Research Article

Partial hepatectomy vs. transcatheter arterial chemoembolizationfor resectable multiple hepatocellular carcinoma beyond

Milan criteria: A RCT

Lei Yin1, Hui Li2,�, Ai-Jun Li1,�, Wan Yee Lau1,3, Ze-ya Pan1, Eric C.H. Lai1,3, Meng-chao Wu1,Wei-Ping Zhou1,⇑

1The Third Depart ment of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital, Second Military Medical University, Shanghai 200438, China;2The Department of Dermatology, Changhai Hospital, Second Military Medical University, Shanghai 200438, China; 3Faculty of Medicine,

The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong Special Administrative Region

See Editorial, pages 3–4

Background & Aims: The aim of this randomized comparativetrial (RCT) is to compare partial hepatectomy (PH) with trans-catheter arterial chemoembolization (TACE) to treat patients withresectable multiple hepatocellular carcinoma (RMHCC) outside ofMilan Criteria.Methods: This RCT was conducted on 173 patients with RMHCCoutside of Milan Criteria (a solitary tumor up to 5 cm or multipletumors up to 3 in number and up to 3 cm for each tumor) whowere treated in our centre from November 2008 to September2010. The patients were randomly assigned to the PH group orthe TACE group. The primary outcome measure was overallsurvival (OS) from the date of treatment. A multivariate Coxproportional hazards regression analysis was performed to assessthe prognostic risk factors associated with OS.Results: The 1-, 2-, and 3-year OS rates were 76.1%, 63.5%, and51.5%, respectively, for the PH group compared with 51.8%,34.8%, and 18.1%, respectively, for the TACE group (Log-rank test,v2 = 24.246, p <0.001). Multivariate Cox proportional hazards

Journal of Hepatology 20

Keywords: Multiple hepatocellular carcinoma; Partial hepatectomy; Transcath-eter arterial chemoembolization; Overall survival; RCT.Received 4 August 2013; received in revised form 10 February 2014; accepted 10March 2014; available online 17 March 2014q DOI of original article: http://dx.doi.org/10.1016/j.jhep.2014.04.005.⇑ Corresponding author. Tel.: +86 21 81875521/65341828; fax: +86 2181875529.E-mail address: EHPHWP@126.com (W.-P. Zhou).

� These authors contributed equally to this work.Abbreviations: RCT, randomized comparative trial; PH, partial hepatectomy; HCC,hepatocellular carcinoma; RMHCC, resectable multiple hepatocellular carcinoma;TACE, transcatheter arterial chemoembolization; OS, overall survival; LT, livertransplantation; BCLC, Barcelona Clinic Liver Cancer; IEC, Institutional EthicsCommittee; CTMB, Clinical Trial Monitoring Board; USG, ultrasonography; CT,computed tomography; MRI, magnetic resonance imaging; HBsAg, hepatitis Bsurface antigen; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; DNA, deo-xyribonucleic acid; AFP, serum alpha-fetoprotein; CEA, carcinoembryonic anti-gen; CA 19-9, carbohydrate antigen 19-9; AST, aspartate aminotransferase; ALT,alanine aminotransferase; IOUS, intraoperative ultrasonography; mRECIST, Mod-ified Response Evaluation Criteria in Solid Tumors; CR, complete response; PR,partial response; PD, progressive disease; SD, stable disease; PRFA, percutaneousradiofrequency ablation; PEI, percutaneous ethanol injection.

regression analysis revealed the type of treatment (hazard ratio,0.434; 95% CI, 0.293 to 0.644, p <0.001), number of tumor (hazardratio, 1.758; 95% CI, 1.213 to 2.548, p = 0.003) and gender (hazardratio, 0.451; 95% CI, 0.236 to 0.862, p = 0.016) were significantindependent risk factors associated with OS.Conclusions: PH provided better OS for patients with RMHCCoutside of Milan Criteria than conventional TACE. The numberof tumor and gender were also independent risk factors associ-ated with OS for RMHCC.� 2014 European Association for the Study of the Liver. Publishedby Elsevier B.V. All rights reserved.

Introduction

Hepatocellular carcinoma (HCC) is a common malignancy espe-cially in East-Asian countries [1]. Partial hepatectomy (PH) isthe most commonly used curative therapy for HCC [2]. Theresults of PH for small HCC are good [3]. However, its role formultiple HCC is less well-defined. Multiple HCC have been shownto have a poor impact on survival after PH [4]. The use of MilanCriteria to select patients for liver transplantation (LT) producesgood results for a solitary HCC up to 5 cm or for multiple HCCup to 3 in number and up to 3 cm for each tumor [5]. In mostcenters, LT is not recommended for HCC beyond Milan Criteria[6,7].

