volume 11 supplement 1 march 2009 - ahpba...s. roayaie, md, d. bassi, s. hiotis, md, d. labow, md...

V O L U M E 1 1 • S U P P L E M E N T 1 • M A R C H 2 0 0 9

www.hpbjournal.com

THE OFFICIAL JOURNAL OF THE

International Hepato-Pancreato-Biliary Association

American Hepato-Pancreato-Biliary Association

European Hepato-Pancreato-Biliary Association

IN THIS ISSUE

Abstracts of the Ninth Americas Hepatopancreatobiliary Congress

12–15 March 2009

Miami Beach, FL, USA

HPB

VO

LUM

E 11 • SUPPLEM

ENT 1 • M

ARC

H 2009

HPB_v11_is1_cover_B_5.0mm.indd 1 2/19/2009 12:08:43 PM

Editor-in-ChiefO. J. Garden, UK

Associate EditorsM. Callery, USAS. Connor, New Zealand

Honorary Regional EditorsY. Nimura, JapanH. Obertop, the NetherlandsS. Strasberg, USA

Editorial BoardR. Andersson, SwedenC. Bassi, ItalyJ. Belghiti, FranceG. Belli, ItalyM. Buechler, GermanyR. Chari, USAM.-F. Chen, TaiwanM. Choti, USAC. Christophi, AustraliaP.-A. Clavien, SwitzerlandJ. E. Cunha, BrazilE. de Santibanes, Argentina

C. Dervenis, GreeceJ. Espat, USAL. Fernández-Cruz, SpainY. Fong, USAA. Frilling, SwitzerlandS. Gallinger, CanadaJ.-F. Gigot, BelgiumH. Gooszen, the NetherlandsD. Gouma, the NetherlandsS. Helton, USAM. Henderson, USAJ. Izbicki, Germany

P. Jagannath, IndiaB. Jeppsson, SwedenL. Kow, AustraliaM. Krawczyk, PolandJ. Krige, South AfricaP. Lai, ChinaD. Mahvi, USAM. Makuuchi, JapanD. Nagorney, USAR. Padbury, AustraliaT. Pappas, USAR. Parks, UK

W. Pinson, USAH. Pitt, USAA. Principe, ItalyM. Rees, UKM. Ryska, Czech RepublicM. Selzner, CanadaF. Sutherland, CanadaT. Takada, JapanY. Tekant, TurkeyT. van Gulik, the NetherlandsJ. Windsor, New Zealand

HPB Editors and Editorial Board

Aims and Scope. HPB is an international forum for clinical, scientifi c and edu-cational communication. Eight issues a year bring the reader leading articles, expert reviews, original articles, images, editorials, and reader correspondence encompass-ing all aspects of benign and malignant hepatobiliary disease and its management. HPB features relevant aspects of clinical and translational research and practice. Specifi c areas of interest include HPB diseases encountered globally by clinical practitioners in this specialist fi eld of gastrointestinal surgery. The journal addresses the challenges faced in the management of cancer involving the liver, biliary system and pancreas. While surgical oncology represents a large part of HPB practice, submission of manuscripts relating to liver and pancreas transplantation, the treat-ment of benign conditions such as acute and chronic pancreatitis, and those relating to hepatobiliary infection and infl ammation are also welcomed. There will be a focus on developing a multidisciplinary approach to diagnosis and treatment with endoscopic and laparoscopic approaches, radiological interventions and surgical techniques being strongly represented. HPB aims to help its readers – surgeons, physicians, radiologists and basic scientists – to develop their knowledge and prac-tice. HPB will be of interest to specialists involved in the management of hepat-obiliary and pancreatic disease however will also inform those working in related fi elds.

ISSN (Print) 1365-182X; ISSN (Online) 1477-2574

Disclaimer. The Publisher, the International Hepato-Pancreato-Biliary Associa-tion and Editors cannot be held responsible for errors or any consequences arising from the use of information contained in this journal; the views and opinions expressed do not necessarily refl ect those of the Publisher, the International Hepato-Pancreato-Biliary Association and Editors, neither does the publication of advertisements constitute any endorsement by the Publisher, the International Hepato-Pancreato-Biliary Association and Editors of the products advertised.

Journal compilation © 2009 International Hepato-Pancreato-Biliary Association. All rights reserved. No part of this publication may be reproduced, stored or transmitted in any form or by any means without the prior permission in writing from the copyright holder. Authorization to photocopy items for internal and personal use is granted by the copyright holder for libraries and other users registered with their local Reproduction Rights Organisation (RRO), e.g. Copyright Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923, USA (www.copyright.com), pro-vided the appropriate fee is paid directly to the RRO. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising and promotional purposes, for creating new collective works or for resale. Special requests should be addressed to: [email protected].

Information for subscribers. HPB is published in 8 issues per year. Subscrip-tion prices for 2009 are:Print + Online: US$920 (The Americas), #593 (Europe), $920 (Rest of World). Prices are exclusive of tax. Australian GST, Canadian GST and European VAT will be applied at the appropriate rates. For more information on current tax rates, please

go to www.wiley.com, click on Help and follow the link through to Journal subscrip-tions. The institutional price includes online access to the current and all online back fi les to January 1st 1997, where available. For other pricing options, including access information and terms and conditions, please visit www.interscience.wiley.com/journals.

Delivery Terms and Legal Title. Prices include delivery of print journals to the recipient’s address. Delivery terms are Delivered Duty Unpaid (DDU); the recipient is responsible for paying any import duty or taxes. Legal title passes to the customer on despatch by our distributors.

Back issues. Single issues from current and recent volumes are available at the current single issue price from [email protected]. Earlier issues may be obtained from Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. Tel: +1 518 537 4700, Fax: +1 518 537 5899, Email: [email protected].

HPB (ISSN 1365-182X), is published eight times a year. US mailing agent: Mercury Airfreight International Inc., 365 Blair Road, Avenel, NJ 07001, USA. Periodical postage paid at Rahway, NJ, USA.

Postmaster: Send all address changes to HPB, Journal Customer Services, John Wiley & Sons Inc., 350 Main St., Malden, MA 02148-5020, USA.

HPB is published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK. Tel: +44 1865 776868, Fax: +44 1865 714591.

Journal Customer Services. For ordering information, claims and any enquiry concerning your journal subscription please go tointerscience.wiley.com/support or contact your nearest offi ce:Americas: Email: [email protected]; Tel: +1 781 388 8598 or 1 800 835 6770 (Toll free in the USA & Canada)Europe, Middle East and Africa: Email: [email protected]; Tel: +44(0) 1865 778 315Asia Pacifi c: Email: [email protected]; Tel: +65 6511 8000

Production Editor: David Christie (email: [email protected]).

For submission instructions, subscription and all other information visit: www.hpbjournal.com

This journal is available online at Wiley InterScience. Visit www3.interscience.wiley.com to search the articles and register for table of contents and e-mail alerts.

Printed in the UK by the Charlesworth Group.

Abstracting and Indexing Services. The Journal is indexed by EMBASE/Excerpta Media, SIIC database.

HPB_v11_is1_ifc_B_5.0mm.indd 1 2/19/2009 12:23:12 PM

HPB

CONTENTS Volume 11, Supplement 1, March 2009

Abstracts of the Ninth Americas Hepatopancreatobiliary Congress

Annual Scientifi c Session and Postgraduate Program

March 12–15 2009

Eden Roc Renaissance Resort

Miami Beach, FL, USA

Disclaimer

This abstract book has been produced using authors-supplied copy. Editing has been restricted to some corrections of spelling and style where appropriate. No responsibility is assumed for any claims, instructions, methods or drug dosages contained in the abstracts: it is recommended that these are verifi ed independently.

(fm_i_ii)HPB_v11_is1_toc.indd fm_i 2/19/2009 12:24:23 PM

(fm_i_ii)HPB_v11_is1_toc.indd fm_ii 2/19/2009 12:24:23 PM

Abstracts 1 to 8President’s Plenary Session

Friday, March 13, 2009 7:45–9:30 am

1

Second hepatic resection for recurrenthepatocellular carcinoma

S. ROAYAIE, MD, D. BASSI, S. HIOTIS, MD,D. LABOW, MD and M. SCHWARTZ, MDMount Sinai Hospital, New York, NY, USA

While several Asian studies have shown promising resultsfor second resection for recurrent hepatocellular carci-noma (HCC), there are no Western series on this topic.The purpose of this study was to determine the outcomeof patients undergoing a second hepatic resection forrecurrent HCC at a Western center and to identifyprognostic variables.Methods: A retrospective review of all patients under-going hepatic resection for HCC from 1/1990 to 1/2008was conducted. Patients underwent second resection ifthey had a single tumor on imaging, Child’s A liverfunction, platelets > 100 000 and no extrahepatic dis-ease.Results: During this period, 487 patients underwentresection with 221 having documented recurrence. Ofthese, 30 underwent second resection. Underlying liverdisease included HBV (n = 18), HCV (n = 6), none(n = 4), and other (n = 2). Mean tumor size was 9.6 cmat first resection and 4.0 cm at second resection. Medianinterval between first resection and recurrence was15.5 months. There were no perioperative mortalities and1, 3, and 5 years survivals were 88%, 65%, and 65%.Recurrence rate at 3 year was 80%. Univariate predic-tors of survival included < 1year interval from firstresection to recurrence, gross vascular invasion at secondresection, tumor > 5cm at second resection, and bloodtransfusion at second resection. Multivariate analysisfound gross vascular invasion at second resection as theonly independent predictor of mortality (HR 76.9,P = 0.002). Patients without gross vascular invasion at

2nd resection had median survival of 76.8 months and5 year survival of 74%.Conclusions: Second resection for recurrent HCC hasexcellent outcomes in well selected patients. Grossvascular invasion is the only independent predictor ofoutcome after second resection.

