Journal of Surgical Oncology 2004;87:121–125

Primary Sarcoma of the Breast

MANOJ PANDEY,1* ALEYAMMA MATHEW,2 ELIZABETH K. ABRAHAM,3 AND BALAKRISHNAN RAJAN4

1Division of Surgical Oncology, Regional Cancer Centre, Medical College, Trivandrum, Kerala, India2Division of Epidemiology, Regional Cancer Centre, Medical College, Trivandrum, Kerala, India

3Division of Pathology, Regional Cancer Centre, Medical College, Trivandrum, Kerala, India4Division of Radiation Oncology, Regional Cancer Centre, Medical College, Trivandrum, Kerala, India

Background and Objectives: Primary sarcoma occurring in breast is rare andcomprises 0.5–1% of all breast neoplasm. Majority of the series include both stromaland cystosarcoma phyllodes, only a few hundred cases of sarcomas other then cysto-sarcoma are reported.Patients and Methods: We carried out a retrospective analysis of 19 patients withprimary sarcoma of the breast treated between 1982 and 2002.Results: Mean age of the patients was 38.6 years (12–70 years). Gradually pro-gressive swelling was the commonest presenting feature. There were eight cases ofangiosarcoma, four cases of spindle cell sarcoma, two each of pleomorphic sarcomaand stromal sarcoma, and one each of malignant fibrous histiocytoma, embryonalrhabdomyosarcoma, and sarcoma (NOS). Eight of these were high-grade (42%). Eightpatients underwent either radical or modified mastectomy, three underwent wideexcisions, and one underwent quadrantectomy. Ten (52.6%) patients received post-operative adjuvant radiation. Two patients received chemotherapy. After a meanfollow-up time of 34.5 months (median 25 months), eight patients failed. Failure waslocal in five, opposite breast in one, and both local and distant in two. The disease freesurvival at 3-year was 39%. In univariate analysis only the margin of first surgery wasfound to be a significant predictor of survival (P¼ 0.05).Conclusions: Primary sarcomas of the breast are aggressive tumors. Surgicaltreatment should consist of at least simple mastectomy. All attempts should be made toachieve a negative margin as this appears to be the only factor influencing survival inthese patients.J. Surg. Oncol. 2004;87:121–125. � 2004 Wiley-Liss, Inc.

KEY WORDS: mesenchymal tumors; neoplasms; soft tissue tumor;angiosarcoma; rhabdomyosarcoma; spindle cell neoplasm; malignant

fibrous histiocytoma; stromal tumors

INTRODUCTION

Primary sarcoma occurring in breast is rare andconstitute 0.2–1.0% of all mammary malignancies [1].About a few hundred cases have been reported in worldliterature. With increased reporting and clinical experi-ence the perceived differences between breast sarcomaand other soft tissue sarcomas are decreasing [2]. Theclinical presentation is usually a gradually progressiveswelling with or without pain. Ulceration occurs if leftuntreated. The spread is commonly blood borne whilelymphatic spread is low [3,4]. Classification of the breastsarcoma is controversial with most of the reported seriesincluding cystosarcoma phyllodes and carcinosarcoma

with other primary sarcomas. As in the case of sarcoma ofother body parts, treatment depends on histological type,degree of differentiation, and tumor size [5]. Surgery isthe primary modality of treatment with adjuvantchemotherapy and radiotherapy indicated in limited stage

*Correspondence to: Dr. Manoj Pandey, MS, Associate Professor, SurgicalOncology, Regional Cancer Centre, Medical College, P.O. Trivandrum695 011, Kerala, India. Fax: þ91-471-2447 454.E-mail: manojpandey@rcctvm.org

Accepted 12 July 2004

DOI 10.1002/jso.20110

Published online in Wiley InterScience (www.interscience.wiley.com).

� 2004 Wiley-Liss, Inc.

disease. We report here a series of 19 cases of primarybreast sarcomas treated at our center over 21-year period.