In non-randomized studies, PH has the potential to improvethe quality and quantity of survival in patients with advancedT-stage but resectable HCC who have no extrahepatic metastasis[8]. PH can be followed by regional or systemic therapy to dealwith possible micro-metastases left in the liver remnant.

The Barcelona Clinic Liver Cancer (BCLC) recommendsmultiple HCC of BCLC B which exceeds Milan Criteria to receivetranscatheter arterial chemoembolization (TACE). However,retrospective studies suggested that PH produced better survivalthan TACE [9,10]. This randomized comparative trial (RCT) wasdesigned to compare PH vs. TACE in patients with resectablemultiple HCC (RMHCC) outside of Milan Criteria.

14 vol. 61 j 82–88

JOURNAL OF HEPATOLOGY

Patients and methods

Study design

Patient’s decision to participate in the study was voluntary and informed consentwas obtained from each patient that their data would be used for research pur-poses. The study protocol was approved by the Institutional Ethics Committee(IEC) of the Eastern Hepatobiliary Hospital. The Clinical Trial Monitoring Board(CTMB), which was affiliated with the IEC, monitored the whole process, and col-lected and stored data of the study. An interim analysis was carried out when halfof the patients were enrolled. The CTMB had the choice to continue or stop thestudy according to the outcomes of the interim analysis, but it decided that thestudy should be continued because there was no significant difference betweenthe 2 groups due to the short follow-up period.

The study was registered with the U.S. NIH Clinical Trials Registry at http://clinicaltrials.gov and the registration number is NCT00820157.

Participants

From November 2008 to September 2010, 2502 HCC patients were treated at TheThird Department of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital.Only patients who met the following inclusion criteria were enrolled into thisstudy: (1) good surgical risk patients >18 years and 670 years of age; (2) at leasttwo rounds of radiological imaging showing characteristic features of HCC, or oneradiological imaging associated with alpha fetoprotein (AFP) >400 lg/L, or cyto-logical/histological evidence of HCC [11]; (3) resectable HCC with tumors outsideof Milan Criteria; (4) adequate liver remnant size after liver resection; (5) noradiological evidence of vascular invasion or extrahepatic metastasis; (6) tumorsresectable with a single or multiple liver resections; (7) liver function status ofChild-Pugh A-B, with serum bilirubin 61.5 times the upper limit of normal, ala-nine aminotransferase and aspartate aminotransferase 62 times the upper limitof normal; (8) no major organ dysfunction; (9) hemoglobin P90 g/L, white bloodcell P3000 cells/mm3, platelets P80.000 cells/mm3; (10) informed consent hadbeen obtained. Patients were excluded if they met any of the following criteria:(1) cardiac, pulmonary, cerebral and renal dysfunction; (2) a history of othermalignancy; (3) extrahepatic metastasis, portal vein or other major vascularinvolvement; (4) liver functional status of Child-Pugh C; (5) any other contraindi-cation like: active gastrointestinal bleeding, refractory ascites, coagulopathy,severe portal hypertension; (6) no prior TACE or liver resection.

The study was investigator-designed and investigator-driven, and it receivedno support from any pharmaceutical companies.

Sample size

The sample size was computed using the overall survival (OS) as the primary end-point. According to published data [12], there was a difference of 20% (21.7% vs.4.7%) in 5-year survival rates between the 2 groups. Assuming a type-I error of5% (a = 0.05), a power of 80% for a 2-tailed log-rank test (b = 0.2), 59 patientsper group were needed. Considering possible post-randomization exclusion andloss to follow-up of patients, we decided to include 180 patients, with 90 patientsin each arm.

Randomization

The double-blind technique was not used here because this study was a clinicaltrial about two completely different therapies. The included patients (1:1 ratio)were randomly assigned to either the PH group or the TACE group after comple-tion of preoperative evaluation using random numbers generated from a com-puter by a research nurse who was not involved in this study.

Preoperative investigation

All patients had a chest X-ray, ultrasonography (USG) of abdomen, and contrastenhanced computer tomography (CT) or magnetic resonance imaging (MRI) ofabdomen. Laboratory blood tests including hepatitis B surface antigen (HBsAg)and its antibody, hepatitis B e antigen (HBeAg) and its antibody, hepatitis Bvirus-deoxyribonucleic acid (HBV-DNA), hepatitis C antibody, serum alpha-feto-protein (AFP), carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA19-9), serum albumin, serum total bilirubin, aspartate aminotransferase (AST),alanine aminotransferase (ALT), and prothrombin time were obtained, and theChild-Pugh’s score was determined.