2

Incorporating an HPB fellowship does notdiminish surgical residents’ HPB experience in ahigh-volume training program

N. J. ZYROMSKI, MD, L. TORBECK, PHD,D. F. CANAL, MD, K. D. LILLEMOE, MD andH. A. PITT, MDIndiana University, Indianapolis, IN, USA

Background: A major paradigm shift in surgical educa-tion has recently been instituted by the American Boardof Surgery and the Surgical Council on Resident Edu-cation (SCORE). Specific surgical procedures have beendefined as ESSENTIAL (Common/Uncommon – specificprocedural competency required by the end of training)and COMPLEX (generic competence required, but notcompetence in individual procedures). Importantly, vir-tually all elective HPB procedures fall into the SCORE‘COMPLEX’ category. Trainees who wish to practiceHPB surgery will therefore be required to obtainadvanced training. As this training paradigm evolves, it isequally crucial that incorporation of an HPB fellowshipinto an established surgical residency program does notdiminish surgical residents’ exposure to complex HPBprocedures. We hypothesized that incorporation of anHPB fellowship into a high volume clinical trainingprogram would not detract from residents’ HPB experi-ence.Methods: Our institution incorporated an HPB trainingprogram in 2005–2006. Resident operative case logs(provided by the American Council of Graduate MedicalEducation) and HPB fellow case logs were reviewed.Resident exposure to complex HPB procedures for3 years prior to and 3 years after fellowship incorpo-ration were compared. Student’s t-test was appliedwhere appropriate; P < 0.05 was accepted as statisticallysignificant.Results: The ACGME requires graduating residents’exposure to a minimum of three pancreas and fourliver cases (complex biliary cases are not subdividedfrom the category ‘alimentary’). In 2007, the nationalaverage exposure of graduating chief residents was:pancreas – 11; liver – 9; biliary – 5. The InternationalHepatopancreatobiliary Association (IHPBA) guidelinesfor HPB fellowship training call for exposure to:pancreas – 30; liver – 25; and biliary – 20. Our institu-tional resident and fellow HPB experience is shown inthe Table 1.

0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0 P

Table 1. Resident and fellow HPB experience 2002–2008.

Pancreas Liver Complex BiliaryResident Fellow Resident Fellow Resident Fellow

2002–2003 20 – 13 – 8 –2003–2004 25 – 22 – 11 –2004–2005 27 – 15 – 9 –2005–2006 23 100 14 40 4 382006–2007 23 109 14 42 8 442007–2008 19 114 13 40 ** 56P value 0.39 0.34 0.17

**Data will be available 12/2008.

Conclusions: These data show that an HPB fellowshipprogram can be incorporated into a high volume clinicaltraining program without detracting from resident HPBexperience. Prior to initiating an HPB fellowship pro-gram, individual training programs must carefully assesstheir capability to provide an adequate clinical experiencefor fellows without diminishing resident exposure tocomplex HPB procedures.

3

Improved survival with aggressive resection ofhilar cholangiocarcinoma

Y. L. CHEAH, MD, M. AKOAD, MD, K. VAKILI,MD, J. J. POMPOSELLI, MD, PHD, E. A. POMFRET,MD, PHD, W. D. LEWIS, MD and R. L. JENKINS,MDLahey Clinic, Burlington, MA, USA

Background: Complete surgical resection with negativehistologic margins is a major determinant of long-termsurvival for patients with hilar cholangiocarcinoma. Asour experience with surgical resection for hilar cholan-giocarcinoma evolved, we adopted a more radical surgi-cal approach with the addition of hepatectomy andcomplex vascular reconstruction to achieve negativesurgical margins. In this study we sought to examine theresults of this aggressive surgical approach and its impacton patient survival.Methods: All patients with the diagnosis of hilar cho-langiocarcinoma evaluated and treated by a single teamof surgeons between 1986 and 2007 were identified fromour computerized database. A total of 120 patientsunderwent resection with curative intent; 65 patientsbetween 1986 and 1998 at the former New EnglandDeaconess Hospital (period 1), and 55 patients between1999 and 2007 at the Lahey Clinic (period 2). Patientdemographics, extent of surgical resection and outcomeswere retrospectively analyzed and compared betweenperiods 1 and 2. Survival is calculated using the Kaplan–Meir method and curves were compared using the logrank test.Results: The overall 1, 3, and 5-year survival rates ofpatients with R0 resection were 85.5%, 63.7% and35.5% compared to 56.6%, 8.2% and 0% for R1 resec-tion (P < 0.001). Patients in period 2 received moreextensive surgical resections; hilar combined with hepaticresection was performed in 69.1% of patients in period 2compared to 48.6% in period 1. Concomitant vascularresection was performed in 17 patients (three in period1 and 14 in period 2), with portal vein resection in 13

patients, hepatic artery resection in two patients, andboth hepatic artery and portal vein resection in twopatients. Negative margins were achieved in 81.8%in period 2 compared to 50.0% in period 1 (P < 0.05).The perioperative mortality rate improved from 6% inperiod 1 to 1% in period 2. The 1, 3 and 5 year survivalin period 1 was 67.1%, 39.4%, and 18.5% compared to79.6%, 48.4%, and 31.1% in period 2 (P < 0.05). Post-operative complications occurred in 22 (40%) patientsin period 2; the most common complication was bile leakin eight patients.Conclusion: Aggressive surgery with the addition ofpartial hepatectomy and vascular resection in treatinghilar cholangiocarcinoma improves patient survival withacceptable morbidity and mortality.

4

Validation of a predictive algorithm to maxi-mize resectability of pancreatic adenocarcinoma

P. BAO, MD, D. POTTER, PHD, J. YOUNG,D. EISENBERG, MD, D. LENZNER, MS, K. LEE,MD, H. ZEH, MD, M. SANDERS, MD, S. HUGHES,MD and A. J. MOSER, MDUniversity of Pittsburgh, Pittsburgh, PA, USA; UPCIBiostatistics Facility, Pittsburgh, PA, USA

Introduction: The surgeon’s major contribution topatients with localized pancreatic cancer is a marginnegative resection. We hypothesized that a predictionalgorithm based on preoperative computed tomography(CT) and endoscopic ultrasound (EUS) could maximizethe rate of R0 resection while reducing the risk of non-therapeutic surgery.Methods: 197 patients with biopsy-proven pancreaticadenocarcinoma (157 head; 40 body/tail) underwentexploration with intent to resect from 2002 to 2007. Allpatients had staging helical CT and 143 had EUS. Aprediction model was developed from the imaging data of65 patients during 2002–2005. This algorithm classifiedpatients as high or low risk for noncurative surgery. Itwas validated in a subsequent cohort of 78 patientsbetween 2005 and 2007. Model performance was eva-luated using contingency table and survival analysis.

0.0 0.0

0.2

0.4

0.6

0.8

1.0

1.0 2.0 3.0 Years

Cu

m S

urv

ival

P

Results: Predictors of resectability in the developmentcohort were: any evidence of vascular involvement on CTscan, EUS stage IIB by AJCC criteria, and EUS tumorsize greater than 2.6 cm. The resectability rate in thevalidation cohort was 77% with a 58% R0 resection rate.Compared to these outcomes, selecting operative patientsclassified by the model as favorable for curative surgerywould have increased the resectability rate from 77% to92% (P = 0.03) and the R0 resection rate from 58% to73% (P = 0.08). The model was 71% accurate with 78%sensitivity and 61% specificity for R0 resection. Therewas a 40% difference in R0 resectability between highrisk and low risk patients (P = 0.001). High risk classi-fication was associated with unresectable locally-ad-vanced disease (P = 0.007), metastasis (P = 0.012), andthe need for mesenteric vein resection (P = 0.012).Compared to patients at low risk, those at high predictedrisk had a 67% rate of noncurative surgery and experi-enced significantly shorter median survival (12.3 vs.20.6 months, P = 0.022).

Conclusion: This validated prediction algorithm usesstandard CT and EUS criteria to identify patients mostlikely to benefit from resection for pancreatic adenocar-cinoma. The model significantly increased the R0 resec-tion rate and reduced the rate of nontherapeutic surgery.Prediction models may be used to classify patientsentering neoadjuvant therapy trials.

5

Serial molecular analysis of EUS guided FNAaspiration of pancreatic cysts: quantitativemolecular evidence detects neoplastic cystprogression, stability or regression

D. MALLAT, MD, J. P. TAZELAAR, C. SPENCE,PHD, E. M. ELDER, SCD and S. D. FINKELSTEIN,MDBaylor University, Dallas, TX, USA; RedPath IntegratedPathology, Pittsburgh, PA, USA

Introduction: The biological behavior of pancreaticcystic neoplasms, particularly those of mucinous origin,is difficult to predict. Progression rates of 30% andhigher have been reported in mucinous cystic processesleading to recommended surgical excision in all cases.Regression of mucinous cysts over time is suspectedto occur, however objective support has been lacking.

Molecular analysis of aspirated pancreatic cyst fluid(point mutation, loss of heterozygosity) is used here toevaluate the presence and rate of that mutational changeover time.Design: The fluid from 52 patients with EUS-FNApancreatic cysts underwent molecular analysis con-sisting of measurement of 1) DNA quantity, 2) DNAquality (extent of degradation), 3) Kras point mutation,4) allelic imbalance (loss of heterozygosity determina-tion [LOH]) for a panel of 16 markers and 5) degree ofclonal expansion of DNA alterations when present(PathFinderTG�). In each patient, serial aspirationswere available (40-two serial, 8-three, 4-four) totaling 120separate cyst fluid genotyping reactions. The intervalbetween serial analyses ranged from 3 to 36 months.Mucinous cysts were defined as cysts containing gross ormicroscopically visible mucin, elevated CEA, Kras pointmutation and/or multiple LOH alterations. Cysts notmeeting these criteria were classified as serous/reactiveprocesses.Results: Forty-two of 52 patients had cysts meeting thecriteria for mucinous etiology. 20 (48%) manifestedneoplastic regression by reduction in DNA amount, shiftfrom good to poor quality, elimination of some or all ofthe point mutations/LOH change present previously and/or lowering in the degree of clonality. The mucinous cystsof 19 patients (45%) remained essentially stable whilethree patients (7%) manifested molecular evidence ofprogression at 6, 8 and 10 months. All 10 non-mucinouscyst fluids showed stable indolent molecular features (lowDNA, poor quality, 0–2 low clonality alterations).Mucinous cysts with indolent molecular features did notmanifest neoplastic progression.Conclusions: Molecular analysis provides a useful ancil-lary tool with which to characterize aspirated pancreaticfluid. The vast majority of mucinous cysts did notmanifest neoplastic progression for up to 36-monthfollow-up. Only 7% of mucinous cysts showed mole-cular progression which was predicted in the initialsample by aggressive molecular changes. The resultssupport integrated molecular pathology risk stratificationof patients with pancreatic cysts including mucinouscystic lesions.

6

A simple risk score predicts inpatient mortalityafter liver resection for hepatocellularcarcinoma

J. P. SIMONS, MD, S. C. NG, MS, J. S. HILL, MD,S. A. SHAH, MD, Z. ZHOU, MD, PHD andJ. F. TSENG, MD, MPHUniversity of Massachusetts Medical School, Worcester,MA, USA

Objective: To develop an integer-based risk score basedon national data to estimate the risk of in-hospitalmortality in patients undergoing procedures for hepato-cellular carcinoma.Background: There is a wide spectrum of disease burdenin hepatocellular carcinoma accompanied by severaloptions for surgical management. However, the associ-ated mortality of such procedures is not well-defined.Accurate predictions of patients’ perioperative riskwould be helpful to guide decision-making.

