Surgical Contribution to Recurrence-Free Survival in Patients with...

SPHSAJ

Hcdw

DTSGTI

R2FNi(CHG8

©P

urgical Contribution to Recurrence-Free Survival inatients with Macrovascular–Invasion�Negativeepatocellular Carcinoma

hinji Tanaka, MD, PhD, FACS, Kaoru Mogushi, PhD, Mahmut Yasen, MD, PhD, Norio Noguchi, MD, PhD,tsushi Kudo, MD, PhD, Toshiaki Kurokawa, MD, PhD, Noriaki Nakamura, MD, PhD,

ohji Inazawa, MD, PhD, Hiroshi Tanaka, PhD, Shigeki Arii, MD, PhD

BACKGROUND: Macroscopic vascular invasion (MVI) is a well-known indicator of recurrence of hepatocellularcarcinoma (HCC) even after curative hepatectomy, but the clinicopathologic and molecularfeatures of the recurrence remain unclear in MVI-negative HCC.

STUDY DESIGN: Two hundred seven consecutive patients with confirmed primary MVI-negative HCC wereretrospectively assessed after curative resection, with special emphasis on the importance ofanatomically systematized hepatectomy. HCC tissues were also analyzed for genome-wide geneexpression profile of each tumor using a microarray technique.

RESULTS: Univariant analysis of HCC recurrence revealed multiple tumors (p � 0.001), moderate to poordifferentiation (p � 0.044), Child-Pugh B/C (p � 0.047), �-fetoprotein elevation (p � 0.007), andnonanatomic hepatectomy (p � 0.010) as risk factors. According to Cox hazard multivariantanalysis, multiple tumors (p � 0.002), �-fetoprotein elevation (p � 0.001), and nonanatomichepatectomy (p � 0.002) were identified as independent factors of the recurrence. In the recurrentcases after anatomic hepatectomy for HCC, local recurrence was significantly infrequent comparedwith those after nonanatomic hepatectomy (p � 0.001). Network expression analysis using cDNAmicroarray revealed distinct signaling pathways of epithelial-mesenchymal transitions are associatedwith recurrence after anatomically systematized hepatectomy.

CONCLUSIONS: Anatomically systematized hepatectomy might contribute to recurrence-free survival of HCC pa-tients of HCC without MVI. Local recurrence could be mostly averted by anatomic hepatectomy,although specific epithelial-mesenchymal transitions signaling might regulate the biologic aggres-siveness of HCC. (J Am Coll Surg 2009;208:368–374. © 2009 by the American College of
Surgeons)
htreccmsrhatmtHnt

epatocellular carcinoma (HCC) is one of the mostommon malignancies, accounting for nearly 1 millioneaths per year.1 The incidence is still increasing world-ide, even in the US, as a result of the high prevalence of

isclosure Information: Nothing to disclose.his work was supported by Special Coordination Funds for Promotingcience and Technology (Japan Science and Technology Agency), and arant-in-Aid from Ministry of Education, Culture, Sports, Science and,echnology of Japan. Dr Tanaka is a recipient of the Japan Cancer Societyncitement Award and Japan Society for the Promotion of Science Prize.

eceived September 18, 2008; Revised October 9, 2008; Accepted October9, 2008.rom the Department of Hepato-Biliary-Pancreatic Surgery (Tanaka, Yasen,oguchi, Kudo, Kurokawa, Nakamura, Arii), Information Center for Med-

cal Sciences (Tanaka, Mogushi, Yasen, Tanaka), Medical Research InstituteInazawa), Tokyo Medical and Dental University, Tokyo, Japan.orrespondence address: Shinji Tanaka, MD, PhD, FACS, Department ofepato-Biliary-Pancreatic Surgery, Tokyo Medical and Dental University,raduate School of Medicine, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-

r519, Japan. email: [email protected]

3682009 by the American College of Surgeons

ublished by Elsevier Inc.

