cancer patients. Ann R Coll Surg1992;74:9.

38. Poulsen HS, Nielsen OS, Klee M, et al. Palliativeirradiation of bone metastases. Cancer TreatRev 1989;16:41.

39. Harrington KD. Orthopaedic management ofmetastatic bone disease. St Louis: CV Mosby;1988. p. 7.

40. Yazawa Y, Frassica FJ, Chao EYS, et al. Metasta-tic bone disease: a study of the surgical treat-ment of 166 pathologic humeral and femoralfractures. Clin Orthop 1990;251:213.

41. Morton AR, Cantrill JA, Pillai GV, et al. Sclerosisof lytic bone metastases after disodium amino-hydroxypropylidene bisphosphonate (APD) inpatients with breast carcinoma. BMJ 1988;297:772.

42. van Holten-Verzantvoort ATM, Kroon HM,Bijvoet OLM, et al. Palliative pamidronatetreatment in patients with bone metastasesfrom breast cancer. J Clin Oncol 1993;11:491.

43. Robinson RG, Blake GM, Preston DF, et al. Stron-trium-89: treatment results and kinetics inpatients with painful metastatic prostate andbreast cancer in bone. Radiographic 1989;9:271.

44. Recht A, Schnitt SJ, Connolly JL, et al. Prognosisfollowing local or regional recurrence afterconservative surgery and radiotherapy for earlystage breast carcinoma. Int J Radiat Oncol BiolPhys 1989;16:3.

45. Andry G, Suciu S, Vico P, et al. Locoregionalrecurrences after 649 modified radical mastec-tomies: incidence and significance. Eur J SurgOncol 1989;15:476.

46. Dahlstrom KK, Andersson AP, Andersen M, et al.Wide local excision of recurrent breast cancerin the thoracic wall. Cancer 1993;72:774.

47. Kenda R, Lozza L, Zucali R. Results of irradiationin the treatment of chest wall recurrent breast

cancer [abstract]. Radiother Oncol 1992;24Suppl:S41.

48. Aberizk WJ, Silver B, Henderson IC, et al. Theuse of radiotherapy for treatment of isolatedlocoregional recurrence of breast carcinomaafter mastectomy. Cancer 1986;58:1214.

49. Borner M, Bacchi M, Goldhirsch, et al. First iso-lated locoregional recurrence following mas-tectomy for breast cancer: results of a phase IIImulticenter trial comparing systemic treatmentwith observation after excision and radiation. JClin Oncol 1994;12:2071.

50. Hull D, Clark G, Osborne K, et al. Multiple estro-gen receptor assays in human breast cancer.Cancer Res 1983;43:413–6.

51. Wittliff JL. Steroid-hormone receptors in breastcancer. Cancer 1984;53:630.

52. Jaiyesimi IA, Buzdar AU, Decker DA, HortobagyiGN. Use of tamoxifen for breast cancer: twentyeight years later. J Clin Oncol 1995;13:513–29.

53. Gradishar WJ, Jordan VC. Clinical potential of newantiestrogens. J Clin Oncol 1997;15:840–52.

54. Sunderland MC, Osborne CK. Tamoxifen in pre-menopausal patients with metastatic breast can-cer: a review. J Clin Oncol 1986;9:1283–97.

55. Fisher B, Costantino J, Wickerham D, et al.Tamoxifen for prevention of breast cancer:report of the national surgical adjuvant breastand bowel project p-1 study. J Natl Cancer Inst1998;90:1371–88.

56. McDonald CC, Alexander FE, Whyte BW, et al.Cardiac and vascular morbidity in womenreceiving adjuvant tamoxifen for breast cancerin a randomized trial. The Scottish Cancer Tri-als Breast Group. BMJ 1995;311:977–80.

57. Howell A, Dodwell DJ, Anderson H, Redford J.Response after withdrawal of tamoxifen andprogestins in advanced breast cancer. AnnOncol 1992;3:611–7.

58. Dombernowsky P, Smith I, Falkson G, et al. Letro-zole, a new oral aromatase inhibitor foradvanced breast cancer: double-blind random-ized trial showing a dose effect and improvedefficacy and tolerability compared with mege-strol acetate. J Clin Oncol 1998;16:453–61.

59. Buzdar A, Jonat W, Howell A, et al. Anastrozoleversus megestrol acetate in the treatment ofpostmenopausal women with advanced breastcarcinoma. Cancer 1998;83:1142–52.

60. Muss H, Case D, Cappizzi RL, et al. High versusstandard dose megesterol acetate in womenwith advanced breast cancer: a phase III trial ofthe Piedmont Oncology Association. J ClinOncol 1990;8:1797–1805.

284 BREAST CANCER

61. Cavalli F, Goldhirsch A, Jungi F, et al. Randomizedtrial of low versus high dose medroxyproges-terone acetate in the induction treatment ofpostmenopausal patients with advanced breastcancer. J Clin Oncol 1984;2:414–9.

62. Taylor CW, Green S, Dalton WS, et al. Multicenterrandomized clinical trial of goserelin versus sur-gical ovariectomy in premenopausal patients withreceptor-positive metastatic breast cancer: anintergroup study. J Clin Oncol 1998;16:994–9.