Low Risk73% R0

(+) (–)

(–) (+)

(–) (+)

EUS Stage 1A

EUS Stage 2B/3

(–) (+) Size>2.6 cm

Prediction algorithm for Risk of Nontherapeutic Laparotomy

CT vessel

High-Risk67% Not R0

Figure 1.

Abstracts 3

� 2009 The AuthorsJournal Compilation � 2009 Hepato-Pancreato-Biliary Association, HPB, 11 (Suppl. 1), 1–104

Methods: The Nationwide Inpatient Sample, the largestall-payer discharge database in the U.S., was queriedfrom 1998 to 2005. A cohort of patient-discharges forhepatic procedures with a diagnosis of primary liverneoplasm was assembled. Procedures were categorized ashepatic lobectomy, wedge resection, or enucleation/ablation. Logistic regression and bootstrap methods wereused to create an integer risk score for estimating the riskof in-hospital mortality using procedure type, patientdemographics, comorbidities, and hospital type. A ran-domly selected sample of 80% (n = 2274) of the cohortwas used to create the risk score with validation of thescore conducted in the remaining 20% (n = 560).Results: A total of 2834 patient-dischargeswere identified.Overall in-hospital mortality was 6.53%. Factors includedin the final model were age, gender, Charlson comorbidityscore, procedure type, and teaching hospital status. Inte-ger values were assigned to these characteristics, and thenused for calculating an additive score (Figure, right panel).Four clinically relevant score groups were then assembledto stratify risk of in-hospital mortality, with a 13-foldgradient of mortality ranging from 1.6 to 21.5% (Figure,left panel; P < 0.0001). In the derivation set, as inthe validation set, the score discriminated well, with ac-statistic of 0.74 and 0.72, respectively.Conclusion: An integer-based risk score can be used topredict in-hospital mortality after surgical procedure forhepatocellular carcinoma, and may be useful for preop-erative risk stratification and patient counseling.

7

SNPs in RecQL predict survival in pancreaticadenocarcinoma

M.-C. GINGRAS, PHD, D. LI, PHD, S. E. HODGES,R. A. GIBBS, PHD, F. C. BRUNICARDI, MD andW. E. FISHER, MDBaylor College of Medicine, Houston, TX, USA; M. D.Anderson Cancer Center, Houston, TX, USA

Introduction: RecQL is a DNA helicase involved in DNAmismatch repair. The RecQL 3’UTR A159C genotypehas previously been associated with overall survival ofpatients with pancreatic cancer. However, the functionalsignificance of this SNP remains unknown.Hypothesis: The RecQL A159C SNP is in linkagedisequilibrium with other functional SNPs of the gene.Somatic mutations of the RecQL gene in tumors mayinfluence the clinical outcome of pancreatic cancer.Methods: We sequenced the entire coding regions of theRecQL gene in paired blood and tumor DNA of 35patients with resectable pancreatic cancer treated withsurgery and adjuvant chemoradiation. DNA was isolatedfrom blood using the PAXgene Blood DNA kit (Pre-AnalytiX) and from matched tumors using the QIAamp

DNA Mini kit (Qiagen). Primer sets were designed tocover the 15 RecQL exons with their surroundingintronic regions. Sequencing was performed on ABI 3700DNA Sequencers. SNPs and somatic mutations werevalidated with Biotage pyrosequencing. SNPs thatshowed significant association with overall survival werefurthered tested in blood DNA samples of 120 patientswith resectable pancreatic adenocarcinoma who receivedneoadjuvant chemoradiation. Univariate analysis of theeffect of genotype on time to recurrence and overallsurvival was performed using the Cox proportionalhazards models.Results: Several previously reported and newly identifiedSNPs but no nonsynonymous SNPs were found inthis study. In addition to the RecQL A159C SNP(rs13035), 2 SNPs, located in introns 2 and 11(rs10841834, rs2159943) showed a significant associationwith overall survival. The three SNPs are part of thesame haplotype block (linkage disequilibrium). The var-iant allele of each SNP had a similar effect on overallsurvival of patients receiving either adjuvant or neoad-juvant therapy. No mutations were detected in thetumors in the exonic regions of the RecQL gene.

Conclusion: The SNPs from the 3’UTR and intronic 2and 11 regions (CC, TT, and AA genotypes) of RecQLare associated with improved overall survival of patientswith resectable pancreatic cancer. The functional signi-ficance of these SNPs warrants further investigation.

8

Biliary complications including a single donormortality: experience of 207 adult-to-adultliving donor liver transplantation

M. A. EL-METEINI, MD, A. F. HAMZA, MD,A. A. ABDALAAL, MD, M. F. FATHY, MD,M. BAHAA, MD, A. MOKHTAR, MD,F. ABOUELFETOUH, MD, I. MOSTAFA, MD andA. EL-DORRY, MDAin-Shams University, Cairo, Egypt; Wady ElNeelHospital, Cairo, Egypt

Introduction: Donor safety is crucial in Living donorliver transplantation (LDLT) with ‘do no harm’aphorism printed in the transplant team mind. Biliaryanomalies are more common in right liver compared toleft liver grafts. Biliary complication is the main cause ofmorbidity following right lobe donation. Mortalityfollowing right lobe donation has been estimated to beless than 0.5%.Patients and methods: Between November 2001 to date,207 adult-to-adult LDLT has been done using right lobegrafts. The donors included 173 men and 34 women withmean age 28.4 ± 5.2 years. Siblings were144 (69.6%)cases while unrelated donors were 63 (30.4%) with amean body mass index 25.2 ± 2.4. Liver biopsy is

Score Calculation

Age groups 0 ≤553 56-65 3 66-75 6 >75 0 0 Charlson score

Procedure type

Gender

Hospital type N(%): 554(24.4) 859(37.8) 624(27.4) Score

Est

imat

ed m

orta

lity

(%)

0–5 6–10 11–15 16–26

21.47

8.77

3.78 1.65

0

5

10

15

20

25

237(10.4)

2 1 5 2 10 3 0 Enucleation 2 Wedge 7 Lobectomy 0 Female 1 Male 0 Teaching 2 Non-teaching

Figure 1.

rs10541534 (C/T)

Log-rank P value: 0.005 Log-rank P value: 0.01 Log-rank P value: 0.001

rs2159943 (T/C) rs13035 (A/C)

Figure 1.

Abstracts4


routinely done and steatosis less than 15% is accepted.Single and multiple right bile duct (RBD) were present in82 (39.6%) and 125 (60.3%) donor, respectively. Multi-ple RBD included 2 or 3 RBD in 116 (56%) and 9 (4.3%)cases, respectively. The mean operative time was360 ± 50 min. with an estimated blood loss around950 ± 450 mL and returned cell-saver amount of450 ± 334 mL. Two donors (0.9%), each received twoblood bank units. Donor remnant liver volumes (RLV)were 33.5 ± 3.2%. Mean ICU stay was 3 ± 0.7 daysand mean hospital stay was 14 ± 3.5 days.Results: The overall biliary complications occurred in 27(13.04%) cases. Following modified Clavien classifica-tion, biliary complications were graded as grade I(n = 10), grade II (n = 2), grade III (n = 14) and gradeV (n = 1). Grade I and II (n = 12) biliary complicationswere successfully managed conservatively. Grade IIIcases were treated by ultrasound guided aspiration,

ERCP and surgery in 10, 2 and 2 donors, respectively.The later 2 donors were treated by t-tube insertion in onecase and duct-to-duct anastomosis following transactionof the CBD in the other case. The later case neededultrasound guided dilatation and 10F stent insertiontwice with normal liver profile thereafter. Single donormortality (Grade V) (0.4%) occurred following biliaryleakage from RBD stump followed by peritonitis thatnecessitated exploration on day 11, re-exploration onday 31 and ERCP with stent insertion on day 38 andthe donor succumbed on day 43 due to uncontrolledsepsis.Conclusion: Donor biliary complications accounts for upto 43% of all complications. The majority of biliarycomplications is minor and could be managed conser-vatively. However, uncontrolled biliary leakage is aserious morbidity that should be avoided as it could leadto a mortality.

Abstracts 5


Abstracts 9 to 16Free Papers – Liver I

Saturday, March 14, 2009 8:10–10:10 am

9

Surgical downstaging and neo-adjuvant therapyin metastatic colorectal carcinoma with irino-tecan drug eluting beads: a multi-institutionalstudy

M. BOWER, MD, K. ROBBINS, MD, D. TOMALTY,MD, C. R. SCOGGINS, MD, K. M. MCMASTERS,MD, PHD and R. C. MARTIN, MDUniversity of Louisville, Louisville, KY, USA; BaptistHealth, Little Rock, AR, USA; Huntsville Hospital,Huntsville, AL, USA

Background: Neoadjuvant chemotherapy for potentiallyresectable metastatic colorectal cancer (MCC) is becom-ing a more common treatment algorithm. However, withthe ever increasing hepatic toxicity with systemic che-motherapy a more target approach with maximumresponse and minimal hepatic toxicity is needed. The aimof this study was to evaluate the efficacy of precisionhepatic arterial Irinotecan therapy in potentially resect-able MCC.Methods: An open-label, multi-center, multi-nationalsingle arm study of MCC patients, who receivedhepatic arterial Irinotecan. Primary endpoints weresafety, tolerance and metastatic tumour resection.Results: Fifty-five patients with metastatic colorectal tothe liver underwent a total of 90 hepatic arterialIrinotecan treatments. Thirty (55%) Women, 25 (45%)Men, 39 (71%) Caucasian, with a median age of52 years (range 42–75) were treated. The extent ofliver involvement was 41 (75%) of patients had< 25% Tumor Replacement, 15% (< 26–50% tumorreplacement), 10% (> 50% replacement), with mediannumber of hepatic lesions being 4 (range 1–20), totalsize of all target lesions being 9 cm (range 5.5–28 cm),and 50% of patients having bilobar tumor distribution.Median number of irinotecan treatments were two(range: 1–5), median treatment dose was 100 mg (range100–200), with total hepatic treatment of 200 mg(range 200–650), with 86% of treatments being per-formed in a lobar infusion treatment, and 30% ofpatients treated with concurrent simultaneouschemotherapy. Eleven (20%) patients demonstratedsignificant response and downstage of their disease ordemonstrated stable disease without extra-hepaticdisease progression that they under went resection,ablation, or resection and ablation. There were nodeaths with morbidity occurring in 20% ofpatients, with none of them being hepatic related.Non-tumorous liver resected demonstrated no evi-dence of steatohepatitis from the Irinotecan arterialinfusion.Conclusions: Hepatic arterial infusion with Irinotecanwas safe and effective in the treatment of MCC asdemonstrated by a minimal infusion complication rate,acceptable tumor response and sustained diseasestabilization. Hepatic arterial infusion is an acceptabletherapy for evaluating the biology of metastatic colo-rectal to the live prior to planned hepatic resection.