epatitis virus infection. Although surgical resection ishe effective treatment modality for HCC, rapid recur-ence of the tumors remains frequent even after appar-ntly curative resection.2,3 Hepatic recurrence has beenlassified as intrahepatic metastasis and multicentric re-urrence, but longterm outcomes are affected mainly byetastatic recurrence.4,5 Because tumor vascular inva-

ion is regarded as a direct cause of metastatic recur-ence, evidence of tumor invasion in major portal andepatic veins is a known determinant of poor outcomesfter resection for HCC.6-8 In accordance with the archi-ecture of the portal and hepatic veins, anatomic seg-entectomy or lobectomy has been developed as one of

he feasible methods of surgical treatment.9,10 In cases ofCC without apparent macrovascular invasion (MVI),

onanatomically partial hepatectomy is often selected ashe alternative procedure for surgical treatment to avoid
isk of postoperative hepatic failure.11,12 To determine
ISSN 1072-7515/09/$36.00doi:10.1016/j.jamcollsurg.2008.10.031

mailto:[email protected]

trsf

aastsfNNmts

MPTaTivtbtMitpttRaphcspt

pswimmdFiasTrt

RmPsmwnCfp(Um

NmTmwptaUubaegapgtpvp

369Vol. 208, No. 3, March 2009 Tanaka et al Anatomic Hepatectomy for HCC without MVI

he importance of systematized hepatectomy in recur-ence of MVI-negative HCC, the tumor characteristicshould be analyzed by aspects of the clinicopathologicactors.

Genome-wide gene expression analysis by microarray offerssystematic approach to unfold comprehensive information

bout the transcription profiles.13 In addition, such studieshould potentially lead to development of novel molecular-argeting therapy of HCC.14 To identify the most prominentignaling pathways, we examined the network analysis of dif-erentially expressed genes, using Biomolecular Interactionetwork Database (BIND) and Pajeck Program for Largeetwork Analysis.15,16 In this study, the clinicopathologic andolecular features were evaluated with special emphasis on

he importance of anatomically systematized hepatectomy foruch HCC.

ETHODSatients and tissue sampleswo hundred sixty consecutive patients underwent initialnd curative hepatectomy for HCC from 2000 to 2006 atokyo Medical and Dental University Hospital. Written

nformed consent from these patients and institutional re-iew board approval were obtained. Preoperative evalua-ion including liver function and operative procedures haveeen described elsewhere.2,3-8 Indications for hepatic resec-ion and operative procedures were determined based on

akuuchi’s criteria.17 Preoperative imaging for tumor stag-ng included abdominal ultrasonography, CT, hepatic ar-eriography, indirect arterial portography, CT arteriogra-hy, CT arterial portography, and MRI. MVI indicatinghe presence of tumor thrombus in the major branches ofhe portal or hepatic vein were comprehensively evaluated.esected tissue was fixed in 10% formaldehyde solutionnd embedded in paraffin for histopathologic analysis byathologists. Evidence of tumor invasion into the portal/epatic veins was evaluated both macroscopically and mi-roscopically. MVI on gross examination of the resectedpecimen confirmed the preoperative diagnosis in all of theatients examined. On histopathologic examination of

Abbreviations and Acronyms

BIND � Biomolecular Interaction Network DatabaseEGFR � epidermal growth factor receptorEMT � epithelial-mesenchymal transitionsHCC � hepatocellular carcinomaMVI � macrovascular invasionTGF-� � transforming growth factor-�

he resected specimen, microscopic invasion indicated the t

resence of clusters of cancer cells floating in the vascularpace line by endothelial cells. Patients were followed upith assays of serum level of �-fetoprotein and protein

nduced by vitamin K absence or antagonists II everyonth, and with ultrasonography, CT, and MRI every 3onths. When tumor recurrence was suspected, precise

iagnostic imaging was performed using CT angiography.inally, the exact diagnosis of recurrence was made with themaging. The recurrent tumor arising in the same segments the initial tumor, or within 2 cm from the surgicaltump, was regarded as a “local” recurrence, as described byakayama and colleagues.18 The other intrahepatic recur-ence was considered a “distant” one. Mean observationime was 3.8 years.