63. Harvey HA, Lipton A, Max DT, et al. Medicalcastration produced by the GnRH analogueleuprolide to treat metastatic breast cancer. JClin Oncol 1985;3:1068–72.

64. Ingle JN, Green SJ, Ahmann DL, et al. Random-ized trial of tamoxifen alone or combined withaminoglutethimide and hydrocortisone inwomen with metastatic breast cancer. J ClinOncol 1986;4:958–64.

65. Taylor SG, Gelber RD. Experience of the EasternCooperative Oncology Group with doxoru-bicin as a single agent in patients with previ-ously untreated breast cancer. Cancer TreatRep 1982;66:1594.

66. Bronchud MH, Howell A, Crowther D, et al. Theuse of granulocyte colony-stimulating factor toincrease the intensity of treatment with dox-orubicin in patients with advanced breast andovarian cancer. Br J Cancer 1989;60:121.

67. Myers CE, Chabner BA. Anthracyclines. In:Chabner B, Collins JM, editors. Cancerchemotherapy: principles and practice.Philadelphia: JB Lippincott; 1990. p. 356.

68. Cowan JD, Neidhart J, McClure S, et al. Random-ized trial of doxorubicin, bisantrene, andmitoxantrone in advanced breast cancer: aSouthwest Oncology Group study. J Natl Can-cer Inst 1991;83:1077.

69. Hainsworth JD, Andrews MG, Johnson DH, et al.Mitoxantrone, fluorouracil, and high-dose leu-covorin: an effective, well-tolerated regimenfor metastatic breast cancer. J Clin Oncol 1991;9:1731.

70. Lyass O, Uziely B, Heching NI, et al. Doxil inmetastatic breast cancer (MBC) after priorchemotherapy: therapeutic results in two con-secutive studies. Proc Am Soc Clin Oncol 1998.

71. Reichman BS, Seidman AD, Crown JPA, et al.Paclitaxel and recombinant human granulocytecolony-stimulating factor as initial chemother-apy for metastatic breast cancer. J Clin Oncol1993;11:1943–51.

72. Chan S, Friedrichs K, Noel D, et al. A randomizedphase III study of taxotere (T) versus doxoru-bicin (D) in patients (pts) with metastaticbreast cancer (MBC) who have failed an alky-lating containing regimen: preliminary results.ASCO 1997;540.

73. Riva A, Fumoleau P, Roche H, et al. Efficacy andsafety of different corticosteroid (C) premed-ications (P) in breast cancer (BC) patients(PTS) treated with taxotere (T). Proc Am SocClin Oncol 1997;16:188a.

74. Gianni L, Munzone E, Capri G, et al. Paclitaxel by3-hour infusion in combination with bolusdoxorubicin in women with untreated metasta-tic breast cancer: high antitumor efficacy andcardiac effects in a dose-finding and sequence-finding study. J Clin Oncol 1995;13:2688–99.

75. Huan S, Pazdur R, Singhakowinta A, et al. Low-dose continuous infusion 5-fluorouracil. Can-cer 1989;63:419–22.

76. Swain SM, Lippman ME, Egan EF, et al. Fluo-rouracil and high-dose leucovorin in previouslytreated patients with metastatic breast cancer. JClin Oncol 1989;7:890–9.

77. Fraschini G, Yap HY, Hortobagyi N, et al. Five-daycontinuous-infusion vinblastine in the treat-ment of breast cancer. Cancer 1985;56:225–9.

78. Romero A, Rabinovich MG, Vallejo CT, et al.Vinorelbine as first-line chemotherapy formetastatic breast carcinoma. J Clin Oncol1994;12(2):336–41.

79. Carmichael J, Possinger K, Phillip P, et al.Advanced breast cancer: a phase II trial withgemcitabine. J Clin Oncol 1995;13:2731–6.

80. Muss HB, Case LD, Richards F. Interrupted ver-sus continuous chemotherapy in patients withmetastatic breast cancer. N Engl J Med 1991;325:1342.

81. Cobleigh MA, Vogel CL, Tripathy D, et al. Effi-cacy and safety of herceptin (humanized anti-HER2 antibody) as a single agent in 222women with HER2 overexpression whorelapsed following chemotherapy for metasta-tic breast cancer. Proc Am Soc Clin Oncol1998;17:97a.

82. Slamon D, Leyland-Jones B, Shak S, et al. Additionof herceptin (humanized anti-HER2 antibody)to first line chemotherapy for HER2 overex-pressing metastatic breast cancer (HER2+/MBC) markedly increases anticancer activity: arandomized, multinational controlled phase IIItrial. Proc Am Soc Clin Oncol 1998b;17:98a.