10

Exsanguination protocol improves survival aftermajor hepatic trauma

V. ZAYDFUDIM, MD, W. D. DUTTON, MD, I. D.FEURER, PHD, B. K. AU, B.SC., C. W. PINSON,MD, MBA and B. A. COTTON, MDVanderbilt University Medical Center, Nashville, TN,USA

Background: Hepatic injury remains an important causeof exsanguination after major trauma. Recent studieshave noted a dramatic reduction in mortality amongseverely injured patients when trauma exsanguinationsprotocols (TEP) are employed. We hypothesized thatutilization of our institution’s TEP at the initiation ofhospital resuscitation would improve survival in patientswith significant hepatic trauma.Methods: TEP, which involves immediate and sustainedrelease of blood products to the hemodynamicallyunstable trauma patients in the operating room inpre-defined ratios, was initiated in February 2006. Allpatients who (1) underwent immediate operative inter-vention, (2) sustained intra-abdominal hemorrhage withGrade 3–5 hepatic injury, and (3) received the TEPbetween February 2006 and January 2008 were pro-spectively identified. A pre-TEP cohort was retrospec-tively identified from among all trauma patients admittedbetween February 2004 and January 2006 who (1)received immediate operative intervention, (2) weretreated for major intra-abdominal hemorrhage with aGrade 3-5 hepatic injury, and (3) received more than 10units packed red blood cells in the first 24 h. Univariateand multivariate analyses were used to examine the ef-fects of TEP on blood product use during the first 24 h ofresuscitation and evaluated the effects of demographicand clinical covariates on survival.Results: Seventy-five patients were included in the anal-ysis: 39 constituted the pre-TEP cohort (31% 30-daysurvival) and 36 were treated with the TEP protocol(53% 30-day survival). There were no differences in age,gender, mechanism, grade of injury, lobar involvement,or injury to major hepatic vasculature between the twocohorts (all P > 0.25). However, patients treated withTEP had higher injury severity scores (P < 0.01). While24-hour use of blood products did not differ betweencohorts, TEP patients received more FFP and plateletsduring operative intervention and significantly less crys-talloid (all P < 0.01). After adjusting for age, gender,mechanism, and injury severity score; Grade 5 injury andinvolvement of major hepatic vasculature had significantnegative effects on survival (both P £ 0.02), while utili-zation of TEP improved the odds of 30-day survival by78% (OR = 0.22, CI: 0.06–0.81, P = 0.02).Conclusions: An exsanguination protocol allows formore effective utilization of FFP and platelets duringintra-operative management of major hepatic injury.TEP was associated with a significant improvement in30-day survival among patients treated for intra-abdominal hemorrhage associated with significanthepatic trauma.


11

Evaluation of perioperative chemotherapy usinga prognostic nomogram for survival followingresection of colorectal liver metastases

S. K. REDDY, MD, M. W. KATTAN, PHD, C. YU,PHD, E. P. CEPPA, MD, S. G. DE LA FUENTE, MD,Y. FONG, MD, B. M. CLARY, MD and R. R. WHITE,MDDuke University Medical Center, Durham, NC, USA;Cleveland Clinic, Cleveland, OH, USA; Memorial SloanKettering Cancer Center, New York, NY, USA

Introduction: Nomograms are statistical tools designedto predict outcomes and risk stratify patients for clinicaltrials and multi-modality therapy. The aim of this studywas to evaluate the effects of perioperative chemotherapyon disease-specific survival (DSS) after resection ofcolorectal liver metastases (CLM) using a prognosticnomogram incorporating demographic and clinicopath-ologic variables established at Memorial-Sloan KetteringCancer Center (MSKCC).Methods: An external cohort comprised of 203 consec-utive patients who underwent resection of CLM between1996 and 2006 at Duke University Medical Center wasused to validate the nomogram and to evaluate the effectsof perioperative chemotherapy on DSS after resection.Results: Similar to the MSKCC population, the externalcohort included patients with node-positive primarydisease (61.1%), rectal primary tumors (22.2%), disease-free interval

13

Resection of colorectal cancer (CRC) livermetastases: What is an adequate margin?

R. T. PADBURY, MD, PHD, D. VANDEWEYER,MBBS, G. J. MADDERN, MD, PHD and J. W. CHEN,MD, PHDFlinders Medical Centre, Adelaide, Australia; The QueenElizabeth Hospital, Adelaide, Australia.

Traditionally a 1cm margin has been accepted as the goalwhen resecting CRC liver metastases. Evidence is emerg-ing that a lesser margin may provide good outcomes, buta critical margin, below which recurrence is higher andsurvival poorer, has not been universally agreed. In arecent publication1 we reported peri-operative morbidityand clear margin as the two independent prognosticfactors. In this study we defined a clear margin as anabsence of tumor cells within 1 mm of the transectedsurface. The aim of the current study is to further analysethe effect of the width of the surgical margin on patientsurvival to determine whether a margin of 1 mm isadequate.Methods: Two hundred and sixty-one consecutive pri-mary liver resections for CRC mets from 1992 to 2007were analysed (including 197 patients from1). 163(62.5%) were male. The median age was 64 (22–92) years.The 30 day (inhospital) mortality was 1.5%. The primaryCRC Duke staging include 11 As, 70 Bs, 110 Cs and 70Ds. Initial analysis was performed on five groupsaccording to the resection margins; involved, minimalmargin (0–1mm), > 1–4mm, > 4 to < 10mm and‡10mm. Subsequent analysis was based on two groups:margin £ 1mm and > 1mm.Results: With a median follow-up of 4.7 years, theoverall 5 year patient & disease free survival (DFS) were38% & 22% respectively. The 5 year patient and DFSwas significantly better in patients with ‡ 10mm resectionmargin compared with those of involved margin (43.4%vs. 19.4% P < 0.03; 28.7% vs. 16.9% P < 0.02 respec-tively). There was no significant difference in patient orDFS between the three groups with margin > 1mm. The5-year patient survivals in these groups were similar.When a comparison is made between patient with eitherinvolved or £ 1mm margin with patients with > 1mmmargin, there is a significant 5 year patient survival dif-ference of 25% versus 43% (P < 0.04). The DFS dif-ference however did not reach statistical significance(P = 0.14)

Conclusions: In this cohort with a medium follow-up of5 years, we can demonstrate that a margin of > 1 mm isassociated with significant better 5 year survival. Thepossible beneficial effect of greater margin beyond 1mmcould not be clearly demonstrated in this cohort.Reference:1. Scheisser M, Chen JW, Maddern GJ, Padbury RT.

Perioperative morbidity affects long term survival inpatients following liver resection for colorectal metas-tases. J Gastrointestinal Surg 2008; 12: 1054–60.

14

Locally advanced intrahepatic cholangiocarci-noma: survival benefit with aggressive hepaticresection

S. H. TEH, MD, D. CUSATI, MD, J. GRAMS, MD,E. ABOIAN, MD, L. CHANG, MD, S. CHA, PHD,L. BURGART, MD and D. M. NAGORNEY, MDSacred Heart Medical Center, Eugene, OR, USA; MayoClinic, Rochester, MN, USA; Mayo Clinic, Rochester,MN, USA

Introduction: Intrahepatic cholangiocarcinoma (ICC) israre and often presents as locally advanced disease. Wehypothesized that aggressive major hepatic resection mayimprove the survival of patients with ICC.Methods: All consecutive patients who underwenthepatic resection for ICC from Jan 1993 to Jan 2003 wereretrospectively reviewed. All pathological specimens werere-examined by two independent pathologists.Results: There were 100 patients (male = 40, female =60) with a mean age of 63 years old. Eighty-six patientshad major hepatic resections, of which 30 were extendedhepatic resections and eleven involved resection andreconstruction of an extrahepatic structure (vena cava =7, main portal vein = 2, and common bile duct = 2).The mean tumor size was 7.7 cm; multifocal disease waspresent in 27 and regional invasion in 17. There were 89R0 resections and eight R1 resections despite an extendedhepatic resection. Perioperative morbidity and mortalitywere 37% and 3%, respectively. Mean hospitalizationwas 9 days. Multivariate analysis demonstrated that thefactors significantly predicting survival were R0 resection(P < 0.0001, HR 3.92), tumor size (P = 0.0001, HR3.74), metastasis to lymph node(s) (P = 0.0001, HR3.241), older age (P = 0.002, HR 3.11), and tumormultiplicity (P = 0.02, HR 1.78). The median survivalrate for R0 and R1 resections was 4.97 years vs.8.9 months, respectively. Overall 1-, 3-, and 5- year sur-vival rates for R0 and R1 resection were 84% vs. 45%,57% vs. 18% and 49% vs. 0%, respectively. Diseaserecurred in 59% of patients after a mean follow up of4 years (hepatic recurrence = 53%, extrahepatic =47%).Conclusions: Multivariate analysis demonstrated bothtumor and patient factors as critical predictors of longterm outcome in the management of locally advancedICC. However, the most significant factor predictingsurvival was ability to achieve an R0 resection. Effec-tive preoperative multimodality therapy is needed todownstage tumors, and patients should undergoaggressive major hepatic resection with the goal of R0resection.

0 1 2 3Patient Survival (years)

Resection margins

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

P1 mm

4 5 6

Figure 1.

Abstracts8


15

Long-term outcome following sequentialresections of liver and lung metastases fromcolorectal carcinoma

R. MARUDANAYAGAM, MS,R. KRISHNAMOORTHY, MS, V. SHANMUGAM,MS, C. COLDHAM, NONE, S. BRAMHALL, MD,MBBS, D. MAYER, MD, MBBS, J. BUCKELS, MD,MBBS and D. MIRZA, MSQueen Elizabeth Hospital, Birmingham, UK

Background: Surgical resection of colorectal livermetastases (CLM) is an established form of treatment.Limited data exists on the value of sequential hepatic andpulmonary metastatectomy. We analysed patients whounderwent sequential liver and lung resections for CLM.Methods: A total of 910 patients who underwent liverresection for CLM between January 2000 and December2007, were analysed to identify patients with resectablepulmonary metastases (n = 43; 4.7%). Patient demo-graphics, overall survival and survival difference betweensynchronous and metachronous pulmonary metastatec-tomy group and between ‘liver and lung resection’ groupand matched ‘liver resection only’ group without pul-monary metastases (matched for age, primary diseasestage, interval to liver resection, and liver disease stage)were analysed.Results: Forty-three patients underwent sequential liverand lung resections. The median age was 62 years. Themedian number of liver lesion detected was three and themost commonly performed procedure was right hemi-hepatectomy (41.9%). Right upper lobe was the pre-dominant site of lung metastasis (51%). Ten patients hadsynchronous lung metastasis. The median interval be-tween liver and lung metastatectomy was 25 months. The1-, 3- and 5-year overall survival rates after first metas-tatectomy were 100%, 87.1% and 53.9% respectivelywith a median survival of 42 months. Metachronouspulmonary metastatectomy group had better 1-, 3- and5-year survival rates than the synchronous group (100%,88.9% and 60.9% vs. 100%, 75% and 0% respec-tively).There was no significant survival differencebetween the ‘liver and lung resection’ and the ‘liverresection only’ groups.Conclusion: Sequential liver and lung resections formetastases from colorectal carcinoma have good long-term survival for selected patients. Presence of synchro-nous lung and liver metastases was not associated withlong term survival.