Journal of Hepatology 20

Partial hepatectomy (PH)

Partial hepatectomy (PH) included single or multiple liver resections aiming toremove all macroscopic tumors [8,13]. The interval from randomization to PHwas within one week. PH was carried out through a bilateral subcostal incision.During surgery, the abdominal cavity was carefully searched for the extent oflocal disease, extrahepatic metastases, and peritoneal seeding. After mobilizationof the liver, intraoperative ultrasonography (IOUS) was performed to assess thenumber, size of the lesions, and the relation of the tumors to vascular structures.Pringle’s maneuver was applied to occlude the blood inflow of the liver. Theentire hepaticoduodenal ligament was encircled and tightened with a rubbertourniquet. Liver resection was carried out by a clamp crushing method. Whenthe liver was normal, the occlusive time was continuous for up to 30 min. Whenthe liver was cirrhotic, the occlusion was done using cycles of clamp/unclampintervals of 15/5 minutes.

Routine investigations were carried out at 4 weeks after PH and then onceevery 3 monthly thereafter. Patients with intrahepatic recurrence were treatedactively with re-resection, local ablation, TACE or conservative treatment depend-ing on the site of the tumor, the liver function, and the general condition of thepatient [14]. Sorafenib was given to patients with extrahepatic recurrence.

TACE

The interval from randomization to TACE was within one week. A vascular cath-eter was inserted through a femoral artery using the Seldinger technique to thehepatic artery, and hepatic angiography was carried out. The catheter tip wasinserted selectively into the tumor-feeding artery whenever technically possible.An emulsion of 5-fluorouracil (1 g), mitomycin C (20 mg), cisplatin (5 mg), andlipiodol 10 to 30 ml (1 to 2 ml/cm diameter of the tumor) was injected. Gelfoamfragments were then injected to embolise the tumor-feeding vessel. No antibioticprophylaxis was given.

The radiographic response was evaluated 4–6 weeks after TACE, according tothe Modified Response Evaluation Criteria in Solid Tumors (mRECIST) [15].

Complete response (CR) was defined as disappearance of any intra-tumoralarterial enhancement in all target lesions on CT or MRI after TACE. Partialresponse (PR) was defined as at least a 30% decrease in the sum of diameters ofthe viable (contrast enhancement in the arterial phase) target lesions, taking asreference the baseline sum of the diameters of the target lesions. Progressive dis-ease (PD) was defined as an increase of at least 20% in the sum of the diameters ofthe viable (enhancing) target lesions, taking as reference the smallest sum of thediameters of the viable (enhancing) target lesions recorded since the treatmentstarted. Stable disease (SD) was defined as any cases that do not qualify for eitherPR or PD.

LT or PH was recommended to the patients with CR, depending on the generalcondition of the patient and the liver functional status of the patients. Patientswith PR or SD were treated with repeat TACE and/or local ablative treatment ifthe liver function was Child A or B. TACE was stopped for patients with PD or withChild C liver functional status. Adverse events related to TACE were gradedaccording to the National Cancer Institute Common Terminology Criteria forAdverse Events v4.0.

Follow-up

All patients were followed-up by the same team of surgeons. Routine investiga-tions included serum AFP, liver function, complete blood count, ultrasonography,chest X-ray, and CT or MRI. Routine investigations were carried out at 4 weeksafter treatment and then once every 3 month thereafter.

Outcomes

The primary outcome measure was OS measured from the date of treatment.

Statistical methods

Continuous variables were expressed as mean (sd) or median (range), as appro-priate. Categorical variables were compared by the v2 test or Fisher exact test,and continuous variables by the student’s t test. p values of <0.05 were consideredas significant. The overall survival curves were constructed using the Kaplan-Meier method and compared with the log-rank test. The relative prognosticsignificance of the variables in predicting overall survival was assessed by univar-iate and multivariate Cox proportional hazards regression models. The statisticalanalyses of the data were performed using the SPSS 18.0 statistical software.

14 vol. 61 j 82–88 83

Allergic to iohexol (n = 5)

Lost to follow-up (n = 5)

Did not meet the inclusion criteria (n = 2296) Decline to participate (n = 26)

Patients within inclusion criteria

(n = 2502)

Randomized to TACE (n = 90)

Intention-to-treat analysis (n = 85)

Received TACE (n = 85)

Randomized to PH (n = 90)

Intention-to-treat analysis (n = 88)

Received PH (n = 88)

Patients enrolled in study

(n = 180)

Violation of study protocol

(n = 2)

Lost to follow-up (n = 2)

Fig. 1. Flow diagram.