Adenoid cystic carcinoma, 102–103Adenopathy, 81–82Adenosis, 52

calcifications of, 49Adjuvant therapy, 201–218, 253

benefits versus toxicity, 212–214candidates for, 202chemotherapy, 204–208complications, 213guidelines, 208, 214hormone receptor status, 212myelosuppression, 213patient age, 212principles of, 201regimens, 205side effects, 213–214trials, 204–209

Adriamycin, cardiac morbidity, 219Age of onset, 3, 5, 8Alcohol consumption, 31Alkylating agents, 280–281Aminoglutethimide, 210Anastrozole, 210Androstenedione, 24Aneuploidy, as biomarker, 23Angiogenesis, 226Angiosarcoma

postradiotherapy, 108–109primary, 107prognosis, 109

Anthracycline, 158–159, 206, 280Antiestrogens, 27–28Antimetabolites, 281Apocrine carcinoma, 101–102Architectural distortion, 43, 57, 84

progressive, 58Areolar thickening, 43Argyrophilic cytoplasmic granules,

104Aromatase inhibitors, 210, 278Ashkenazi Jews, 2, 5

BRCA1 mutation, 21risk, 2

Ataxia-telangiectasia, 3, 22

Atypical cells, 93pagetoid spread of, 93

Atypical hyperplasiaas biomarker, 23as indicator for mastectomy, 33, 85as risk factor, 19

“Atypical vascular lesions,” 108Augmented breast

evaluation of, 58–60Autologous reconstruction, 181–193

advantages, 182lateral thigh flaps, 191–192latissimus dorsi flap, 188–189Rubens flap, 189–190superior gluteal flap, 190–191TRAM flap, conventional,

182–185TRAM flap, free, 185–188

Axilla, management of, 146–148dissection of, 146–147side effects, 146–147staging, 147

Axillary adenopathy, 43Axillary irradiation, 159Axillary lymph nodes

metastases in, 201–202outcome, 202

Axillary node dissection, 157–159arguments against, 158

Bilateral reconstruction, 192–193mesh, 193

Biomarkersassessment, 30–31intermediate, 23, 91

Biopsyadequacy of excision, 97automated Tru-cut™ type, 71–72devices, 70–72fibroadenoma, 75fine-needle aspiration (FNA),70–71, 81–82frozen section, 97image-guided, 65–87

“long throw,” 71of microcalcifications, 74needle localization, 65–68open surgical, indication for, 74,

84–85scarring, 65specimen handling, 92, 97–98,

144specimen orientation, 144stereotactic equipment, 68–79stereotaxis and, 66techniques, 66–79vacuum-assisted, 75–77wire localization, 65–68wound closure, 68

Bisphosphonates, 276Bloom’s syndrome, 3Bone marrow transplantation, 161,

204, 207Bone metastases, 271, 276

bisphosphonates, 276impending fractures, 276pamidronate, 276radiation therapy, 276strontium 276

Bone scans, 271, 275for follow-up, 267–268

Brachial plexopathy, 231Brain metastases, 272–273

clinical features, 272radiation therapy, 272, 273stereotactic radiosurgery, 272–273treatment, 272

BRCA1/BRCA2 genes, 1, 21, 166breast cancer risk, 2description, 21male carriers, 2mutations, 1–5other cancers, 2ovarian cancer risk, 2, 8–9and p53, 122pathophenotype, 7–8probability, 3prophylactic mastectomy, 33

287

Index

psychological assessment, 13radiation and, 8risk, 11

Breast conservation therapy (BCT), 221–225, 232age, 223chemotherapy with, 225contraindication to, 220, 224defect correction, 193–195mastectomy versus, 221nipple-areolar complex, 222Paget’s disease, 222patient variables, 222–223radiation therapy and, 8rare malignancies and, 99recurrence after, 133, 158, 232recurrence risk factors, 221, 222,

223self-esteem, 219

Breast preservation therapy, 141–144central lesions and, 142contraindications for, 142contraindications for, 155ductal versus lobular, 143efficacy of, 141incision placement, 145after induction chemotherapy,

154–159Paget’s disease, 142patient selection, 141–143recurrence risk factors, 144recurrence study, 144

Breast-gastrointestinal tract cancersyndrome, 3

c-erbB-2, 119–121, 282abnormal, 120assays, 120–121CAF and, 121description, 120overexpression, 23, 89, 120, 122as predictive marker, 121, 202scoring, 120as therapeutic target, 121

CA-15-3, 266Calcifications, 47–55

in acini, 52adenosis, 49analysis of, 48–55benign, 50–52biopsy of, 49branching, 53, 54characteristics, 48clustered, 52

in DCIS, 54, 134detection of, 47–48distribution, 49dystrophic, 52fat necrosis, 51–52, 57in fibroadenoma, 51fine, 53indeterminate, 52–55linear, 53, 54lobular, 52malignant, 52–55number, 49pleomorphic, 53on preradiation mammography,

134rod-like, 52secretory, 51, 52shape, 49size, 48–49skin, 50tubular pattern, 50, 51, 52underestimation of, 53vascular, 50See also Microcalcifications

Cancer and Steroid Hormone(CASH) Study, 25

Capecitabine, 281Carcinoembryonic antigen (CEA), 118

for follow-up, 266–267Carcinosarcoma, 101Cardiovascular disease

estrogen replacement therapy and,26, 254

HRT and, 27Chemoprevention, 19, 27–32

agents, 27–32definition, 19strategies, development of, 19studies, 19–21trials, 27–31