PATIENTS AND METHOD

Regional Cancer Centre, Trivandrum, is a 350-bedsuper specialty hospital serving a population of 30 millionin the state of Kerala and adjoining part of states of TamilNadu. It maintains an electronic database of over 100,000patients treated since 1982 on Ingress software (Ingress,Inc., Los Angeles, CA). Data was extracted from thisdatabase using ICD-9 and ICD-0 code for breast andhistological code for sarcomas. The general classificationof soft tissue sarcomas was followed [6]. Patients withmalignant cystosarcoma phyllodes were excluded basedon the presence of epithelial element and the presentreport concentrate purely on soft tissue sarcomas. Theterm primary stromal sarcoma was used for the tumorsarising from the specialized hormone sensitive breaststroma purely on morphological ground based on the cellof origin. Multiple sections were studied to excludetumors with epithelial elements. Grading was based oncellular pleomorphism, mitotic activity, and presence ofnecrosis; hence these factors were not looked at separa-tely. Tumors with <5 mitotic figures/10 HPF, absent tomild atypia, moderate cellularity, and absent necrosiswere considered low-grade. Tumors with >5 mitoticfigures/10 HPF, moderate to severe atypia, dense cellu-larity, and presence of necrosis were considered high-grade. A total of 58 cases of sarcomatous tumors wereidentified and reviewed by a single pathologist. Of these,there were 37 cases of malignant phyllodes tumor, 2 casesof carcinosarcoma, and 19 cases of sarcoma other thancystosarcoma, which form the basis of this report. Theclinical treatment and follow-up details were obtainedfrom patients case records. Survival analysis was carriedout by Kaplan–Meier method [7]. The survival resultswere compared univariately using log-rank and Breslow’stest. Hazard ratios were calculated.

RESULTS

Between 1982 and 2002 (21 years), 19 patients withprimary sarcoma of the breast were identified. Mean agewas 38.6 years (range 12–70 years; SD 15.5). Progres-sive lump in the breast was the main presenting symptomin 18 (94.7%) while pain was the presenting feature in 1(5.3%). The mean duration of symptoms was 7.6 monthsranging from 1 to 18 months (Table I).

Histologically there were eight (42.1%) cases ofangiosarcoma, four (21%) spindle cell sarcoma, two(10.5%) cases each of pleomorphic sarcoma and stromalsarcoma, and one case each of malignant fibrous histio-cytoma, embryonal rhabdomyosarcoma, and sarcoma(NOS). Eight of the tumors were high-grade (42.1%) T

AB

LE

I.C

lin

ica

lP

rese

nta

tio

n,

Pa

tho

log

y,a

nd

Tre

atm

ent

inP

ati

ents

Wit

hB

rea

stS

arc

om

a

Sl.

no.