Research Article

Results

Patients

Among the 2502 patients with a diagnosis of HCC who wereadmitted to our Department from November 2008 to September2010, 180 patients met the inclusion criteria and were random-ized to the two groups. Of 180 patients who were randomizedto the PH group and the TACE group, 2 patients in the PH groupand 5 patients in the TACE group were excluded from this studyafter randomization because of violation of study protocol orbecause of allergy to iohexol. The remaining 173 patients, includ-ing 88 patients in the PH group and 85 patients in the TACEgroup, were the subjects of the study (Fig. 1). These 173 patientswere included for the intention-to-treat analysis.

The baseline characteristics were well matched between the 2groups. 162 (93.6%) of these patients had viral hepatitis. Amongthem, 158 (91.3%) were positive for hepatitis B surface antigen,2 (1.2%) for hepatitis C antibody and 2 (1.2%) for both hepatitisB surface antigen and hepatitis C antibody. 143 (82.7%) patientswere diagnosed as HCC with cirrhosis (Table 1).

In the PH group, additional new lesions were diagnosed in theliver on intraoperative ultrasound in 5 patients. As these newlesions were close to the main tumor, the preoperatively plannedoperations could be carried out to include these lesions into theresected specimens. Histopathology showed R0 resection in 69patients and R1 resection in 19 patients. Complications in thePH group included perioperative bleeding (n = 2), liver failure(n = 1), bile leak (n = 4), and infection (n = 2). For the 62 (70.5%)patients who had recurrent HCC in the PH group, 5 patients

84 Journal of Hepatology 20

underwent reoperation, which included right hepatectomy(n = 3), segments resection (n = 1), and pulmonectomy (n = 1),50 patients received local ablation and 5 patients received sorafe-nib (Table 2).

For the 85 patients in the TACE group, the number of TACEsessions that the patients received were one session (n = 17), 2sessions (n = 7), 3 sessions (n = 24), 4 sessions (n = 19), 5 sessions(n = 9), 6 sessions (n = 6), 7 sessions (n = 1), and 8 sessions (n = 2).Thus, 68 patients received repeat TACE at an interval of one ses-sion of TACE per 2–3 months. The mean session of TACE was 3.3.Using the predetermined definition and at the time of censor ofthis study, 10 patients had CR, 11 had PR or SD, and 64 had PD.Of the 10 patients with CR, 2 patients received one session ofTACE while the remaining 8 patients received repeat TACE. 6 of10 patients with CR underwent liver resection. These 6 patientswere still included in the TACE group for analysis and the OSwas calculated from the date of the first TACE. When these 6patients were censored at the time of liver resection (OS calcu-lated from the first session of TACE to the date of liver resection),there was very little impact on the OS rate for the TACE group.When these 6 patients were taken away from the TACE groupfor analysis, the OS for the TACE group became worse. The oper-ations included left hepatectomy (n = 2), right hepatectomy(n = 2), and 2 segments resection (n = 2). The remaining 4patients refused surgery. For the other patients with PR, SD orPD, 58 patients received repeat TACE. In addition, 18 patientsreceived percutaneous radiofrequency ablation (PRFA), 28patients received percutaneous ethanol injection (PEI) and 20received sorafenib. There were no grade 4 or 5 adverse effectsrelated to TACE (Table 3).

14 vol. 61 j 82–88

Table 1. Clinical features.

Characteristics PH (n = 88) TACE (n = 85) p valueAge (years) (mean ± sd) 51.6 ± 9.0 54.0 ± 9.5 0.083Sex (male/female) 82/6 79/6 0.999HBsAg (positive/negative) 81/7 77/8 0.944HBeAg (positive/negative) 29/59 33/52 0.518HBV-DNA (positive/negative)* 63/25 53/32 0.258HCV Ab (positive/negative) 3/85 1/84 0.329Cirrhosis (positive/negative)Child-Pugh classification (A/B)

69/19 74/11 0.19387/1 80/5 0.088

AFP (ng/ml)† 0.801≥400 68 68<400 20 17

Hemoglobin (g/L) (mean ± sd) 143.3 ± 14.6 140.1 ± 814.8 0.155Platelet (x 109/L) (mean ± sd) 172.3 ± 69.6 154.7 ± 64.7 0.087Prothrombin time (s) (mean ± sd) 12.7 ± 6.4 12.6 ± 1.0 0.874Biggest tumor diameter (cm) (mean ± sd) 7.3 ± 2.5 7.4 ± 2.3 0.693Total tumor diameter (cm) (mean ± sd) 9.5 ± 3.0 10.4 ± 3.3 0.061Number of tumor†† 0.578

2 55 483 22 294 9 65 2 2

Distribution of tumor 0.230Right liver 51 42Left liver 6 3Both livers 31 40

⁄HBV-DNA (positive) was defined as beyond 1 � 103 IU/ml.�AFP, alpha-fetoprotein.��Number of preoperatively-diagnosed tumors.