Chemoresistance, 154Chemosensitivity, predicting, 166Chest wall recurrence, 149Cholesterol, tamoxifen and, 28Chondrosarcoma, 107Choriocarcinomatous differentiation,

104Cisplatin, 166, 281“Clinging carcinoma,” 90Colloid carcinoma, 95

histopathology, 95presentation of, 45, 46

Comedo carcinoma, 49, 53, 54DCIS, 90

histology, 90Comparisons of survival, 123–124Computer-aided diagnosis, 61Cosmesis

importance of, 219optimal, 142, 172tumor size and, 141–142

Cowden’s disease, 3, 21–22Cox model, 124Cox multivariate model, 125Cribriform DCIS, 91

invasive, 96Cyclin, 117Cyclophosphamide, 166, 203, 281Cyclophosphamide, methotrexate,

5-fluorouracil (CMF), 154, 203, 207

for men, 247meta-analyses, 205–206toxicity, 213

Cylindroma, 102Cyst, 44

aspiration, 80–84calcium in, 50–51sonography of, 44, 80–82

Cystosarcoma phyllodes, 46, 104Cytokine support, 203, 207Cytotrophoblast, 104Cytoxan, adriamycin, 5-fluorouracil

(CAF), 121, 207–208ex-CAF, 203trial, 212

Decision matrix, 263Dietary interventions, 31–32, 34Digitization versus film, 71–72Dimpling, 43Docetaxel, 159, 161, 203Doxorubicin, 166, 203, 206–207, 212

complications, 213side effects, 213

Drug resistance, predicting, 166Ductal carcinoma in situ (DCIS),

143, 171adjuvant systemic therapy, 211as biomarker, 23classification, 90–92comedo type, 54, 90, 132, 134cribriform, 91endocrine differentiation in, 104follow-up, 136–137grading, 90–92histopathology of, 89–93incidence, 131–132

288 INDEX

with LCIS, 93lumpectomy with radiation,

132–135, 211, 227in men, 242microcalcifications, 47, 53micropapillary DCIS, 90–91mortality rate, 136multicentric, 59non-comedo, 55, 134occult multicentric, 132papillary, 91pathology report, 92radiotherapy for, 226–227recurrence, 132–136, 227recurrent, 58re-excision, 136solid, 91standard of care, 137surgical management, 131–138survival rates, 135tamoxifen and fenretinide in,

30–31total mastectomy for, 132treatment, 132–136, 141Van Nuys system, 91–92, 134–135

Ductal casts, 53, 54Ductography, 55–56

Earlier detection, 19, 220Early onset, 33Edema

arm, 230breast, 43, 57recurrent, 57

Elston scheme, 116–117Endocrine differentiation, 104Endocrine metaplasia, 104Endogenous hormones, 23–24Endometrial cancer, 26Epidermal growth factor receptor

(EGFR), 119, 121–122overexpression, 122

Epirubicin, 206Epithelial hyperplasia, 105Essential oils, 31Estrogen receptor status, 28,

116–117, 203Estrogen replacement therapy, 26–27,

253–262Alzheimer’s disease, 256benefits, 254–257benefit versus risk, 261breast cancer on, 258cardiovascular disease, 254–255

cholesterol, 254colorectal cancer, 256–257coronary artery disease, 254–255hip fractures, 256menopause and, 26osteoporosis, 255–256progestin, 255recurrence rates, 259studies, 259–260survival analysis, 259urogenital atrophy, 257vasomotor instability, 257

Estrogenpostmenopausal deficiency, 25risk and, 23–27

Estrogen withdrawal, 253Estrone, 24Exogenous hormones, 24–27Extensive intraductal component

(EIC), 55, 59, 222Extracapsular extension, 228–229

False negative ratio, 263False positive ratio, 263–264Family history, 16, 21–22

clinical features, 3as indicator for mastectomy, 33pedigree, 4, 5, 6risk assessment, 3

Fat necrosis, 44, 51–52, 53, 57Fenretinide, 30Fibroadenolipoma, 44Fibroadenoma, 43, 103, 104–105

biopsy of, 75calcifications in, 51, 52follow-up, 84–85

Fibrosarcoma, 107–108Fibrous histiocytoma, 107–108Fine-needle aspiration (FNA), 70

limitation, 140pitfalls of, 71sensitivity of, 71specificity of, 71stereotactic guidance, 70ultrasound-guided, 81–82versus core biopsy, 140

Fish-oil supplements, 31Flow cytometry, 117Fluid collection, on mammogram, 57Fluorodeoxyglucose, 61Fluorouracil, 203, 204, 2125-Fluorouracil, 2815-Fluorouracil, doxorubicin,

cyclophosphamide (FAC), 154

dose-intensive, 155–157induction, 161for men, 247paclitaxel and, 161standard, 155

Follow-up, 263–270alkaline phosphatase, 268benefits of, 263bone scans, 267–268chemical evaluation, 266chest radiography, 267detecting recurrent, 265health-care costs, 268–269legal implications, 269mammography, 267managed care, 268quality of life, 268schedule, 268timing, 266tumor markers, 266