Age

Dura

tion

of

sym

pto

m

Men

opau

se

stat

us

Siz

eof

tum

or

(cm

)N

odes

His

tolo

gy

Gra

de

T

stag

eS

urg

ery

Mar

gin

Ad

juvan

t

RT

Dose

Fra

ctio

ns

Ad

juvan

t

CT

Fai

lure

Mo

nth

s

136

4P

re5

No

Angio

sarc

om

aH

2W

Eþ

LD

Po

siti

ve

Yes

40

20

No

No

25

223

12

Pre

4Y

esA

ngio

sarc

om

aH

1R

efuse

dP

osi

tive

Ref

use

dR

efuse

dN

o0

360

—P

ost

—N

oS

trom

alsa

rcom

aD

on’t

know

xW

Eof

rec

Neg

ativ

eY

es40

20

No

Loca

l84

470

12

Post

18

No

Ple

osa

rcom

aH

2R

MN

oco

mm

ent

No

Liv

erþ

loca

l6

532

3P

re—

No

Spin

dle

cell

sarc

om

aH

XM

RM

Posi

tive

Yes

50

25

ICE

Loca

l100

653

18

Post

10

No

Spin

dle

cell

sarc

om

aL

2N

one

Neg

ativ

eR

efuse

dR

efuse

d0

742

2P

re4.5

No

Sar

com

aN

OS

Don’t

know

2M

RM

No

com

men

tY

es40

20

No

No

69

844

—P

re—

Don’t

know

Angio

sarc

om

aH

XM

RM

Neg

ativ

eN

oN

oN

o0

936

—P

re13.5

Don’t

know

MF

HD

on’t

know

2M

RM

Neg

ativ

eY

es40

20

No

No

2

10

12

12

Pre

men

stru

al4

No

Spin

dle

cell

sarc

om

aL

2W

EN

oco

mm

ent

No

No

No

58

11

26

12

Pre

9N

oA

ngio

sarc

om

aL

2W

EP

osi

tive

No

No

Loca

l1

12

40

4P

re4.5

No

Mal

stro

mal

sarc

om

aI

2W

EN

egat

ive

No

No

Loca

lþlu

ng

34

13

55

5P

ost

12

No

Ple

osa

rcom

aH

2W

EP

osi

tive

No

No

No

0

14

27

18

Pre

18

No

Angio

sarc

om

aI

2R

MN

oco

mm

ent

Yes

50

25

No

opp

bre

ast

9

15

48

1P

ost

—N

oA

ngio

sarc

om

aH

xW

EN

egat

ive

Yes

40

20

No

Loca

l67

16

32

6P

re—

No

Spin

dle

cell

sarc

om

aI

2W

EN

egat

ive

No

No

No

35

17

34

6P

re10

No

Angio

sarc

om

aI

4R

MN

egat

ive

Yes

50

25

No

Loca

l14

18

12

1P

rem

enst

rual

15

Yes

ER

MS

H4

Quad

rante

ctom

yN

egat

ive

Yes

40

20

VA

CN

o73

19

52

6P

ost

—Y

esA

ngio

sarc

om

aL

xM

RM

þL

DN

oco

mm

ent

Yes

50

25

No

No

78

NO

S,nototh

erw

ise

spec

ified

;R

MS

,rhab

dom

yo

sarc

om

a;M

FH

,m

alig

nan

tfi

bro

us

his

tiocy

tom

a;M

RM

,mo

difi

edra

dic

alm

aste

ctom

y;R

M,r

adic

alm

aste

ctom

y;W

E,w

ide

exci

sio

n;L

D,L

atti

smus

dors

ifl

ap;

Pre

,pre

men

opau

sal;

Post

,post

men

opau

sal;

NE

D,no

evid

ence

of

dis

ease

;L

TF,lo

ssto

foll

ow

-up;

Met

s,m

etas

tasi

s;D

F,dis

ease

free

;P

leo,p

leom

orp

hic

;R

ec,re

curr

ence

;o

pp

,o

pp

osi

te.

122 Pandey et al.

while four (21%) were low-grade. Two patients hadpalpable axillary lymph nodes at presentation (Table I).

Three cases were referred primarily to center while 16had underwent surgery elsewhere of these 5 had positivemargin and another 5 had recurrent disease, 6 werereferred after definitive surgery for adjuvant treatment.All patients, except one, were treated by surgery asprimary modality. Ten (52.6%) patients underwent eitherradical (2) or modified radical mastectomy (8), 7 (36.8%)had wide excision; 1 patient underwent quadrantectomy,while 1 patient refused surgery and was lost to follow-up.Three out of five patients undergoing surgery for positivemargin had residual disease. Ten (52.6%) patientsreceived postoperative adjuvant radiation, It was givento five patients who had pectoral muscle involvement,one patient with positive margin, three cases who hadunderwent wide excision only as part of breast conserva-tion while for one patient the indication for radiotherapywas not known. Radiotherapy was delivered to the chestwall using parallel pair six patients received 40 Gy in 20fractions while four received 50 Gy in 25 fractions. Axillaand supraclavicular fossa was not treated in any case.Chemotherapy was used as adjuvant in two patients ofwhich one patient had rhabdomyosarcoma.

Mean follow-up time was 34.5 months (SD 34.8;median) with a median of 25 months. During follow-upeight patients failed. The failure was local in five, oppo-site breast in one (angiosarcoma), and both local anddistant in two. Five out of eight failures were in patientsreceiving adjuvant radiotherapy. Of the eight failuresthree had grade II and three grade III tumors, fourpatients had T2 disease while four of the failures were inpatients with angiosarcoma, two in stromal sarcomawhile one each in pleomorphic and spindle cell sarcoma.All patients with local failure were successfully salvagedby surgery. Two patients died due to progressive disease.The overall disease free survival was 39% at 3-year(median 33 months) (Fig. 1).

Margin of surgical resection was found to be mostimportant prognostic indicator. The survival at 3-yearwas 33.3% in patients with negative margin, 57% in pa-tients where margins was not known, and 0% in patientswith positive margins (P¼ 0.05) (Table II; Fig. 2).Presence of nodes (P¼ 0.27) (Fig. 3), type of previoussurgery (P¼ 0.6), use of adjuvant radiotherapy (P¼ 0.7)(Fig. 4), and grade of the tumor (P¼ 0.65) (Fig. 5) werenot found to be statistically significant (Table II). Pre-sence of positive margin showed a hazard ratio of 7.9,while the hazard for failure in high-grade tumors was 1.9.