JOURNAL OF HEPATOLOGY

Follow-up

This study was censored on April 30, 2013. The median follow-upperiod was 33.3 months (95% CI, 28.1 to 53.8) for the PH groupand 13.5 months (95% CI, 9.5 to 18.4) for the TACE group, respec-tively. At the time of censor, 116 patients (67.1%) had died. In thePH group, 1 patient died of perioperative liver failure, and 46patients died of tumor-related death. In the TACE group, 70patients died of tumor-related death. 5 patients were lost to fol-low-up in the PH group and 2 patients in the TACE group. Therewere no significant differences in the 30-day and in the 90-daymortality after treatment between the 2 groups.

Survival

The 1-, 2-, and 3-year OS rates and median survival were 76.1%,63.5%, 51.5%, and 41 months (range 1–50 months) respectively,in the PH group. The corresponding figures for the TACE groupwere 51.8%, 34.8%, 18.1%, and 14 months (range 5–47 months),respectively. The PH group had significantly better OS than theTACE group (log-rank test, v2 = 24.246, p <0.001) (Fig. 2). TheOS rates were also significantly different between the PH groupand the TACE group for the patients with 2 tumors (log-rank test,v2 = 20.866, p <0.001) (Fig. 3) and those with more than 2 tumors(log-rank test, v2 = 4.591, p = 0.032) (Fig. 4), respectively.

Using univariate analysis, the factors that significantly influ-enced OS were the type of treatment, serum AFP level, total

Journal of Hepatology 20

tumor size, gender and number of tumor. Factors with a p <0.05and those variables with obvious clinical significances, includingthe type of treatment, number of tumor, gender, serum AFP level,total tumor size and Child-Pugh classification were selected formultivariate analysis. On multivariate analysis, the type oftreatment (hazard ratio, 0.434; 95% CI, 0.293 to 0.644,p <0.001), number of tumor (hazard ratio, 1.758; 95% CI, 1.213to 2.548, p = 0.003), and gender (hazard ratio, 0.451; 95% CI,0.236 to 0.862, p = 0.016) remained as significant factors of OS(Table 4).

Discussion

Despite recent advances in the treatment of HCC, the prognosis ofHCC remains poor, especially for patients who are in the interme-diate or late stages of the disease. Recently, sorafenib has beenfound to provide survival benefit for these patients [16,17].

Patients with multiple tumors have a poorer survival outcomethan those with a solitary tumor [4,18,19]. A recent report hasshown TACE to provide survival benefit in asymptomatic patientswith multiple HCC outside of Milan Criteria when compared withconservative treatment, especially for those patients without vas-cular invasion or extrahepatic spread who were not candidatesfor curative treatments [6]. Unfortunately, the evidence support-ing the role of TACE for these patients is still limited [7], and thereis no properly conducted RCT.

14 vol. 61 j 82–88 85

Table 2. Operative variables and perioperative outcomes of the partialhepatectomy group.

PH (n = 88)Operative blood loss (ml) (mean ± sd) 298.3 ± 183.6Perioperative blood transfusion (n) 11Operative time (min) (mean ± sd) 152.5 ± 40.6Type of resection (n)

Resection of 1 segment 6Resection of 2 segment 37Resection of 2 segment + WR* 3Resection of 3 segment 24Resection of 3 segment + WR 2Caudate lobe + 2 segment 1

Right hepatectomy 6Right hepatectomy + WR 1Extended right hepatectomy 1

Left hepatectomy 3Left hepatectomy + WR 2Extended left hepatectomy 2

Resection (n)R0 69R1 19

Complications (n)Bleeding 2Liver failure 1Bile leak 4Infection 2

Perioperative death (n) 1Recurrence (n) 62

Sites of recurrence (n)Intrahepatic 40Extrahepatic† 22

Treatment for recurrence (n)Reoperation†† 5Local ablation††† 50Sorafenib 5Conservative treatment 2

⁄WR, wedge resection.�Including 12 patients with both intrahepatic and extrahepatic recurrences.��Including 4 patients with re-resection in liver and 1 patient in right lung.���Including percutaneous radiofrequency ablation (PRFA) and percutaneousethanol injection (PEI).