Formestane, 210Frond-like pattern, 100

Gail Model, 13, 20Galactocele, 44Galactography, 55–56Gemcitabine, 281Genetic counseling, 9–13, 16

algorithm, 10depression rates, 12discrimination, 9–10patient reaction, 9–13

Genetic predisposition, 19Genetic therapies, 14–15Genetics, 1–16Genistein, 31–32Gonadotropin-releasing hormone,

209–210Goserelin, 210, 211Growth factors/receptors, 119

Halstedian theory, 202Hamartoma, 44Headache, 273Hemangiopericytoma, 107Hematoma, 108, 140

aspiration/biopsy, 81on mammogram, 57

HER-2/neu, 119, 140antibodies against, 166

Hereditary breast cancer, 1–2characteristics of, 21–22high S-phase, 7mammography, 15

INDEX 289

management of, 14–15“other,” 7patient education, 15self examination, 15surveillance, 15testing, 16types, 7–8

Hereditary breast-ovarian cancer syndrome, 1, 21

genetics of, 1, 3management of, 14pedigree, 6

Heterocyclic amines, 31Hodgkin’s disease, 22Hormonal therapy

aromatase inhibitors, 278estrogen antagonists, 278metastatic breast cancer, 277–279ovarian ablation, 279progestins, 278–279trials, 208–211

Hormone replacement therapy(HRT), 26–27benefits, 27breast cancer–related deaths and,

27relative risk, 26–27studies, 26

Hysterectomy, prophylactic, 33–34with oophorectomy, 33–34

Ifosfamide, 281Immunotherapy, 282Implants, 174–181

advantages, 181–182bilateral, 192capsular contracture, 175, 180, 231capsular fibrosis, 181delayed reconstruction, 177device failure, 181disadvantages, 181dual chamber, 174, 178evaluation, 58–60extracapsular extravasation, 176irradiation and, 162lawsuits, 176–177mammography and, 176MRI evaluation of, 60nonadjustable, 174position, 177primary reconstruction, 177rupture, 175–176, 181saline, 174,181silicone gel, 174, 175–176, 181

symmetry, 177ultrasonographic evaluation of, 59

Implants, adjustable, 177–179prior radiation, 178, 190

Implants, expanders, 177, 178–181, 189

bilateral, 192In situ carcinoma

presentation of, 43“Indian files,” 94Induction chemotherapy, 154–160Infiltrating breast carcinoma, 93–97Infiltrating ductal carcinoma

architectural patterns, 94cytology, 94–95EGFR in, 122endocrine differentiation in, 104histopathology, 94imaging of, 44–45, 58, 59

Infiltrating lobular carcinoma, 45, 94Inflammatory breast cancer, 159–161Intracystic carcinoma, 99Intracytoplasmic lumens, 93Intraductal carcinoma, 222Invasive breast carcinoma

biomarkers of, 23categories, 93–96grading, 96–97histopathology, 93–97

Invasive cribriform carcinoma, 96Invasive ductal carcinoma, 58, 97Invasive lobular carcinoma, 106, 143

mitotic rate in, 117

Juvenile carcinoma, 103

Kaplan-Meier survival curves,123–124, 160

Keratin pearls, 103Keratohyaline granules, 103Ki-67, 117

Lactating women, imaging, 42, 80Lateral thigh flaps, 191–192Latissimus dorsi flap, 162, 163,

188–189advantage, 162for defect repair, 195disadvantage, 162expander, 189indications for, 188

Lead-time bias, 264Leiomyosarcoma, 107Length-time bias, 264–265

Leptomeningeal carcinomatosis, 273Lesion

fat containing, 44hyperplasia, 23intraductal, 46metastatic, 47nonpalpable, solid, 81precancerous, 23preinvasive, 23risks, 23

Letrozole, 210Leuprolide, 210Li-Fraumeni syndrome, 3, 4, 22, 122Limonene, 31Lipoma, 44Liposarcoma, 107Lobular acini, 93“Lobular cancerization,” 93Lobular carcinoma in situ (LCIS)

as biomarker, 23with DCIS, 93definition, 92–93endocrine differentiation in, 104histopathology, 92–93in men, 242recurrent, 222

Lobular neoplasia, 7Locally advanced breast cancer,

153–169axillary node dissection, 156–157breast conservation, 153, 157–159definition, 153goals, 153irradiation, 158multimodality therapy for, 229radiotherapy for, 227–229recurrence, 158survival, 158TNM, 153

Log-rank test, 123Lumpectomy, 66, 221, 226, 227

boost irradiation, 224–225cosmetic outcome, 224for DCIS, 135–136with radiation, 132–135recurrence risk factors, 144–145for Stage I, II, 141

Lymph node involvement, 147Lymph node metastases

as marker, 115Lymphatic drainage patterns, 147Lymphedema, 147Lymphocytic infiltrate, 8

periductal, 90

290 INDEX

Lymphoplasmacytic infiltrate, 96Lymphoscintigraphy, 148Lynch syndrome, 8

Magnetic resonance imaging (MRI), 60–61

of benign lesions, 60evaluation of implants, 60false positives, 60limitations of, 60–61of malignant lesions, 60measuring tumor response, 166of occult in situ carcinoma, 60sensitivity of, 60specificity, 60