DISCUSSION

Primary sarcomas occurring in breast are rare and anumber of histopathological varieties can be seen as in

soft tissue sarcomas at other regions [6]. Cystosarcomaphyllodes, on the other hand, are more common andinclusion of these tumors in clinico-pathological analysisdilutes the data as the behavior, management, andoutcome of these are different from pure sarcomas. Thepresent study reports the results of 19 primary sarcomas,other then cystosarcoma. Controversies exist on theclassification of these tumors, however, the classificationproposed for extremity and truncal sarcomas are general-ly applied to breast sarcomas as well [8]. Controversy stillexists on the term ‘‘stromal sarcoma.’’ Berg et al. (1962)[9] unified all breast sarcomas other than cystosarcomaphyllodes under stromal sarcoma to indicate theircommon origin from breast stroma. Callery et al.(1985) [10], however, proposed that the term ‘‘stromal

Fig. 1. Disease free survival among patients with primary sarcoma ofthe breast.

TABLE II. Survival and Predictors of Survival in Breast Sarcoma(Median Follow-Up 25 Months)

Variable 3-year DFS (%) P value Hazard

Disease free survival 39 — —

Clinical nodes

Positive 100 — —

Negative 25.9 0.27 —

Margin

Positive 0.00 0.05 7.9

Negative 33.33 1.0

Don’t know 57 —

Adjuvant radiotherapy

Yes 40.5 0.7 —

No 40 —

Grade of tumor

Low 66.6 0.65 1.0

Intermediate 25 —

High 30 1.9

Breast Sarcoma 123

sarcoma’’ should be used only for tumors arising fromspecialized hormone sensitive breast stroma. It ispreferred, however, to use the histological descriptionby cell of origin of the neoplasm [8], which is followed inthe present communication.

Clinical presentation of these tumors is similar to thatof sarcomas of trunk and extremity, i.e., a graduallyprogressive mass lesion with or without skin changes andpain. In our series, 18 patients presented with painlessprogressive lump. Surgery is the accepted primarymodality of treatment; however, the optimal adjuvanttreatment is still controversial. Postoperative adjuvantradiotherapy has been shown to achieve good local-regional control in a few series [10–13]; however, there

was no difference in survival among patients receivingradiotherapy and those not receiving it (Fig. 4).

Reported 5-year survival rates for patients of breastsarcoma range from 40 to 91% [3,10,11,13]. The higherrates in some series like that of Christensen et al. [13]were due to the inclusion of cystosarcoma phyllodes.Higher survival is also reported by Zelek et al. [14],however, 57/83 women in their series had malignantfibrous histiocytoma. The overall 3-year disease freesurvival in the present series was 39%. There were nofailures beyond third year in the present series.

Grade of the tumor is reported to be the single mostimportant prognostic factor as in soft tissue sarcomaselsewhere. Eight out of 19 (42%) tumors in the present

Fig. 2. Disease free survival by surgical margins in patients withprimary sarcoma of the breast.

Fig. 3. Disease free survival by nodes in patients with primarysarcoma of the breast.

Fig. 4. Disease free survival by adjuvant radiotherapy in patientswith primary sarcoma of the breast.

Fig. 5. Disease free survival by grade of the tumor in patients withprimary sarcoma of the breast.

124 Pandey et al.

series were high-grade, 4 (21%) each were low- andintermediate-grades. Since the composite grading in thepresent report includes mitosis and necrosis, these twofactors were not separately looked up for prognosis. Thetumor size as a prognostic factor could not be properlyevaluated as a number of patients in two stratums were lowand hence a log rank statistic could not be obtained.Adjuvant therapy too appears to have failed to control thelocal disease as 5 out of 10 patients receiving adjuvantchest wall radiotherapy failed, four of these were local andone in opposite breast. Role of chemotherapy is contro-versial, though there are reports of improved survival withchemotherapy in soft tissue sarcoma elsewhere; there arefew reports of its use in breast sarcomas [14–16]. Chemo-therapy was used only in two patients, hence, it is difficultto comment on its usefulness in the present series.

Margin of surgical resection too have been found tohave a prognostic importance [17]. In the present series 6/19 (31.6%) patients had positive margins, of which 3failed. Two failures were observed in patients withnegative margins, while there were three failures amongpatients where margins were not known. These patentshad undergone surgery elsewhere and there was nocomment on margin in their final histopathology reports.Tumor size, however, could have an important bearing onresection margin and hence outcome, this could not beevaluated in the present study.