NA isolation, cRNA preparation, andicroarray hybridization

rimary HCC specimens were obtained from surgically re-ected materials. Total RNA was extracted from tissue speci-ens using RNeasy kit (Qiagen). Integrity of obtained RNAas assessed using Agilent 2100 BioAnalyzer (Agilent Tech-ologies). All samples had RNA Integrity Number �5.0.ontaminant DNA was removed by digestion with RNase-

ree DNase (Qiagen). Using 2 �g total RNA, cRNA was pre-ared using one-cycle target labeling and control reagents kitAffymetrix). Hybridization and signal detection of HG-133 Plus 2.0 arrays (Affymetrix) was performed followinganufacturer’s instruction.

ormalization and statistical analysis oficroarray dataotal microarray datasets were normalized using robustultiarray average method under R 2.4.1 statistical soft-are together with BioConductor package, as describedreviously.14 Estimated gene expression levels were log2-ransformed, and 62 control probe sets were removed fordditional analysis. For each 54,613 probes on HG-133 Plus 2.0 arrays, fold-change values were calculatedsing ratio of geometric means of gene expression levelsetween the two groups. Wilcoxon rank sum test waslso performed to estimate the significance levels of genexpression differences between them. Then we selectedenes meeting criteria for both log2(fold-change) valuesnd p values of Wilcoxon rank sum test, as describedreviously.14 Hierarchical clustering with the selectedenes was performed on R software using Euclidean dis-ance and complete linkage method. For clustering, ex-ression data were standardized as z scores (mean 0 andariance 1) for each probe. Differentially expressedrobe sets were overlaid on a cellular pathway map using
he BIND database (http://bond.unleashedinformatics.
http://bond.unleashedinformatics.com/index.jsp

cs(

SSfe

tvuuhtw

oor; pic; p

370 Tanaka et al Anatomic Hepatectomy for HCC without MVI J Am Coll Surg

om/index.jsp).15 The resulting networks were repre-ented in graphic format using the Pajek softwarehttp://vlado.fmf.uni-lj.si/pub/networks/pajek/).16

tatistical analysistatistical comparisons of clinicopathologic characteristicsor significance were made using chi-square test or Fisher’s

Figure 1. Overall recurrence-free survival cuvascular invasion�negative hepatocellular cathe actual numbers of patients at risk (in italicdifferences of the recurrence-free survivals weChild A, thin � Child B/C; p � 0.047), �-femAU/mL; p � 0.007), tumor number (D) (bolddifferentiation (E) (bold � well, thin � mod/ptomy (F) (bold � anatomic, thin � nonanatom

xact test with a single degree of freedom, and Student’s m

-test was used to analyze differences between continuousalues. Cumulative patient survival rates were determinedsing the Kaplan-Meier method, and for comparisons wesed the log rank test. A p value �0.05 was considered toave statistical significance. To investigate factors that con-ributed to prediction ability of the aggressive recurrence,e performed multivariant analysis by logistic regression

of 207 patients with primary macroscopicma after curative hepatectomy (A). A table ofwn below each survival curve. The significanttected in Child-Pugh classification (B) (bold �tein (C) (bold � 100 mAU/mL, thin � 100ingle, thin � multiple; p � 0.001), histologic

� 0.044), and systematization of hepatec-� 0.010).

rvesrcino) shore detopro

� s

odel.