Males, breast cancer in, 21, 228, 239–252

age, 240androgen deficiency, 241androgen-receptor gene, 240axillary metastases, 245biopsy, 244BRCA1/BRCA2, 240comorbidities, 249DCIS, 242ductal hyperplasia, 241EGFR in, 122epidemiology, 240–242estrogen metabolism in, 24estrogen receptor status, 246evaluation, 244family history, 240gynecomastia, 241, 243–244head injury, 241history of, 239hyperprolactinemia, 241incidence, 240Klinefelter’s syndrome, 24, 240,

241LCIS, 242liver disease and, 240localized, 246–248lung carcinoma, 242mammography, 243metastatic, 248nipple discharge, 242node number, 245, 246Paget’s disease, 242papillary carcinoma in, 99pathology, 242presentation, 242–244prognosis, 239, 244, 249prostatic carcinoma, 241receptor status, 244, 248

risk factors, 240–242sarcomas, 242similarities to women, 239–240skin involvement, 245survival, 244–249tamoxifen, 247–248testicular function, 241TNM stage, 243, 245transsexuals, 241treatment, 246–248tumor markers, 247types, 242ultrasound, 243–244

Malignancyclassification of, 89–98secondary signs, 43signs of, 43sonographic features of, 44

Malignant ascites, 275Malignant pleural effusions, 274–275Mammography

assessment categories, 42baseline, 68with biopsy, 66–68breast cancer abnormalities, 43decision categories, 42diagnostic, 41follow-up, 267implants and, 58–60localization, 66for long-term follow-up, 56–57magnification, 41, 55, 56–57for monitoring, 56postoperative, 56postradiation, 56–60radiation risk, 22radiotherapy, 68recommendations, 41reporting, 42screening, 41–44specimen, 92spot compression, 41

Margin distance, optimal, 144Margin involvement, 144Margins, biopsy

clear, 223intraoperative inking of, 66–67,

223negative, 223radiation boost, 225

Margins, tumorassessment of, 43, 66, 92ill-defined, 43imaging, 45

spiculated, 43, 44status, 133–134, 223ultrasound of, 44

Markers, 115–125, 119evaluation of new, 114, 125goal of using, 118guidelines, 115predictive, 114, 119prognostic, 114, 118–119, 125, 225statistical issues, 122–125studies, 118–119trials, 122–123types, 114

Mass, palpable, 42Mass

analysis of, 43–44discrete, 43hypoechoic, 44intracystic, 46, 47irregular, 43mammography of, 43–44margin of, 43shape, 43

Mastectomy defect, 193–195classification, 194flap choice, 195inferior, 194inferior pole, 195recurrence risk, 193supra-areolar, 195surveillance, 195technique, 194upper quadrant, 194

Mastectomy, prophylactic, 32–33, 146

bilateral, 32, 146candidates for, 33indications for, 33unsuspected cancers found, 33

Mastectomy, radical, 146, 221Mastectomy, skin-sparing, 172,

173–177complications, 174esthetic results, 174indications for, 173recurrence with, 174

Mastectomy, subcutaneousdefined, 32prophylactic, 33

Mastectomy, total, 144–145for DCIS, 132defined, 32incision placement, 145indications for, 144

INDEX 291

with reconstruction, 145skin sparing, 145

Medroxyprogesterone, 210Medullary carcinoma, 95–96

atypical, 96EGFR in, 122histopathology, 95–96presentation of, 45, 46

Melphalan, 203, 212Menstrual factors, 24Mesh, 187–188

prolene, 193Metaplastic carcinoma, 95, 103, 107

histopathology, 101Metastatic, treatment, 271–285

biopsy, 271–272bone involvement, 271, 276bone scans, 271, 275brain, 272–273chemotherapy, 279–281immunotherapy, 281leptomeningeal carcinomatosis,

273–274local, 276–279maintenance, 281malignant effusions, 274–275peritoneal, 275prognosis, 272recurrence rate, 271sites, 271

Methotrexate, 121, 156, 203, 281Methotrexate, 5-fluorouracil,

vinblastine (MFVb), 155Microcalcifications

follow-up, 84as indicator for mastectomy, 33monitoring, 56–57visualization of, 41

Microlumens, secondary, 91Micropapillary DCIS, 90–91Milk of calcium, 51Mitotic count, 7–8

as marker, 117Monoclonal antibodies, 119, 121, 166Monomorphism, 90–91Monoterpenes, 31Mortality, decline in, 19Mortality rate, 264Mucin globules, 93Mucinous carcinoma, 46

endocrine differentiation in, 104histopathology, 95

Multicentric disease, 143, 155,221–222

Multiple myeloma, 49Multivariate analyses, 125Mutation search, 2–3

Necrosis, 91, 95Necrotic neoplasm, 44Negative predictive value, 264Neoadjuvant therapy, 211Nicotine use, 184Nipple retraction, 43Nipple thickening, 43Nipple-areolar reconstruction,

195–196, 222Nipple discharge

ductography, 55–56galactography, 55–56in males, 242malignant, 99serosanguinous, 45, 56