16

Predictors of blood transfusion requirement inelective liver resection

A. J. COCKBAIN, MBBS, T. MASUDI, MBBS,G. J. TOOGOOD, MD, J. P. LODGE, MD andK. RAJ PRASAD, MDSt James’ University Hospital, Leeds, UK

Background: Liver resection for primary and secondarymalignancy remains major surgery frequently requiringintra-operative blood transfusion. The balance betweensafer surgical techniques and more extensive resections

has direct implications on blood transfusion require-ments. In addition to increasing shortage of blood stocks,and the cost and side effects of blood transfusion, there isconcern that perioperative blood transfusion has animmunomodulatory effect which adversely affectstumour recurrence and prognosis.Aim: To identify predictors of perioperative bloodtransfusion.Methods: A retrospective review of a prospectively col-lected database of all elective hepatic resections under-taken in a tertiary referral centre over a 4-year period wasperformed. Data analysed included patient demograph-ics, comorbidities, underlying liver disease, haematolog-ical parameters, preoperative radiological intervention,chemoradiotherapy, previous liver resection, number oftumours, extent and method of resection, use of hae-mostatic agents and histological diagnosis and grade.Number of units of blood crossmatched and transfusedwere collected from the hospital’s blood bank database.Multivariate regression analysis was performed on sig-nificant factors on univariate analysis to determineindependent predictors of blood transfusion in theimmediate perioperative period (48 h).Results: 599 patients were identified with a median age of64 years and a male: female ratio of 8:5. In the periop-erative period patients were crossmatched a median of10 units blood. Ratio of units crossmatched: unitstransfused was 13:1. Fifteen percent of patients received ablood transfusion with a median transfusion of two units.Transfusion requirement varied by operation from 12%of patients undergoing hemihepatectomy or metastecto-my to 48% of patients undergoing trisectionectomy.Multivariate regression analysis identified seven inde-pendent factors predictive of transfusion requirement(Table 1).

Table 1. Independent predictors of blood transfusionrequirement.

Predictor P-valueOddsratio 95% CI

Coronary artery disease 0.009 2.769 1.287–5.960Preoperative biliary drainage < 0.001 6.120 2.276–16.452Previous liver resection 0.001 4.450 1.830–10.824Preoperative platelet count 0.002 1.005 1.002–1.008No. of segmentsresected: 3–4

0.507 1.300a 0.599–2.824

a) compared to 1–2 segments resectedNo. of segments resected: 5+ < 0.001 4.182a 1.874–9.328Hepatocellular carcinoma 0.03 2.443 1.092–5.464Preoperative haemoglobin< 10 g/dL

0.021 6.556b 1.325–32.440

b) compared to Hb > 12.5 g/dLPreoperative haemoglobin10–12.5 g/dL

0.041 2.019b 1.030–3.958

Conclusions: Results from this single centre study suggestthat major liver surgery may be safely performed withfewer crossmatched units of blood. Here we have iden-tified seven independent predictors of transfusionrequirement. These factors could be used as criteria forcross matching of blood, with group and save being asafe and more cost effective measure in certain groups ofpatients.

Abstracts 9


Abstracts 17 to 24Free Papers – Basic Science

Saturday, March 14, 2009 8:10–10:10 am

17

Targeting focal adhesion kinase inhibitspancreatic cancer growth and metastasis

J. B. STOKES, MD, R. W. TILGHMAN, PHD,E. DAN HERSHEY, MS, J. K. SLACK-DAVIS, PHD,J. THOMAS PARSONS, PHD and T. W. BAUER, MDUniversity of Virginia, Charlottesville, VA, USA

Background: Focal adhesion kinase (FAK) is a cyto-plasmic protein tyrosine kinase involved in the regulationof cellular signaling, migration, apoptosis, and cell cycleprogression. FAK has been shown to activate prolifera-tion and inhibit apoptosis in cancer; however, its role inpancreatic cancer is not well understood. We analyzedthe effects of PF-562,271, an inhibitor of FAK, on pan-creatic cancer growth, invasion and metastasis.Methods: The human pancreatic adenocarcinoma celllines L3.6pl and MPanc 96 were investigated in vitro,while MPanc 96 was used for in vivo investigation. FAKphosphorylation was assessed by Western blot analysis.Transwell migration assays were used to evaluate the roleof FAK inhibition in pancreatic cancer cell migrationand invasion after stimulation with growth factors andextracellular matrix proteins. To evaluate the effects ofFAK inhibition in vivo, we used an orthotopic mousemodel in which 33 mg/kg of PF-562,271 was adminis-tered twice daily by gastric lavage. Tumor volume wasevaluated by magnetic resonance imaging (MRI) andtumor size, retroperitoneal invasion, and metastasis wereevaluated at necropsy. PF-562,271 was generously pro-vided by Pfizer Inc.Results: In pancreatic cancer cell lines FAK phosphor-ylation was significantly reduced by PF-562 271 admin-istration exhibiting an IC50 of 0.1 uM. Stimulation byIGF-I resulted in a 7-fold increase (P < 0.05) in cellmigration which was inhibited 94% by pretreatment withPF-562 271. Coating of transwell migration chambermembranes with collagen-I increased migration 6-foldand this was inhibited 32% by PF-562 271. In micebearing orthotopic pancreatic tumors, PF-562,271 ther-apy (vs. control) significantly inhibited tumor growth byMRI volumetric analysis (58.75 mm3 vs. 127 mm3,P < 0.05) and by pathologic examination (325 mm3 vs.816 mm3, P < 0.05). Treatment with PF-562,271 alsoyielded significantly fewer abdominal metastases (29%

vs.100%, P < 0.05), less retroperitoneal invasion(0% vs. 83%, P < 0.05), and showed a trend towardsfewer liver metastasis (0% vs. 50%, P = 0.06), comparedto control.Conclusions: Our study demonstrates the important roleof FAK in pancreatic cancer growth, invasion andmetastasis. These findings support FAK as a potentialtarget for therapy in patients with pancreatic cancer.

18

The molecular mechanism of HIF-1independentVEGF expression in hepatocellular carcinomacell line

S. B. CHOI, MD, J. B. PARK, PHD, K. S. KIM, MD,PHD and T. J. SONG, MD, PHDYonsei University College of Medicine, Seoul, Republic ofKorea;Research Institute and Hospital, National CancerCenter, Ilsan, Republic of Korea; Korea University Collegeof Medicine, Ansan, Republic of Korea

Purpose: Hypoxia-inducible factor-1 (HIF-1) is a mastertranscription factor that plays a central role in hypoxicexpression of various genes. The aim of this study was toprovide molecular pathway of vascular endothelialgrowth factor (VEGF) expression of HIF-1 independentpathway in hepatocellular carcinoma cell line (Hep3B).Methods: HIF-1a, HIF-2a dominant negative lentiviralvector was introduced to decrease the expression of HIFin Hep3B cell line. Cells were incubated at 37�C undernormoxic and hypoxic condition. We performed VEGFELISA using supernatant, and Western Blotting todemonstrate the difference of protein expression innormoxic or hypoxic condition. To validate the HIF-1dependent or HIF-1 independent pathway, we treated thecells with PI3K inhibitor and Erk kinase inhibitor.Results: VEGF level was increased under the hypoxiccondition. HIF-1a protein expression was induced inHep3B cell line after 24 h of exposure of hypoxia. Weused siHIF-1a and siHIF-2a transfected Hep3B cell linewhich was incubated in the normoxic or hypoxic condi-tion. The production of the VEGF was done by theHIF-1 pathway. However significant portion of theVEGF was produced by HIF-1 independent pathway.We treated the vector, siHIF-1a and siHIF-2a transfectedcells with ERK inhibitor before incubating normoxic orhypoxic condition. The VEGF expression was not dif-ferent in each cell lines in hypoxic condition. Thereforethe regulation of VEGF expression was not influenced byERK pathway. We treated the vector, siHIF-1a andsiHIF-2a transfected cells with PI3K inhibitor beforeincubating normoxic or hypoxic condition. The PI3Kinhibitor decreased VEGF expression in the siHIF-1atransfected cells in the hypoxic condition. We treated thevector, siHIF-1a and siHIF-2a transfected cells withsiSP1 transient transfection before incubating normoxicor hypoxic condition. The siSP1 transient transfectiondecreased VEGF expression in the siHIF-1a transfectedcells in the hypoxic condition. Therefore the VEGFregulation of Hep3B cell line was mainly controlled by

160

140

120

100

80

60

40

20

0Week 0 Week 1 Week 2

Control

PF-562,271

Time

MRI Growth Curves

Tum

or

Volu

me

(mm

3 )

Week 3

Figure 1.


Akt/PI3K and SP1 pathway which was independent onHIF-1 in hypoxic condition.Conclusions: Under the hypoxic condition, VEGFexpression was controlled by HIF-1. However the VEGFexpression by Akt/PI3K and SP1 pathway is not medi-ated by HIF-1 in the Hep3B cell line.