http://vlado.fmf.uni-lj.si/pub/networks/pajek/

http://bond.unleashedinformatics.com/index.jsp

RRvAaowar(vf0Pac(ap

RsToatarss

btocwctwfbadmrlhbts

DTbptrt

TaH

V

A

G

H

C

A

P

T

T

P

H

S

*AHb


ESULTSisk factors of recurrence in HCC withoutascular invasionmong the consecutive 260 patients who underwent initialnd curative hepatectomy for HCC from 2000 to 2006 atur hospital, the 207 patients with MVI-negative HCCere analyzed in this study; 28 patients at TNM stage 1, 88t stage II, 69 at stage III, and 22 at stage IV.19 Cumulativeecurrence-free survivals of these 207 patients were 57.8%3-year) and 35.3% (5-year), as shown in Figure 1A. Uni-ariant analysis of HCC recurrence was shown inTable 1. Riskactors of recurrence were revealed as multiple tumors (p �.001), moderate/poor differentiation (p � 0.044), Child-ugh B/C (p � 0.047), �-fetoprotein elevation (p � 0.007),nd nonanatomic hepatectomy (p � 0.010) (Fig. 1B–F). Ac-ording to Cox hazard multivariant analysis, multiple tumorsp � 0.001), �-fetoprotein elevation (p � 0.002), and non-natomic hepatectomy (p � 0.002) were identified as inde-endent factors for recurrence (Table 2).

ecurrence of HCC after anatomicallyystematized hepatectomyo analyze the differences in recurrence patterns, localizationf the recurrent tumors was compared between cases afternatomic hepatectomy and those after nonanatomic hepatec-omy. Local recurrence was detected in 27 of 45 recurrent casesfter the nonanatomic hepatectomy, but in only 7 of 47 recur-ent cases after anatomic hepatectomy (Table 3). There wasignificant infrequence of local recurrence after anatomicallyystematized hepatectomy (p � 0.001).

Recurrence after anatomic hepatectomy was evaluatedy network analysis of the genome-wide gene expression ofhe primary HCC tissues (Suppl. Table 1 and Table 2,nline only). According to network expression analysis onomparison between the recurrent and nonrecurrent casesithin 2 postoperative years (Fig. 2A), distinct signals, in-

luding activating transcription factor 2 downstream ofransforming growth factor-� (TGF-�)–SMAD pathways,ere activated in cases of recurrence. In addition, the dif-

erences between local and distant recurrence were assessedy network expression analysis using the BIND database15

nd Pajeck Program,16 as described elsewhere.20,21 Figure 2Bemonstrated the E2F1 pathways downstream of epider-al growth factor receptor (EGFR)-Src signaling might

egulate the biologic aggressiveness of HCC,22 potentiallyeading to local recurrence after anatomically systematizedepatectomy. Because both of the pathways have recentlyeen revealed as critical factors of epithelial-mesenchymalransitions (EMT) for cancer metastasis,23-25 the specific EMT
ignaling might regulate the biologic aggressiveness of HCC.26 b
ISCUSSIONhe concept of anatomically systematized hepatectomy haseen proposed on the basis of hepatic sectors along Glisson’sedicles.9,10 The systematized hepatectomy has been reportedo substantially improve longterm outcomes,17,27-29 but severaleports found that major hepatectomy does not affect long-erm outcomes.30,31 Because patient survival is affected mainly

able 1. Univariate Analysis on Risk Factors of Recurrencefter Resection for Macroscopic Vascular Invasion�Negativeepatocellular Carcinoma

ariables of primary HCC n

Recurrenceprobability(3-year) p Value*

ge (y) 0.70465 or younger 78 0.609Older than 65 129 0.573

ender 0.695Male 150 0.586Female 57 0.667epatitis virus 0.508HBV 33 0.529HCV 107 0.633NBNC 67 0.558

hild-Pugh 0.047A 186 0.539B/C 21 0.716

FP (mAU/mL) 0.007�100 67 0.522�100 140 0.732

IVKA-II (mAU/mL) 0.146�100 109 0.494�100 98 0.578

umor number �0.001Single 66 0.497Multiple 141 0.789

umor size (cm) 0.714�3 65 0.555�3 142 0.609

athologic vascular invasion 0.281Negative 144 0.556Positive 63 0.634istologic differentiation 0.044Well 43 0.483Moderate/poor 164 0.621

ystematization of hepatectomy 0.010Anatomic 128 0.484Nonanatomic 79 0.741

Log rank test.FP, �-fetoprotein; HBV, hepatitis B virus; HCC, hepatocellular carcinoma;CV, hepatitis C virus; NBNC, non-B non-C; PIVKA-II, protein induced

y vitamin K absence or antagonists-II.