“No special type” (NST) invasivecarcinomas, 7, 94, 97

Node number, 154Noninvasive intracystic carcinoma, 99Normal breast anatomy, 81Nuclear grading, 116Nuclear medicine techniques, 61Nulliparity, 25

Obesity, postmenopausalas risk factor, 24

Omega-3 polyunsaturated fatty acids,31

Omega-6 polyunsaturated fatty acids,31

Oncogenes, 119Oophorectomy, 203, 209–210

prophylactic, 11, 13, 15, 33–34protective effect of, 23–24trials, 209–210

Oral contraceptives, 259as risk factor, 24–25, 27studies, 24–25survival, 259

Osseous metaplasia, 101Osseous trabeculae, 107Osteogenic sarcoma, 107Osteoporosis

estrogen replacement therapy and, 26

tamoxifen and, 28–29

Osteosarcoma, 108Ovarian carcinoma, 8–9

BRCA1/BRCA2 mutations and, 8surveillance, 15

p53, 19, 140adenovirus for, 166as biomarker, 23, 122, 202mutations, 122tamoxifen and, 225

Paclitaxel, 159, 161, 203Paget’s disease, 43, 142, 222

in men, 242Pamidronate, 276Papillary carcinoma, 44–45, 47, 95,

99–101Papillary DCIS, 91Parenchyma, irregular, 57Pectoralis muscle, invasion of, 146Pericardial effusions, 274

tamponade, 274Perillic acid, 31Perillyl alcohol, 31Peritoneal metastases, 275Phyllodes tumor, 46–47, 48,

103–106classification, 105differentiation, 106excision, 106histopathology, 105incidence, 104–105presentation, 104

Polarity, cell, 90–91Positive predictive value, 264Positron emission tomography

(PET), 61, 147as predictor of response, 166sensitivity, 61specificity, 61

Postexcision monitoring, 56Postradiotherapy angiosarcoma, 108Predictive value, 264Pregnancy, 258–259

abortion as risk factor, 24BCT, 224imaging, 42, 80protective factor of, 24radiation, 224

Prevalence of disease, 264Primary lymphoma, 106Probability (p) values, 124

significance, 125

292 INDEX

Progestins, 26, 255, 278–279, trials, 210

Progression, 23Proliferating cell nuclear antigen, 117Proliferation rate, 115

as marker, 117Proliferative breast disease, 23Prophylactic mastectomy, 13–14

attitudes toward, 12Proportional hazard statistical

models, 124Pseudocalcification, 50Pseudosarcomatous metaplasia, 101Pseudosarcomatous stroma, 101

Radial scar, 46, 48Radiation, 219–238

for bone metastases, 276boost doses, 223, 230brain, 273BRCA1/BRCA2 mutations and, 8carcinogenic potential, 231carcinogenic risk, 22cardiac injury, 230, 232cardiac toxicity, 227chemotherapy and, 229chest wall, 227–229collagen vascular diseases, 224complications, 230–231contralateral breast cancer, 231for DCIS, 226–227environmental, 22factors, 224local recurrence, 231for locally advanced breastcancer, 227–230lumpectomy and, 221orthovoltage, 228post complete response, 229–230postmastectomy, 231pregnancy and, 224pulmonary injury, 230recurrence rates, 221sarcomas, 231stereotactic, 273–274tangential, 230, 230techniques, 229–230toxicity of, 220

Radiation-induced DNA damage, 14Radiation injury, 142Radiography

magnification, 48

routine for follow-up, 267Raloxifene

as chemoprevention, 29, 34with tamoxifen, 30trials, 29–30

ras Gene, 119Reconstructive surgery, 161–166,

171–200adjuvant chemotherapy, 172autologous, 181–193bilateral, 192–193delayed, 145, 172, 185expanders, 177, 178–181, 189flaps, 182–193immediate, 171–173, 179–181,

185, 195implants, 145, 162, 173, 174–181mastectomy defect, 193–195nipple-areolar, 195–196, 222radiation with, 173

Rectus abdominis myocutaneousflaps, 162

Recurrenceafter BCT, 221–225cytologic factors, 225–226detecting, 265mammographic indications of,

57–58metastatic, 276–282risk factors for, 221

Relapsepatterns of, 202–203radiation therapy, 202risk of, 201–202survival, 202

Reporting, standardizing, 989-cis-Retinoic acid, 29–30Retinoids, as chemoprevention, 28,

30, 34Rhabdomyosarcoma, 107“Rigid bridges,” 91Risk assessment, 12, 19–27

absolute risk, 20–21family history, 16, 21–22genetic counseling, 9–13predisposition testing, 9quantification, 20

Risk biomarkers, 20Risk factors, 19Risk, modulation of, 32“Roman bridges,” 90Rubens flap, 189–190, 191

Sarcoma, breast, 107–108in men, 242

Sarcoma, postradiotherapy, 108Scarff-Bloom-Richardson system,

96–97Scintigraphic imaging, 61Sclerosing adenosis, 52Sclerosing papillomatosis, 46Secretory carcinoma, 103–104Seizures, 272Selective estrogen receptor

modulators (SERMs), 29–30, 34,213

Sentinel lymph nodebiopsy, 159, 231mapping, 229

Sentinel lymphadenectomy, 147–148advantages, 148isosulfan blue, 148radiocolloid, 148