Table 3. Outcomes of the TACE group.

TACE (n = 85)Adverse events* (n)

Nausea/vomiting 83Pain 48Leukopenia 33Increase in ALT/AST 56Increase in GGT 34Decrease in albumin 25Increase in bilirubin 45

⁄Adverse events were defined as adverse effects related to TACE of grade 3 orbelow.No grade 4 or 5 adverse effects related to TACE were observed.ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, c glutamyltranspeptidase.

1.0

0.8

0.6

0.4

0.2

0.0

0 12 24 36 48

Ove

rall

surv

ival

rate

Log-rank, χ2 = 24.246, p <0.001

PHPH-censorTACETACE-censor

Number at riskPH

TACE8885

6743

5527

3911

40

Time (mo)

Fig. 2. Overall survival curves for PH and TACE.

1.0

0.8

0.6

0.4

0.2

0.0

0 12 24 36 48

Ove

rall

surv

ival

rate

Log-rank, χ2 = 20.866, p <0.001

PHPH-censorTACETACE-censor

Number at riskPH

TACE5548

4426

4018

329

30

Time (mo)

Fig. 3. Overall survival curves for patients with 2 tumors after PH or TACE.

Research Article

Whether TACE is the treatment of choice for resectable multi-ple HCC outside of Milan Criteria is still uncertain.

In carefully selected patients with low surgical risks, PH hasbeen shown to prolong survival and provide good symptomaticrelief for patients with resectable multiple HCC [20] in non-ran-domized studies [9,12]. The surgery aims to remove macroscopictumors, possibly allowing microscopic foci to persist in the liverremnant, which can then be treated with regional or systemictherapy. Theoretically, these perioperative therapies may helpto eradicate any residual tumors, if present. Retrospective studieshave also indicated that PH provided better survival for patientswith resectable multiple HCC than TACE [9,10].

This randomized study showed PH to be better than TACE as afirst line treatment for resectable multiple HCC outside of MilanCriteria. TACE can down-stage tumors but this has to be followed

86 Journal of Hepatology 20

by salvage surgery as viable tumor cells are likely to be present inthe necrotic areas even after CR to TACE. Unfortunately, salvagesurgery cannot be carried out in a significant proportion of

14 vol. 61 j 82–88

1.0

0.8

0.6

0.4

0.2

0.0

0 12 24 36 48

Ove

rall

surv

ival

rate

Log-rank, χ2 = 4.591, p = 0.032

PHPH-censorTACETACE-censor

Number at riskPH

TACE3337

2317

159

72

10

Time (mo)

Fig. 4. Overall survival curves for patients with >2 tumors after PH or TACE.

JOURNAL OF HEPATOLOGY

patients after tumor downstaging by TACE because TACE can alsodamage the non-tumorous part of the liver with resultant deteri-oration in liver function. Also, HCC commonly progresses onTACE, making initially resectable HCC not resectable.

One major criticism of the present study is the poor overallsurvival in the TACE group when compared with the results pub-lished in recent medical literature [21–24]. Drug-eluting beadshave been shown to give better survival than conventional TACE[25]. Unfortunately, this treatment is not available to us becausethe State Food and Drug Administration has not approved of itsuse in China.

A RCT from Japan showed a median OS of 21 months in theTACE group [26]. The median maximum tumor diameter of thisJapanese study was 35 mm as compared with 73 mm in ourstudy. The proportion of HBsAg (+) in the HCC patients was13.9% in the Japanese study while it was 90.6% in our study. Thusthe etiological factors of HCC in the two studies were significantly

Table 4. Univariate and multivariate analysis of prognostic factors of OS.

Variableχ2 value (L

Treatment (PH vs. TACE) 24.246Number of tumor (>2 vs. 2) 12.800Sex (female vs. male) 11.282Serum AFP* (ng/ml) (>400 vs. ≤400) 6.204Total tumor size (cm) (>10 Child-Pugh classification (B vs. A)

vs. ≤10) 5.2350.001

Age (yr) (>50 vs. ≤50) 1.478HBsAg (positive vs. negative) 0.273HBeAg (positive vs. negative) 0.981HBV-DNA (IU/ml) (>1000 vs. ≤1000) 0.233HCV (positive vs. negative) 0.038Cirrhosis (normal vs. cirrhosis) 0.010Tumor location (same hepatic segment vs. different hepatic segment)

0.001

⁄AFP, alpha-fetoprotein.n.a., not applicable.