y metastatic recurrence through tumor vascular invasion,4

sdttiMItHa

oTtktppirsro

canidcAytHstcSt

wiittchcttp

mtfwAustkdttrptp

Hawptiov

TcI

VH

CAMM

N

A

TaVSh

ANT

TwtT*


urgical procedures and outcomes should be analyzed depen-ent on the status of the vascular invasion. Because the exis-ence of apparent MVI obviously requires anatomic hepatec-omy for HCC, systematization of resection should be arguedn the patients without MVI. We focused on the cases of

VI-negative HCC to determine the effects of hepatectomy.n our study, the anatomic hepatectomy contributed substan-ially to recurrence-free survival of patients with MVI-negativeCC (Fig. 1F), not only by univariate analysis (Table 1), but

lso by multivariate analysis (Table 2).Next, the recurrence patterns were assessed after anatomic

r nonanatomic hepatectomy. Local recurrence, as shown inable 3, was significantly suppressed by anatomic hepatec-omy, compared with nonanatomic cases (p � 0.001). Na-ashima and colleagues32 reported the histopathologically de-ailed analysis of 219 small HCCs without MVI, and foundathologic invasion in the portal veins in �25% and intrahe-atic micrometastasis in 10% within the sector.32 These find-ngs suggested that metastatic foci within the sector cannot beemoved with nonanatomically partial hepatic resection. Theystematized anatomic hepatectomy might contribute toecurrence-free survivals, possibly as a result of the inhibitionf local recurrence potentials of HCC.

It is open to debate whether the biologic ability of the re-urrence is identified in the primary MVI-negative HCC evenfter anatomic hepatectomy. In this study, we assessed theetwork expression analysis using cDNA microarray data us-

ng the BIND database15 and Pajeck Program.16 Signal trans-uction pathways play a key role in the regulation of keyellular processes of cancer, including invasion and metastasis.s demonstrated for recurrent cases within 2 postoperativeears (Fig. 2A),TGF-�-SMADs�activating transcription fac-or 2 signaling pathways were upregulated in the primaryCC tumors. TGF-� is a polypeptides with dual tumor-

uppressive and oncogenic effects, signaling through serine/hreonine kinase receptor complexes, which phosphorylateytoplasmic mediators, the SMADs.33 On phosphorylation,MADs translocate to the nucleus and regulate the transcrip-

able 2. Cox Multivariable Analysis on Risk Factors of Re-urrence after Curative Resection for Macroscopic Vascularnvasion�Negative Hepatocellular Carcinoma

ariables of primaryCC Coefficient

Oddsratio

95%Confidence

intervalp

Value

hild-Pugh B, C 0.622 1.863 0.954–3.640 0.069FP �100 mAU/mL 0.723 2.299 1.449–3.648 0.002ultiple tumors 0.833 2.061 1.300–3.266 �0.001oderete or poordifferentiation 0.657 1.928 0.688–5.404 0.212onanatomic 0.686 1.986 1.275–3.095 0.002

FP, �-fetoprotein; HCC, hepatocellular carcinoma.

ional factors, including activating transcription factor 2, o

hich mediates both transcription and DNA damage controln the malignant tumor development.34 During initial tumor-genesis, malignantly transformed cells often lose the responseo tumor-suppressive effects of TGF-�, which, in turn, startso act as an autocrine tumor-promoting factor by enhancingancer invasion and metastasis.33,35 More important, TGF-�as been noted as one of the main inducers of EMT,23 a pro-ess to convert epithelial cells into mesenchymal cells and con-rol cell adhesion, motility invasion, survival, and differentia-ion.26 Growing evidence points to changes in TGF-� signalingathway that occur during HCC progression at the late stage.36