Seromaaspiration/biopsy, 81on mammogram, 57

Shadowing, 44Signs of malignancy, 43Skin changes, 108Skin thickening, 43, 57–58Small-cell undifferentiated

carcinoma, 104Smoking, 31, 184Solid DCIS, 91

histopathology, 91Sonography

features of malignancy, 44of masses, 44

Soy-based products, 31“Special type,” tumors, 94–96Spindle-cell carcinoma, 101Squamous cell carcinoma, 101

histopathology, 103Squamous metaplasia, 103, 105Stage I and II

adjuvant therapy, 149diagnostic evaluation, 139–140fine-needle aspiration, 140recurrence, 149staging, 140, 141stereotactic core biopsy, 139therapy, 140–149treatment algorithm, 141tumor size, 141–142

Stem cell transplant, 207–208, 215

INDEX 293

Stereotactic biopsy, 68–79anesthetic, 72automated Tru-cut™, 71–72complications, 77dedicated prone, 69–70, 69difficulties in, 77–79digital imaging, 71–72equipment, 68–72indications for, 84pull-back depth, 73, 74scout image, 72stroke margin, 77, 79targeting errors, 77, 78technique, 72–75upright add-on, 68

Stereotaxis, 66, 70standardization of, 71

Steroid receptors, 116–117as markers, 116positivity, 116progesterone receptors, 116–117scoring systems, 116

Strontium 89, 276Study of Tamoxifen and Raloxifene

(STAR), 27, 30Subtyping, 117Superior gluteal flap, 190–191Surrogate end-point biomarkers

(SEBs), 19–20, 23trials, 30–31

Surveillance, 12, 263–270data analysis, 263–264Hodgkin’s disease, 22

Surveillance, Epidemiology, and EndResults (SEER) program, 25–26

Survival, apparent increase, 264Syncytiotrophoblast, 104Systemic therapy

improving, 166–167metastatic breast cancer, 277

Tamoxifen, 203, 225as adjuvant treatment, 146benefits, 214bone mineral density, 28–29cardiac events, 29as chemopreventive agent, 27–29,

34, 146chemotherapy plus, 210–211cholesterol, 28clotting factors, 28colorectal cancer, 28complications of, 29

contralateral breast cancer, 28endometrial cancer, 28–29estrogen receptor status and, 116gastrointestinal malignancies, 28invasive breast cancers, 29long-term treatment, 28for men, 247–248noninvasive breast cancer, 29raloxifene with, 29–30resistance, 122side effects of, 29, 213–214studies, 28–29term, 214trials, 208–211uterine cancer with, 213–214vasomotor instability, 257

Taxanes, 203, 280Taxol resistance, 121Technetium-99M (Tc-99m), 61Test evaluation, 265–269Test sensitivity, 263–264Test specificity, 264Therapeutic modeling, 117Thiotepa, 203, 204Thymidine, 117TNM (tumor, nodes, metastases)

staging system, 140, 141, 153Toremifene, 210TRAM, conventional, 182–185

bilateral, 192blood flow in, 182complications, 184–185description, 182macromastia, 185outcome, 186patient selection, 184perfusion in, 182–184perioperative conditions, 183–184postoperative monitoring, 183pre-existing scars, 184radiation of, 173strategy, 184technique, 183zones, 183

TRAM, free, 185–188advantages, 185, 187, 192bilateral, 192contraindications to, 186outcome, 186technique, 185

Transverse rectus abdominismyocutaneous (TRAM) flap, 33,

162–164

benefits of, 190bilateral, 192bipedicled, 188complications, 187–188for defect repair, 195disadvantage, 186fat necrosis, 188free versus pedicled, 186–188hernias, 186–187high-risk patients, 185, 188mesh, 187–188modifications, 187

Trastuzumab, 282Tru-cut™ biopsy, 71–73True negative ratio, 263True positive ratio, 263Tubular carcinoma, 95–96

histopathology, 95presentation of, 46, 47

Tubular-lobular group, 7Tubule formation, 7Tumor

estrogen receptor status, 28–29grade, 92, 116, 202, 222size, 89, 92, 115, 202, 222subtyping, 117

Tumor cellslinear files, 94“targetoid,” 94

Tumor perimeter, 7–8

Ultrasonography, 41, 79–80of augmented breast, 59whole breast, 80

Ultrasound-guidanceadvantages, 84core-needle biopsy, 81, 83cyst aspiration, 81–84selecting, 84

Univariate analyses, 124–125

Vacuum-assisted biopsy, 75–77versus core-needle, 76–77

Van Nuys system, 91–92, 134–135,227

Vertical rectus abdominismyocutaneous (VRAM), 162–164

Vinblastine, 156, 203, 281Vinca alkaloids, 281Vincristine, doxorubicin,

cyclophosphamide, prednisone(VACP), 155

Vinorelbine, 281

294 INDEX

breast cancer atlas of clinical oncology

Documents