19

Proteasome inhibition in combination therapy ofexperimental pancreatic cancer: in vitro andin vivo evaluation

N. AWASTHI, PHD, M. A. SCHWARZ, MD andR. E. SCHWARZ, MDUT Southwestern Medical Center, Dallas, TX, USA

Background: Use of targeted therapy to enhance clinicalbenefits of cytotoxic chemotherapy has failed to improveresponses in pancreatic ductal adenocarcinoma (PDAC).Proteasome inhibition (PI) with agents such as bortezo-mib (B, Velcade) has shown anticancer benefits throughincrease in proapoptotic mechanisms and cell cycle-re-lated antiproliferative effects. This and the possibility foran increased biologic activity of other agents after PIprovide the rationale for PI combination therapy. Wehave previously shown that endothelial monocyte acti-vating polypeptide II (EMAP II, E), an antiangiogeniccytokine, enhances combination treatment effects inexperimental PDAC through VEGF and integrin-relatedmechanisms. Testing E and B for combination benefitsin vivo and in vitro was the purpose of this study.Methods: In vitro studies used WST-1 proliferationassays with three human PDAC lines and HUVECsendothelial cells (ECs) in a dose escalation matrix.Human ASPC PDAC cells were used in a murine xeno-graft survival model. Twelve days after i.p. or s.c. injec-tion of 7.5x10E5 tumor cells, animals underwenttreatment with various combinations of E (80 ug/kg i.p.daily), B (0.8 mg/kg i.p. twice weekly), or gemcitabine(G, 100 mg/kg i.p. twice weekly) for maximally 30 days.The resulting group survival (n = 6/group) wascompared via logrank statistic.Results: In 48-h. assays, B showed strong activity againstvarious PDAC cells (IC50 range: 0.5-50 uM) and ECs(25 nM). In a 72-h. combination assay, IC50s againstASPC (in uM) were: B = 0.1, G = 10, E=not reached at> 50; addition of G and/or E to B at various doses did notaffect proliferation differently. In a 48-hr. combinationassay, IC50s against ECs (in uM) were: B = 0.025,G = 5, E = 10; addition of G or E (at their IC20)increased B toxicity by 20%, and the triple combination by42%, suggesting someadditive effects. After initiation of invivo PDAC therapy, median group survival (d) was: con-trol = 19, B = 18, E = 20, G = 28; of all three mono-therapy groups, only G had a significant survival impact(P = 0.02). All combination groups except B+E led toextended survival compared to monotherapies or control(P = 0.003). However, addition of B to E,G, or E+Gdidnot enhance survival (p=NS for all three comparisons).Conclusions: PI with bortezomib mediates strong anti-proliferative effects against PDACs and ECs in vitro, butfails to enhance survival. Although combinations aremore effective than monotherapies, the in vivo synergisticpotential is limited. The findings highlight both benefitsand challenges of combination therapy approaches, whilesuggesting limitations of cancer cell in vitro assay efficacypredictions.

20

Anti-inflammatory effects of the nigella sativaseed extract, thymoquinone, in pancreaticcancer cells

H. A. ARAFAT, MD, PHD, N. CHEHL, B.SC.,G. CHIPITSYNA, PHD, Q. GONG, MD andC. J. YEO, MDThomas Jefferson University, Philadelphia, PA, USA

Background: Both hereditary and sporadic forms ofchronic pancreatitis are associated with an increased riskof developing pancreatic ductal adenocarcinoma (PDA).Inflammation has been identified as a significant factor inthe development of solid tumor malignancies. Thymo-quinone (Tq), the major constituent of the Nigella sativaoil extract induces apoptosis and inhibits PDA cell pro-liferation. Tq also increases p21WAF1 expression, inhibitshistone deacetylase (HDAC) activity, and induces his-tone hyperacetylation. HDAC inhibitors have beenshown to ameliorate inflammation-associated cancer inseveral animal models.Objective: To evaluate the anti-inflammatory potentialof Tq in PDA cells in comparison to a specific HDACinhibitor, trichostatin A (TSA).Methods: PDA cells (AsPC-1, HS766T; MiaPaca) werecultured and treated with or without Tq (25–75 lM),with or without pre-treatment of TNF- a (30 nM). Theeffect of Tq on the expression of different proinflamma-tory cytokines and chemokines was analyzed by real timePCR. Luciferase-labeled promoter studies evaluated theeffect of Tq on the transcription of monocyte chemo-attractant protein-1 (MCP-1) and nuclear factor-jB(NF-jB). The effect of Tq on the endogenous and TNF-a-induced activation and nuclear translocation of NF-jBwas examined by ELISA and immunohistochemistry.Results: Within 6 h, Tq significantly and dose-depen-dently reduced PDA cell production of TNF-a (P <0.02), interleukin (IL-1b) (P < 0.02), IL-8 (P < 0.05),Cox-2 (P < 0.002), and MCP-1 (P < 0.005). There wasno reduction in interferon-c (IFN-c) in the same cultures.Within the same time period, TSA reduced the produc-tion of Cox-2 (P < 0.02) and MCP-1 (P < 0.05), buthad no effect on TNF-a, IL-8, or IL-1b. Tq, but notTSA, significantly and dose-dependently reduced theintrinsic activity of the MCP-1 promoter. Tq alsoinhibited the intrinsic and the TNF-a-mediated activa-tion of NF-jB in PDA cells and reduced the transport ofNF-jB from the cytosol to the nucleus.Conclusions: Our data demonstrate previously unde-scribed anti-inflammatory activities of Tq in PDA cells,which are paralleled by inhibition of NF-jB. Tq as anovel inhibitor of proinflammatory pathways provides apromising strategy that combines anti-inflammatory andproapoptotic modes of action.

21

A proteomic based platform for markerdiscovery in cholangiocarcinoma

G. K. BONNEY, R. CRAVEN, A. MELCHER,P. SELBY, R. BANKS and R. PRASADSt James’ University Hospital, Leeds, UK

Cholangiocarcinoma (CCA) is the second commonestprimary malignancy of the liver with a rising incidenceworldwide and a dismal prognosis. There remains a lack

Abstracts 11


of sensitive and specific biomarkers for this diseaseparticularly in distinguishing it from patients with benignand predisposing diseases such as Primary SclerosingCholangitis.Bile, directly draining the liver, may contain higherconcentrations of biomarkers than those found in thegeneral circulation, as these may be tumour derived, shedor secreted proteins. The proteomic analysis of bile indisease pathogenesis (i.e. malignancy and gallstone dis-ease), drug metabolism and biomarker discovery hasreceived recent interest. We have developed a reproduc-ible and accurate method of quantifying, desalting anddelipidating bile prior to proteomic analysis. Using thismethod we ran two parallel proteomic based approaches,with the aim of marker discover in CCA.The first, using Differential In Gel Electrophoresis(DIGE), was a study comparing bile from 4 groups ofpatients: healthy liver donors (n = 5), benign obstruc-tion (n = 5), Primary Sclerosing Cholangitis (n = 3)and CCA (n = 5). This study resulted in six proteins thatwere significantly differentially expressed between groups(downregulated n = 5, upregulated n = 1) which werethen identified by mass spectrometry. Downstreamvalidation of the upregulated protein was performed byWestern blotting, which confirmed a significant differ-ence in abundance of this protein between the fourgroups.The second method used a Shotgun based approach,namely GeLC-MS/MS. For this, a sample of bile from apatient with CCA was analysed and over 1200 proteinsidentified. In validations studies, by Western blotting,two of these biliary proteins were differentially upregu-lated in CCA when compared to other disease groups.Using a proteomic based approach we have identified 3differentially expressed proteins in the bile of patientswith CCA compared to normal, benign and predisposedpatient groups.

22

Microenvironment-induced gene expressionchanges in breast cancer liver metastases

S. TABARIÈS, PHD, Z. DONG, MD, F. PÉPIN,M. HALLETT, PHD, A. OMERGLU, MD,M. HASSANAIN, MD, P. METRAKOS, MD andP. M. SIEGEL, PHDMcGill University, Montreal, QC, Canada

Introduction: Breast cancer is the most common canceraffecting Canadian women and is the second leadingcause of cancer related deaths in these patients. Theacquisition of metastatic abilities by breast cancer cells isthe most deadly aspect of the disease. Upon dissemina-tion from the primary tumor, breast cancer cells displaypreferences for specific metastatic sites. The liver repre-sents the third most frequent site for breast cancermetastasis, following bones and lungs. Despite the evi-dence that hepatic metastases are associated with poorclinical outcome in breast cancer patients, little is knownabout the molecular mechanisms governing the spreadand growth of breast cancer cells within the liver.Methods: We have employed 4T1 murine mammarycarcinoma cells that were subjected to three rounds ofsplenic injection, which permitted the isolation of breastcancer cells that aggressively grow in the liver. LaserCapture Microdissection (LCM) was used to identifygene expression changes within breast cancer liver

metastatic cells that occur in situ in response to the livermicroenvironment.4T1 breast cancer cells located at themargin and the centre of several metastatic lesions havebeen individually sampled by LCM and the recoveredRNA subjected to linear amplification, labeling andhybridization to Agilent microarrays. To date, we haveanalyzed four matched sets of margin vs. core livermetastases samples. Moreover, we have complementedthe analyses of 4T1-derived lesions with LCM experi-ments performed on liver metastases isolated from breastcancer patients.Results: Among several candidate genes that were dif-ferentially expressed between the margin and core of theliver metastases, the gene encoding for the leptin receptorappeared to be up-regulated in the margin. The Leptin/LEPR axis has gained considerable attention as animportant regulator of breast cancer progression. Indeed,Leptin and LEPR are over-expressed in breast cancerscompared to normal breast epithelium and Leptin hasbeen shown to stimulate the proliferation of breast cancercells. We have validated LEPR expression in the4T1-derived liver lesions, and more importantly, observeLEPR positivity in multiple liver metastases obtainedfrom breast cancer patients.Conclusion: The identification and functional validationof candidate genes important for the ability of breastcancer cells will provide basic insights into the pathwaysrequired for breast cancer cells to metastasize to the liver.Our results suggest that the LEPR may play an impor-tant role in enabling breast cancer cells to metastasize tothe liver.

23

Validation of a novel, physiologic model ofexperimental acute pancreatitis in the mouse

N. J. ZYROMSKI, MD, T. E. WADE, MD,H. A. PITT, MD, S. WANG, MD andD. A. SWARTZ-BASILE, PHDIndiana University, Indianapolis, IN, USA

Background: Acute pancreatitis is a devastating diseasethat affects 240 000 Americans each year. No specifictreatment for acute pancreatitis currently exists, largelybecause its precise pathophysiology is poorly understood.Murine experimental models are attractive, as completeknowledge of the mouse genome permits precise geneticmanipulation. Unfortunately, current methods used toinduce acute pancreatitis in the mouse (cerulein hyper-stimulation, intravascular bile salt infusion, supraphysi-ologic arginine administration) are of questionableclinical relevance. Therefore, the aim of the current studywas to validate a recently reported murine model of acutepancreatitis that is more representative of the humandisease process.Methods: Twenty C57BL/6J and 11 CF-1 mice werestudied. Under general anesthesia, transduodenal cann-ulation of the pancreatic duct was accomplished with a30-gauge catheter, and 50 lL of 5% Sodium Taurocho-late (NaT) or 0.9 normal Sodium Chloride (NaCl) wasinfused. Mice were euthanized 24 h later. Three observersrated pancreatitis severity by light microscopic evalua-tion of H&E sections. A validated scale incorporatingdegree of edema, vacuolization, and inflammatory cellinfiltrate comprise the total pancreatitis score. Pancreatictissue concentration of the chemoattractant molecule

Abstracts12


monocyte chemoattractant protein-1 (MCP-1) and theproinflammatory cytokine interleukin-6 (IL-6) weredetermined by ELISA. ANOVA and Student’s t-test wereapplied where appropriate; P value < 0.05 was acceptedas statistically significant.Results: Thirteen mice (NaCl – 6; NaT – 7) survived for24 h. The total pancreatitis score was significantly greaterin mice undergoing retrograde pancreatic duct infusionof NaT (Table 1). Pancreata of mice infused with NaTdemonstrated significant necrosis, consistent with severeacute pancreatitis. Pancreatic concentrations of MCP-1and IL-6 are shown in the Table 1.