Hepatic recurrence of HCC is classified as intrahepaticetastasis and multicentric recurrence.4,5 To distinguish

he recurrence patterns after anatomic hepatectomy,6,7 dif-erences between local and distant intrahepatic recurrenceere evaluated by network expression analysis (Fig. 2B).ccordingly, the EGFR-Src-E2F1 signaling pathways werepregulated in a cluster associated with the locally meta-tatic recurrence after anatomically systematized hepatec-omy. It is of interest that EGFR-Src signaling is alsonown as a positive regulator of EMT.26 Additional evi-ences indicated that E2F1 transcription factor, the majorarget of tumor suppressor Rb, might stimulate the EMThrough Zeb1/ZFHX1A,24,25 Snail,24 or Slug repressors,25

eported to advance the metastatic potential of HCC in ourrevious studies.37 On the basis of our investigations, newherapeutic strategies targeting the EMT signals might beroposed for inhibiting HCC recurrence.In conclusion, the surgical contribution to MVI-negativeCC might be emphasized because anatomically system-

tized hepatectomy could suppress the local recurrence. Net-ork gene expression analysis revealed the distinct signalingathways of EMT are associated with recurrence after ana-omically systematized hepatectomy. Limitations of our stud-es are mainly related to the retrospective analysis and shortbservation time. Randomized trials and much longer obser-ation time are required for proof of the critical impact of

able 3. Recurrence Patterns of Hepatocellular Carcinomafter Anatomic or Nonanatomic Hepatectomy for Macroscopicascular Invasion–Negative Hepatocellular Carcinomaystematization ofepatectomy

Localrecurrence

Distantrecurrence Total

natomic 7* 40 47onanatomic 27* 18 45otal 34 58 92

he recurrent tumor arising in the same segment as the initial tumor, orithin 2 cm from the surgical stump was regarded a “local” recurrence, while

he other intrahepatic recurrence was named a “distant” one, as described byakayama and colleagues.18

Chi-square test; p � 0.001.

ncological profiling. Additional attention should be paid to

tf

ASA

A

DC

R

locaial-me


he novel therapeutic strategies targeting the signal networkor adjuvant therapy of HCC.38,39

uthor Contributionstudy conception and design: S Tanaka, Ariicquisition of data: S Tanaka, Mogushi, Yasen, Noguchi,Kudo, Kurokawa, Nakamura

Figure 2. Hierarchical clustering (left) and biomolecular interactiohepatocellular carcinoma tumors. (A) The recurrent cases (red bar)systematized hepatectomy. Red and green nodes represent upreguIn the network panel using Pajek software,16 blue lines indicate tinteractions using Biomolecular Interaction Network Database.15

signaling pathways are upregulated in a cluster associated with erecurrence cases (red bar) and distant recurrence cases (black barepresent upregulated and downregulated expression in the local rsignaling pathways are upregulated in a cluster associated with thetomy. Note that both of the pathways are closely related to epithel

nalysis and interpretation of data: S Tanaka, Arii

rafting of manuscript: S Tanakaritical revision: Inazawa, H Tanaka, Arii

EFERENCES

1. Ince N, Wands JR. The increasing incidence of hepatocellularcarcinoma. N Engl J Med 1999;340:798–799.

tworks (right) on the basis of cDNA microarray data of primaryhe nonrecurrent cases (black bar) within 2 years after anatomicallyand downregulated expression in the recurrent ones, respectively.otein�protein interactions, and arrows indicate the protein�DNAforming growth factor-�–SMADs–activating transcription factor 2currence after anatomically systematized hepatectomy. (B) Local

er anatomically systematized hepatectomy. Red and green nodesence ones, respectively. Epithelial growth factor receptor-Src-E2F1lly metastatic recurrence after anatomically systematized hepatec-senchymal transitions.

n neand tlatedhe prTransarly rer) aftecurr

2. Arii S, Yamaoka Y, Futagawa S, et al. Results of surgical and non-

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3


surgical treatment for small-sized hepatocellular carcinomas: aretrospective and nationwide survey in Japan. The Liver CancerStudy Group of Japan. Hepatology 2000;32:1224–1229.

3. Tanaka S, Noguchi N, Ochiai T, et al. Outcomes and recurrenceof initially resection of hepatocellular carcinoma meeting Milancriteria: rationale for partial hepatectomy as first strategy. J AmColl Surg 2007;204:1–6.