Like all other reported series of soft tissue sarcoma ofthe breast the number of patients in the present series tooare small, heterogeneous, and spread over a long periodwith a short follow-up (median 25 months). Furtherpleomorphic sarcoma and MFH have been listed separa-tely. The former term was used when further sub-classification of pleomorphic sarcoma was not possible.Majority of the patients have been treated by differingtreatment protocols, hence, it is very difficult to come todefinite conclusion about the prognostic factors andstandard treatment approach. However, it appears thatbreast sarcomas too behave like extremity sarcomas andhence, there is rationale for using the same treatmentprotocol. We suggest a simple mastectomy followed byadjuvant radiotherapy and chemotherapy (though con-troversial) for all high-grade sarcomas. Breast conserva-tion with postoperative radiotherapy with or withoutchemotherapy, however, can be tried for low-grade sarco-

mas. Achieving a negative margin of excision is ofparamount importance as the hazard of patients withpositive margin was found to be 7.9. Establishment ofmulti-disciplinary breast specialty clinics, with definitivetreatment protocols will go a long way in optimizingtreatment and improving survival in breast sarcomas.

REFERENCES

1. Haberthur F, Feichter GE, Torhorst JK: Sarcoma of the breast:What is the contribution of immunohistochemistry and flowcytometry? [German] Geburtshilfe Frauenheilkd 1993;53:811–813.

2. Moore MP, Kinne DW: Breast sarcoma. Surg Clin North Am1996;76:383–392.

3. Pollard SG, Marks PV, Temple LN, et al.: Breast sarcoma: Aclinicopathologic review of 25 cases. Cancer 1990;66:941–944.

4. McGowan TS, Cummings BJ, O’Sullivan B, et al.: An analysis of78 breast sarcoma patients without distant metastases atpresentation. Int J Radiat Oncol Biol Phys 2000;46:383–390.

5. Navarrete E, Janser JC, Pusel J, et al.: Sarcoma of the breast. J Chir(Paris) 1988;125:501–509.

6. Brennan MF, Casper ES, Harrison LB: Soft tissue sarcoma. In:DeVita VT, Jr., Hellman S, Rosenberg SA, editors. Cancer—Principles and practice of oncology, 5th edn. Philadelphia:Lippincott-Raven; 1997.

7. Kaplan EL, Meier P: Nonparametric estimation from incompleteobservations. J Am Stat Assoc 1958;53:457–481.

8. Al-Nafussi A: Spindle cell tumors of the breast practical approachto diagnosis. Histopathology 1999;35:1–13.

9. Berg JW, De Crosse JJ, Fracchia AA, et al.: Stromal sarcomas ofthe breast. A unified approach to connective tissue sarcomas otherthan cystosarcoma phyllodes. Cancer 1962;15:418–424.

10. Callery CD, Rosen PP, Kinne DW: Sarcoma of the breast. Studyof 32 patients with reappraisal of classification and therapy. AnnSurg 1985;201:527–532.

11. Johnstone PAS, Pierce LJ, Merino MJ, et al.: Primary soft tissuesarcomas of the breast: Local-regional control with post-operativeradiotherapy. Int J Radiat Oncol Biol Phys 1993;27:671–675.

12. Ludgate CM, Anderson TJ, Langlands AO: Sarcoma of the femalebreast-report of a series of 30 cases. Clin Oncol 1977;3:97–105.

13. Christensen L, Schiodt T, Blichert-Toft M, et al.: Sarcomas of thebreast: A clinico-pathologic study of 67 patients with long-termfollow-up. Eur J Surg Oncol 1988;14:241–247.

14. Zelek L, llombart-Cussac A, Terrier P, et al.: Prognostic factors inprimary breast sarcomas: A series of patients with long-termfollow-up. J Clin Oncol 2003;21:2583–2588.

15. Silverman LR, Deligdisch L, Mandeli J, et al.: Chemotherapy forangiosarcoma of the breast: Case report of 30 year survival andanalysis of the literature. Cancer Invest 1994;12:145–155.

16. Givens SS, Ellerbrock NA, Butle JJ, et al.: Angiosarcoma arisingin an irradiated breast. A case review of the literature. Ancer1989;64:2214–2216.

17. Smola MG, Ratschek M, Amann W, et al.: The impact ofresection margin in the treatment of primary sarcoma of thebreast: A clinico-pathological study of 8 cases with review ofliterature. Eur J Surg Oncol 1993;19:61–69.

Breast Sarcoma 125