Table 1. Pancreatic concentrations of MCP-1 and IL-6.

Pancreatitisscore

MCP-1(pg/mg)

IL-6(pg/mg)

NaCl(n = 6)

1.2 ± 0.4 2350 ± 1386 452 ± 269

NaT(n = 7)

6.3 ± 1.2* 3633 ± 1853 2028 ± 1612*

*P < 0.05 vs. NaCl

Conclusions: Retrograde pancreatic duct infusion ofSodium Taurocholate induces severe acute pancreatitis inthe mouse. Though associated with a discrete learningcurve, this model is likely more representative humanpancreatitis pathophysiology, and therefore provides apowerful tool with which to elucidate clinically importantbasic mechanisms underlying the pathogenesis of acutepancreatitis.

24

The expression of interferon receptor alpha/betain human pancreatic cancer in nude mice isessential for tumor response to interferon alphatreatment

R. F. SAIDI, MD, A. W. AHAD, MD,I. NALBANTOGLU, MD and M. J. JACOBS, MDMassachusetts General Hospital, Boston, MA,USA;Providence Hospital, Southfield, MI,USA;Department of Pathology, St John Hospital, Detroit,MI, USA; Department of Surgery, Providence Hospital,Southfield, MI, USA

Introduction: Adjuvant interferon (IFN) therapy andchemoradiation status post pancreaticoduodenectomy

for patients with pancreatic cancer have renderedpromising results.The aim of this study was to evaluatethe in vivo effect of interferon alpha on human pan-creatic carcinoma implanted orthotopically into nudemice.Material and methods: Human pancreatic cancer celllines MiaPaCa-2 and Panc-1 were used. MiaPaCa-2 isknown to express the interferon alpha/beta receptor andPanc-1 cells do not. The cells were implanted into thepancreas of nude mice and treatment was initiated sevendays later. Regimen I consisted of intraperitoneal single-agent gemcitabine (125-mg/kg biweekly) and Regimen IIconsisted of IFN-alpha (10 000-units daily, subcutane-ously) and gemcitabine biweekly for 30 days. Animalswere sacrificed after 30 days or if they became moribund.Body weight was determined and the primary tumors inthe pancreas were excised, measured, and weighed. Visi-ble metastases or adjacent organ invasion were countedand processed for H&E staining. All macroscopicallyenlarged regional (celiac and para-aortal) lymph nodeswere harvested and the presence of metastatic disease wasconfirmed by histology.Results: The mice that were implanted with MiaPaCa-2cells showed a more dramatic response to Regimen IIwhen compared to Panc-1 implanted mice. The Mia-PaCa-2 group that was treated with Regimen II showedan 87% reduction in tumor volume compared to 50% inthe Panc-1 group treated with the same regimen(P < 0.001). Mice implanted with MiaPaCa-2 andtreated with Regimen II showed less metastasis, less localinvasion, and a longer survival compared to miceimplanted with Panc-1 treated with same regimen.Regimen II was more effective on MiaPaCa-2 comparedto Regimen I (p

Abstracts 25 to 31Free Papers – Transplant

Saturday, March 14, 2009 2:00–4:00 pm

25

Is the use of donors with high donor risk indicescost-effective in liver transplantation?

D. E. MOORE, MD, MPH, I. FEURER, PHD andC. W. PINSON, MD, MBAVanderbilt University, Nashville, TN, USA

Background: The ever widening gap between the avail-ability of donor organs and the number of liver trans-plant candidates requires the maximal use of allavailable grafts, including those with high donor riskindices (DRI). High DRI organs are classified as thosehaving DRIs greater than 2.0, while low DRI organshave DRIs less than 2.0. Organs with a DRI above 2.0account for about 10% of the donor pool. Recipientsreceiving high DRI organs have been shown to havetwice the length of stay and a two fold increase inhospital costs when compared to recipients of lowDRI organs. Unless these grafts are used judiciously,outcomes and transplant resource utilization will benegatively affected. The aim of this study was to eval-uate the cost-effectiveness of an organ allocation schemeusing high and low DRI organs compared to an allo-cation scheme using only low DRI organs. Additionally,the impact of these schemes on wait list mortality will beevaluated.Methods: A Markov-based decision analytic model wascreated to simulate outcomes for liver transplantation inrecipients of organs from low and high DRI donorsversus recipients of organs from low DRI donors. Asecond model was created to simulate outcomes that re-flect alterations in wait list mortality under both organallocation schemes. Baseline values and ranges weredetermined from the UNOS database (19 000 recipientstransplanted between 2002 and 2007) and Medicare costdata. Sensitivity analyses were conducted to test modelstrength and parameter variability.Results: Recipients of organs from low DRI donors hada three year survival of 75% versus 65% for recipients oforgans from high DRI donors. After transplantation,recipients of high DRI organs gained 5.1 QALYs(Quality Adjusted Life Years) at a cost of $75,000/QALYversus 5.9 QALYs at a cost of $52 000/QALY forrecipients of low DRI organs. However, the allocationscheme in which both low and high DRI organs wereused together proved more cost-effective when comparedto the scheme in which only low DRI organs were usedand there was a 25% increase in wait list mortality; 5.8QALYs at a cost of $54 000/QALY versus 5.6 QALYs ata cost of $59 000/QALY.Conclusions: High DRI donors provide a significantnumber of grafts. While these grafts may represent anincreased cost per individual transplant, the overallcontribution of these grafts to the donor pool and asubsequent reduction in wait list mortality make themcost-effective.

26

The development of a classification system forbiliary complications following orthotopic livertransplantation

A. NEVILLE, MD, M. BOUTROS, MD andJ. BARKUN, MDRoyal Victoria Hospital, McGill University, Montreal,QC, Canada

Introduction: Biliary tract complications remain a sig-nificant source of morbidity and mortality followingorthotopic liver transplantation (OLTx). Even thoughthese occur with an estimated incidence of 10–40% andmay lead to graft failure, there exists no universallyaccepted classification system. As such, descriptions ofbiliary complications in the literature often lack consis-tency and clarity which limits evaluation and comparisonof post-transplantation outcomes. We propose a struc-tured classification system for biliary complications fol-lowing choledocho-choledochal anastomosis (CCA) atOLTx. The classification is based on four major com-ponents: leaks (anastomotic or cystic duct, anastomoticdisruption), filling defects (stones/sludge/casts in theintrahepatic, recipient or donor common ducts), stric-tures (unilateral intrahepatic, common hepatic or anas-tomotic) and development of sclerosing cholangitis.Complications are identified and classified based on theirappearance on a contrast cholangiogram, mostly endo-scopic cholangiopancreatography (ERCP). The goal ofthis initial report is to ensure that the proposed classifi-cation is consistent and reliable.Methods: Patients having undergone OLTx with CCA atthe McGill University Health Center from 2004 to 2007and who had at least one contrast cholangiogram post-operatively were candidates for the study. ERCP filmswere assessed by two independent reviewers to determineinter-rater reliability and the kappa coefficient of agree-ment of classification of the complication.Results: One hundred and eight ERCP films werereviewed. Overall inter-rater reliability (percent agree-ment) among the reviewers was 88.5%. Agreement onspecific elements of the classification scheme (leak, stric-ture, filling defect) was similarly high. Agreement amongstudy reviewers was greater than agreement betweenstudy reviewers and the official radiographic report(77.9%), although this was not statistically significant.The kappa coefficient was 0.87 (95% confidence interval0.79-0.94).Conclusion: The proposed classification system demon-strates high reliability and would allow for consistentinterpretation and reporting of biliary complicationsfollowing OLTx. Validation of this classification systemis ongoing through correlation of complications with theend-points of patient survival, graft survival, number ofadditional interventions required, length of hospital stayand number of readmissions.


27

Regionalization of liver transplantation to highvolume centers is associated with diminishedvolume-outcomes relationship and underutiliza-tion of high volume centers by minority patients

E. T. TRACY, MD, E. M. AVIKI, T. N. PAPPAS, MD,B. H. COLLINS, MD, J. E. TUTTLE-NEWHALL,MD, C. E. MARROQUIN, MD, P. C. KUO, MD andJ. E. SCARBOROUGH, MDDuke University, Durham, NC, USA

Introduction: Our purpose was to determine the temporaltrends in liver transplant center volume as well as theeffect of those trends on both the volume-outcomesrelationship and the utilization of high volume centers byminority patients.Methods: We performed a retrospective analysis of35 374 adult orthotopic liver transplant proceduresincluded in the Scientific Registry of Transplant Recipi-ents (SRTR) for three consecutive 30-month time periodsbetween 1999 and 2006. Transplant centers were dividedinto five categories based on annual volume (very low-volume < 36 procedures per year, low volume = 37–55,moderate-volume 56–79, high-volume group = 80–129,very high volume > 130). One-month and 1-yearobserved-to-expected patient death ratios (provided bythe SRTR), were calculated for each time period.Comparisons between volume groups in each time periodwere made using chi square analysis. Trends in thedemographic characteristics were also assessed.Results: The percentage of liver transplants performed atlower volume centers decreased significantly from 51% inPeriod 1 to 33% in Period 3 (P < 0.0001), while thoseperformed at higher volume centers increased signifi-cantly from 27% in Period 1 to 44% in Period 3(P < 0.0001). In Period 1, the 1-month and 1-year O:Epatient death ratios at very low-volume centers (1-monthratio 1.33, 1-year ratio 1.18) were significantly higherthan the ratios at very high volume centers (1-monthratio 0.89, 1-year 0.93, P < 0.0001 at both 1-month and1-year). By Period 3, the disparity between very low andvery high volume centers was no longer statistically sig-nificant at either 1 month or 1 year. Meanwhile, inPeriod 1 the percentage of Black or Hispanic patients atvery high-volume centers (28.4%) was significantlygreater than at very low-volume centers (20.6%,P < 0.0001). By Period 3, the percentage of Black orHispanic patients at very high volume centers hadbecome significantly lower (18.1%) than at low volumecenters (22.7%, P = 0.0004).Conclusions: As the percentage of liver transplants per-formed at higher volume centers has increased over time,the volume-outcomes relationship has become lessprominent. However, regionalization of liver transplan-tation does appear to be associated with a growingunderutilization of high volume centers by minoritypatients.