4. Kumada T, Nakano S, Takeda I, et al. Patterns of recurrence afterinitial treatment in patients with small hepatocellular carci-noma. Hepatology 1997;25:87–92.

5. Shimada M, Takenaka K, Gion T, et al. Prognosis of recurrenthepatocellular carcinoma: a 10-year surgical experience in Japan.Gastroenterology 1996;111:720–726.

6. Arii S, Tanaka J, Yamazoe Y, et al. Predictive factors for intra-heptic recurrence of hepatocellular carcinoma after partial hep-atectomy. Cancer 1992;69:913–919.

7. Poon RT, Fan ST, Lo CM, et al. Intrahepatic recurrence after cur-ative resection of hepatocellular carcinoma: long-term results oftreatment and prognostic factors. Ann Surg 1999;229:216–222.

8. Ohkubo T, Yamamoto J, Sugawara Y, et al. Surgical results forhepatocellular carcinoma with macroscopic portal vein tumorthrombosis. J Am Coll Surg 2000;191:657–660.

9. Takasaki K. Glissonean pedicle transection method for hepaticresection: a new concept of liver segmentation. J HepatobiliaryPancreat Surg 1998;5:286–291.

0. Machado MA, Herman P, Machado MC. A standardized tech-nique for right segmental liver resections. Arch Surg 2003;138:918–920.

1. Fong Y, Sun RL, Jarnagin W, Blumgart LH. An analysis of 412cases of hepatocellular carcinoma at a western center. Ann Surg1999;229:790–799.

2. Fan ST, Ng IOL, Poon RTP, et al. Hepatectomy for hepatocel-lular carcinoma. The surgeon’s role in long-term survival. ArchSurg 1999;134:1124–1130.

3. Iizuka N, Oka M, Yamada-Okabe H, et al. Oligonucleotidemicroarray for prediction of early intrahepatic recurrence ofhepatocellular carcinoma after curative resection. Lancet 2003;361:923–929.

4. Tanaka S, Arii S, Yasen M, et al. Aurora kinase B is a predictivefactor for aggressive recurrence of hepatocellular carcinoma aftercurative hepatectomy. Br J Surg 2008;95:611–619.

5. Bader GD, Donaldson I, Wolting C, et al. BIND—the biomolec-ular interaction network database. Nucleic Acids Res 2001;29:242–245.

6. Batagelj V, Mrvar A. Graph drawing software: Pajek-analysis andvisualization of large networks. In: Junger M, Mutzel P, eds. Math-ematics and visualization. Berlin: Springer; 2004:77–103.

7. Hasegawa K, Kokudo N, Imamura H, et al. Prognostic impactof anatomic resection for hepatocellular carcinoma. Ann Surg2005;242:252–259.

8. Takayama T, Makuuchi M, Hirohashi S, et al. Early hepatocel-lular carcinoma as an entity with a high rate of surgical cure.Hepatology 1998;28:1241–1246.

9. Liver Cancer Study Group of Japan. The general rules for theclinical and pathological study of primary liver cancer. 2nd En-glish ed. Tokyo, Japan: Kanehara & Co., Ltd.; 2003.

0. Franke L, van Bakel H, Fokkens L, et al. Reconstruction of afunctional human gene network, with an application for prior-itizing positional candidate genes. Am J Hum Genet 2006;78:
1011–1025.
1. Woo HG, Park ES, Cheon JH, et al. Gene expression-based recur-rence prediction of hepatitis B virus-related human hepatocellularcarcinoma. Clin Cancer Res 2008;14:2056–2064.

2. Conner EA, Lemmer ER, Omori M, et al. Dual functions ofE2F-1 in a transgenic mouse model of liver carcinogenesis. On-cogene 2000;19:5054–5062.

3. Zavadil J, Böttinger EP. TGF-beta and epithelial-to-mesenchymaltransitions. Oncogene 2005;24:5764–5774.

4. Liu Y, Costantino ME, Montoya-Durango D, et al. The zincfinger transcription factor ZFHX1A is linked to cell prolifera-tion by Rb-E2F1. Biochem J 2007;408:79–85.