Journal of Hepatology 20

different. Another study from Taiwan showed the median OS ofpatients with HCC of up to 5 cm in diameter and with Child-PughA liver function was 25 to 35 months after TACE [27]. Taking intoconsideration the early stage in the Taiwanese study and theintermediate stage of HCC of the patients in our study, thereseems to be little differences in the results in treatment by TACEfrom these studies.

Although the recurrence rate of 68.9% in the PH group wasvery high, exciting OS results were obtained. This finding is rea-sonable. PH removes almost all macroscopic tumors to createthe condition for subsequent comprehensive treatment. Obvi-ously in the current study, a comprehensive strategy includingthe use of PH and the subsequent multidisciplinary treatmentplayed a significant role in obtaining good OS. A preoperativelow AFP level is also known to be a factor which is associatedwith good OS. This study showed PH to be a better first line treat-ment for patients with resectable multiple HCC than TACE.

The number of tumor predicted the degree of intrahepaticspread and it has been shown by studies to be an independentrisk factor of HCC [28,29]. Although more than half of our patients(103/173) had 2 tumors, the OS rates were significantly differentbetween the PH group and the TACE group irrespective ofwhether the number of tumor was 2 or more. Epidemiologicalreports showed the incidence of HCC was higher in males thanin females [30]. Although the reasons for the difference in HCCrisk between men and women are not clear, studies have demon-strated that the gender disparity of HCC can be attributed to theandrogen and estrogen sex hormonal pathways [31].

Our study showed Child-Pugh liver functional status not to bea significant prognostic factor of overall survival on univariateand multivariate Cox proportional hazards regression analyses.The small sample size of this study may well be the reason forthis result. Only 7(4%) patients had Child-Pugh B. This is a short-coming and a defect of this study.

Included into this study are patients who were outside ofMilan Criteria, and they are not candidates for liver transplanta-tion in most transplant centers [32]. The major disadvantage ofPH to treat these patients is the inherent morbidity and mortality

Univariate Multivariateog-rank) p value HR (95% CI) p value

<0.001 0.434 (0.293, 0.644) <0.001<0.001 1.758 (1.213, 2.548) 0.0030.001 0.451 (0.236, 0.862) 0.0160.013 1.362 (0.920, 2.016) 0.1220.022 1.142 (0.776, 1.681) 0.5010.973 1.204 (0.434, 3.344) 0.7220.224 - n.a.0.601 - n.a.0.322 - n.a.0.630 - n.a.0.845 - n.a.0.919 - n.a.0.976 - n.a.

14 vol. 61 j 82–88 87

Research Article

which are associated with PH. Thus, patients should be selectedvery carefully.

In conclusion, the outcome of PH as an initial treatment forpatients with tumor outside of Milan Criteria was superior toconventional TACE. The number of tumor and gender were alsofound to be independent risk factors associated with OS for thesepatients. To achieve good results of PH, patients should be care-fully selected to minimize postoperative mortality and majormorbidity.

Financial support

This study was supported by State Key Infection Disease Projectof China (Project Number: 2012ZX10002010, 2012ZX10002016),Science Fund for Creative Research Groups of China (ProjectNumber 81201940), Youth Fund of Health Bureau of Shanghai(Project Number 201144172) and Natural Science Fund of Shanghai(Project Number 13ZR1450800).

Conflict of interest

The authors who have taken part in this study declared that theydo not have anything to disclose regarding funding or conflict ofinterest with respect to this manuscript.

Acknowledgments

The authors thank Professor Meng Hong (Department of HealthStatistics, Second Military Medical University, Shanghai, China)for her critical revision of the univariate and multivariate analysissections.

References

[1] Bosch FX, Ribes J, Borras J. Epidemiology of primary liver cancer. Semin LiverDis 1999;19:271–285.

[2] Lai EC, Fan ST, Lo CM, Chu KM, Liu CL, Wong J. Hepatic resection forhepatocellular carcinoma. An audit of 343 patients. Ann Surg 1995;221:291–298.

[3] Bruix J, Llovet JM. Prognostic prediction and treatment strategy in hepato-cellular carcinoma. Hepatology 2002;35:519–524.

[4] Chen MF, Tsai HP, Jeng LB, Lee WC, Yeh CN, Yu MC, et al. Prognostic factorsafter resection for hepatocellular carcinoma in noncirrhotic livers: univariateand multivariate analysis. World J Surg 2003;27:443–447.

[5] Lau WY, Lai EC. Salvage surgery following downstaging of unresectablehepatocellular carcinoma – a strategy to increase resectability. Ann SurgOncol 2007;14:3301–3309.