28

Is magnetic resonance imaging superior tocomputed tomography for preoperativeevaluation of cystic lesions of the pancreas?

C. K. CHU, MD, A. E. MAZO, B.SC., C. A. STALEY,MD, J. M. SARMIENTO, MD, D. R. MARTIN, MD,W. E. TORRES, MD, N. V. ADSAY, MD andD. A. KOOBY, MDEmory University School of Medicine, Atlanta, GA, USA

Purpose: Cystic lesions of the pancreas (CLP) represent adiverse range of pathologies. Appropriate managementdepends on accurate diagnostic imaging. Althoughmagnetic resonance imaging (MRI) provides a theoreti-cal advantage in improved soft-tissue contrast, nodefinitive superiority in its diagnostic capability has beenestablished over computed tomography (CT) in evalua-tion of CLP.Methods: Database review identified patients undergoingsurgical management at our institution between 1/1/00and 12/31/07 for CLP as confirmed by postoperativepathologic analysis. Records of patients who alsounderwent same-institution preoperative CT and/or MRIwere selected for study. Diagnostic accuracy and confi-dence were rated on a 5-point scale (Table 1) for eachradiologic study as compared with histologic review.Mann-Whitney U and Chi-square tests were used tocompare scores for CT and MRI.Results: One hundred and forty-eight patients receivedoperative interventions for CLP. Of these, 96 (65%)underwent preoperative CT and/or MRI (43 CT only,41 MRI only, 12 both) at our institution with a total of108 radiologic studies. Histologically, 24 patients (25%)had serous cystadenomas; 23 (24%), intraductal papil-lary mucinous neoplasms (IPMN); 20 (21%), mucinouscystadenoma/adenocarcinomas; 19 (20%), pseudocysts; 7(7%), solid pseudopapillary tumors; 2 (2%), simplescysts; and 1 (1%), lymphoepithelial cyst. No significantdifference between diagnostic scores of CT and MRI wasdetected (median 3 vs. 2, P = 0.14). However, only43.6% of CT studies received scores of 1 or 2, while62.3% of MRI studies achieved the same (P = 0.05).There was a trend towards improved preoperative iden-tification of IPMN using MRI as evidenced by percent ofscore 1 (50% by MRI vs. 13% by CT, P = 0.075). Nosuch trend was observed among cystadenomas.Conclusions: MRI may provide greater diagnostic accu-racy and confidence in preoperative evaluation of cysticlesions of the pancreas. MRI is prudent when uncertaintyregarding diagnosis and management of CT-visualizedlesions arise. Additionally, cystic lesions suspicious forIPMN may be better evaluated for ductal involvementwith MRI than CT.

Abstracts 15


Table 1. Diagnostic scoring system used for CT andMRI.

Score

Degree of accuracy and confidence of radiologicreport results in comparison to surgicalpathology results

1 Definitive: correct answer proposed as single ortop diagnosis

2 Suggestive: mentions correct answer within anarrow range (3 or less) possible diagnoses ofequal likelihood

3 Somewhat suggestive: mentions correct answerwithin a wide range (greater than 3) of possiblediagnoses

4 Vague: does not mention possibility of correctdiagnosis anywhere within report

5 Contradictory: correct diagnosis not proposedas possibility, and top proposed diagnosisinaccurate or contradictory to correct diagnosis

29

Comparative performance of staging systemsfor early hepatocellular carcinoma

H. NATHAN, G. MENTHA, H. P. MARQUES,L. CAPUSSOTTI, P. MAJNO, L. ALDRIGHETTI,C. PULITANO, L. RUBBIA-BRANDT,N. RUSSOLILLO, B. PHILOSOPHE, E. BARROSO,A. FERRERO, M. A. CHOTI and T. M. PAWLIKThe Johns Hopkins University, Baltimore, MD, USA

Introduction: While several staging systems for hepato-cellular carcinoma (HCC) have been proposed for pa-tients (pts) with a wide spectrum of HCC disease, mostsurgical candidates in fact have early HCC. The ability ofstaging systems to discriminate prognosis after surgeryfor early HCC is unknown. We sought to assess theperformance of major HCC staging systems in a largecohort of pts with early HCC.Methods: Data on pts undergoing liver resection (LR) ortransplantation (LT) for early HCC (£ 5 cm, N0M0, nomajor vascular invasion) with well-compensated cirrhosiswere collected from an international group of sevenhepatobiliary centers. We evaluated five staging systems:American Joint Committee on Cancer (AJCC), Interna-tional Hepato-Pancreato-Biliary Assoc. (IHPBA), Japa-nese Integrated Staging (JIS), Cancer of the Liver ItalianProgram (CLIP), and Okuda. A previously proposedscoring system for early HCC was also evaluated; thissystem allots 1 point each for size > 2 cm, multifocality,and microscopic vascular invasion (mVI). The discrimi-native abilities of the systems were quantified via boot-strap-corrected concordance indices (c).Results: Of 366 eligible pts, 236 underwentLRand 130LT.Overall survival was 74% at 3 years and 53% at 5 years(median 63 months). While most pts had size > 2 cm(n = 278, 76%), a minority had multifocality (n = 47,13%)ormVI (n = 76, 21%).Thedistributionofpts amongstages was: AJCC: T1 (n = 250), T2 (n = 116); IHPBA:T1 (n = 67), T2 (n = 273), T3 (n = 26); JIS 0 (n = 51),1 (n = 249), 2 (n = 54), 3 (n = 12); CLIP: 0 (n = 206),1 (n = 81), 2 (n = 19), 3 (n = 2); Okuda: A (n = 69), B(n = 219), C (n = 3); early HCC scoring system: 0

(n = 58), 1 (n = 219), ‡ 2 (n = 89). Except for the AJCCsystem, all established staging systems performed poorly asjudged by the c-statistics (Table 1). The proposed scoringsystem for early HCC demonstrated the best prognosticdiscrimination of all the evaluated systems (Table 1).

Table 1. Survival by stage for five existing staging systemsand a proposed scoring system for early HCC.

5-year survival, by stage/score (%)

AJCC IHPBA JIS CLIP Okuda

Early HCCScoringSystem

T1/Score 0/B 59 59 64 52 45 67T2/Score 1/B 40 52 50 47 53 55Score 2/C No pts. 51 57 50 67 39Score 3 No pts. No pts. 47 100 No pts. No pts.c-Statistics 0.585 0.518 0.502 0.510 0.508 0.592

Conclusions: Most staging systems perform poorly whenused for prognostic stratification of pts with early HCC.Although the AJCC staging provides moderate discrim-inative ability, it only includes two groupings (T1 and T2)that constitute early HCC. A simple scoring system forearly HCC provides superior discriminative power thanexisting staging systems.

30

Liver retransplantation – a 25-year experience

V. SHANMUGAM, C. BHATI, R.MARUTHANAYAGAM, J. BUCKELS, D. F. MIRZAand S. R. BRAMHALLQueen Elizabeth Hospital, Birmingham, UK

Retransplantation (re-LT) is the only viable option forirreversible graft failure after primary transplantation.Due to the growing disparity between the number ofpeople listed for primary Liver transplantation andorgans available and its inferior outcomes when com-pared to primary transplantation, re-LT is certainlycontroversial. The aim of this study was to analyse theindications and outcomes of re-LT since the beginning ofour programme.Methods: The clinical and demographic data for thisretrospective study from 1982 to 2007 was collected fromour dedicated transplant database. During this period atotal of 2420 adult transplants were performed at ourcentre.Results: A total of 196 re-LT patients underwent 225transplantations (8.09%). The mean age was 42.99 years(SEM±0.09) with the median age of 44.85 years (16.4–68.3). The number of retransplants has gone down stea-dily due to the better techniques and newer immunesuppression. Out of 196 re-LT, 23 had second re-LT(0.95%) and 6 had their fourth re-LT grafts(0.24%).More then 1/3 in our study (37.6%) requiredregrafting within the 6 months of primary transplant.The most common indication was Hepatic arterythrombosis (31.5%), followed by chronic rejection(22.6%) and recurrent disease (14.2%). The patient sur-vival and graft survival are as shown in Table 1.Early retransplants (< 7 days) of the primary trans-plant had a worse outcome than late retransplants

Abstracts16


(P = 0.001). Patient survival after retransplantation inpatient with PNF and Haemorrhagic necrosis had pooroutcome in comparison to others (P = 0.008). MELDdid not have any impact on patient survival (P = 0.2)Conclusion: Long terms graft as well as patient survivalin patients with regraft populations is significantly lowerthen patients with primary grafts. Patient survival de-pends on the indication for retransplantation. Retrans-plants within 7 days had a poor outcome than the latetransplants. Higher MELD did not have any impact onpatient’s survival. Retransplantation is necessary but theclinical selection of recipient remains of paramountimportance.

Table 1. Patient (PS) and graft survival (PS).

First ReLTxPS GS

Second ReLTxPS GS

Third ReLTxPS GS

30 days 83% 82% 64% 59% 67% 60%90 days 73% 72% 59% 59% 24% 40%1 year 66% 65% 45% 45% 24% 20%3 years 61% 59% 40% 36% 0%5 years 57% 54% 40% 32% 0%10 years 47% 43% 25% 24% 0%Total noof patients

196 23 6

31

Initial experience using Hepatitis C positiverenal allografts in elderly Hepatitis C negativerecipients

T. R. FLOHR, MD, D. KEITH, MD, P. I. LOBO, MD,B. R. SWENSON, MD, H. BONATTI, MD,T. M. SCHMITT, MD, R. G. SAWYER, MD,T. L. PRUETT, MD and K. L. BRAYMAN, MD, PHDUniversity of Virginia, Charlottesville, VA, USA

Introduction: First year mortality for elderly end stagerenal disease (ESRD) patients is 1.62–2.53 fold higher ascompared to younger ESRD patients. Renal transplan-tation is an option that provides improved survival,

however, the lack of suitable organs, longer wait-timesand worsening co-morbidities are great impedances forthe elderly. Considering these difficulties, the use ofkidneys from donors with serology positive for HepatitisC virus (HCV) has been explored. We report our expe-rience with nine HCV negative elderly patients whomunderwent renal t

volume 11 supplement 1 march 2009 - ahpba...s. roayaie, md, d. bassi, s. hiotis, md, d. labow, md...

Documents