5. Arima Y, Inoue Y, Shibata T, et al. Rb depletion results in dereg-ulation of E-cadherin and induction of cellular phenotypicchanges that are characteristic of the epithelial-to-mesenchymaltransition. Cancer Res 2008;68:5104–5112.

6. Thiery JP. Epithelial-mesenchymal transitions in tumour pro-gression. Nat Rev Cancer 2002;2:442–454.

7. Yamamoto M, Takasaki K, Ohtsubo T, et al. Effectiveness of sys-tematized hepatectomy with Glisson’s pedicle transection at thehepatic hilus for small nodular hepatocellular carcinoma: retrospec-tive analysis. Surgery 2001;130:443–448.

8. WakaiT, Shirai Y, Sakata J, et al. Anatomic resection independentlyimproves long-term survival in patients with T1-T2 hepatocellularcarcinoma. Ann Surg Oncol 2007;14:1356–1365.

9. Eguchi S, Kanematsu T, Arii S, et al. Comparison of the out-comes between an anatomical subsegmentectomy and a non-anatomical minor hepatectomy for single hepatocellular carci-nomas based on a Japanese nationwide survey. Surgery 2008;143:469–475.

0. Kaibori M, Matsui Y, Hijikawa T, et al. Comparison of limitedand anatomic hepatic resection for hepatocellular carcinomawith hepatitis C. Surgery 2006;139:385–394.

1. Yamashita Y, Taketomi A, Itoh S, et al. Longterm favorableresults of limited hepatic resections for patients with hepatocel-lular carcinoma: 20 years of experience. J Am Coll Surg 2007;205:19–26.

2. Nakashima Y, Nakashima O, Tanaka M, et al. Portal vein invasionand intrahepatic micrometastasis in small hepatocellular carcinomaby gross type. Hepatol Res 2003;26:142–147.

3. Massagué J, Seoane J, Wotton D. Smad transcription factors.Genes Dev 2005;19:2783–2810.

4. SanoY, Harada J,Tashiro S, et al. ATF-2 is a common nuclear targetof Smad and TAK1 pathways in transforming growth factor-betasignaling. J Biol Chem 1999;274:8949–8957.

5. Leivonen SK, Kähäri VM. Transforming growth factor-beta sig-naling in cancer invasion and metastasis. Int J Cancer 2007;121:2119–2124.

6. Fransvea E, Angelotti U, Antonaci S, Giannelli G. Blockingtransforming growth factor-beta up-regulates E-cadherin andreduces migration and invasion of hepatocellular carcinomacells. Hepatology 2008;47:1557–1566.

7. Sugimachi K, Tanaka S, Kameyama T, et al. Transcriptionalrepressor snail and progression of human hepatocellular carci-noma. Clin Cancer Res 2003;9:2657–2664.

8. Thomas MB, Abbruzzese JL. Opportunities for targeted therapiesin hepatocellular carcinoma. J Clin Oncol 2005;23:8093–8108.

9. Tanaka S, Sugimachi K, Maehara S, et al. Oncogenic signaltransduction and therapeutic strategy for hepatocellular carci-
noma. Surgery 2002;131:S142–S147.

Sh

374.e1Vol. 208, No. 3, March 2009 Tanaka et al Anatomic Hepatectomy for HCC without MVI

upplementary Table 1. List of genes differentially expressed in cases of the recurrence after anatomically systematizedepatectomy for MVI-negative HCC

(continued)

S

374.e2 Tanaka et al Anatomic Hepatectomy for HCC without MVI J Am Coll Surg

upplementary Table 1. Continued

(continued)

S



(continued)

S



S



(continued)

S



Sa


upplementary Table 2. List of genes differentially expressed in cases of the local recurrence after anatomically system-tized hepatectomy for MVI-negative HCC.

(continued)

S



(continued)

S



Surgical Contribution to Recurrence-Free Survival in Patients with...

Documents

Transcript of Surgical Contribution to Recurrence-Free Survival in Patients with...