[6] Forner A, Llovet JM, Bruix J. Hepatocellular carcinoma. Lancet 2012;379:1245–1255.

[7] Oliveri RS, Wetterslev J, Gluud C. Transarterial (chemo)embolisation forunresectable hepatocellular carcinoma. Cochrane Database Syst Rev 2011;16:CD004787.

[8] Lai EC, Tang CN, Ha JP, Tsui DK, Li MK. Cytoreductive surgery in multidis-ciplinary treatment of advanced hepatocellular carcinoma. ANZ J Surg2008;78:504–507.

[9] Nagashima J, Okuda K, Tanaka M, Sata M, Aoyagi S. Prognostic benefit incytoreductive surgery for curatively unresectable hepatocellular carcinoma– comparison to transcatheter arterial chemoembolization. Int J Oncol1999;15:1117–1123.

88 Journal of Hepatology 20

[10] Inoue K, Nakamura T, Kinoshita T, Konishi M, Nakagohri T, Oda T, et al.Volume reduction surgery for advanced hepatocellular carcinoma. J CancerRes Clin Oncol 2004;130:362–366.

[11] Bruix J, Sherman M. Management of hepatocellular carcinoma: an update.Hepatology 2011;53:1020–1022.

[12] Wakabayashi H, Ushiyama T, Ishimura K, Izuishi K, Karasawa Y, Masaki T,et al. Significance of reduction surgery in multidisciplinary treatment ofadvanced hepatocellular carcinoma with multiple intrahepatic lesions. JSurg Oncol 2003;82:98–103.

[13] Zhou WP, Lai EC, Li AJ, Fu SY, Zhou JP, Pan ZY, et al. A prospective,randomized, controlled trial of preoperative transarterial chemoemboliza-tion for resectable large hepatocellular carcinoma. Ann Surg 2009;249:195–202.

[14] Ishizawa T, Hasegawa K, Aoki T, Takahashi M, Inoue Y, Sano K, et al. Neithermultiple tumors nor portal hypertension are surgical contraindications forhepatocellular carcinoma. Gastroenterology 2008;134:1908–1916.

[15] Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepato-cellular carcinoma. Semin Liver Dis 2010;30:52–60.

[16] Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF, et al. Sorafenib inadvanced hepatocellular carcinoma. N Engl J Med 2008;359:378–390.

[17] Cheng AL, Kang YK, Chen Z, Tsao CJ, Qin S, Kim JS, et al. Efficacy and safety ofsorafenib in patients in the Asia-Pacific region with advanced hepatocellularcarcinoma: a phase III randomised, double-blind, placebo-controlled trial.Lancet Oncol 2009;10:25–34.

[18] Chen MF, Hwang TL, Jeng LB, Jan YY, Wang CS, Chou FF. Hepatic resection in120 patients with hepatocellular carcinoma. Arch Surg 1989;124:1025–1028.

[19] Lau H, Fan ST, Ng IO, Wong J. Long term prognosis after hepatectomy forhepatocellular carcinoma: a survival analysis of 204 consecutive patients.Cancer 1998;83:2302–2311.

[20] Lau WY, Leung TW, Leung KL, Ho S, Leung N, Chan M, et al. Cytoreductivesurgery for hepatocellular carcinoma. Surg Oncol 1994;3:161–166.

[21] Llovet JM, Real MI, Montana X, Planas R, Coll S, Aponte J, et al. Arterialembolisation or chemoembolisation vs. symptomatic treatment in patientswith unresectable hepatocellular carcinoma: a randomised controlled trial.Lancet 2002;359:1734–1739.

[22] Burrel M, Reig M, Forner A, Barrufet M, de Lope CR, Tremosini S, et al.Survival of patients with hepatocellular carcinoma treated by transarterialchemoembolisation (TACE) using Drug Eluting Beads. Implications forclinical practice and trial design. J Hepatol 2012;56:1330–1335.

[23] Dhanasekaran R, Kooby DA, Staley CA, Kauh JS, Khanna V, Kim HS. Prognosticfactors for survival in patients with unresectable hepatocellular carcinomaundergoing chemoembolization with doxorubicin drug-eluting beads: apreliminary study. HPB (Oxford) 2010;12:174–180.

[24] Takayasu K, Arii S, Ikai I, Omata M, Okita K, Ichida T, et al. Prospective cohortstudy of transarterial chemoembolization for unresectable hepatocellularcarcinoma in 8510 patients. Gastroenterology 2006;131:461–469.

[25] Song MJ, Chun HJ, Song do S, Kim HY, Yoo SH, Park CH, et al. Comparativestudy between doxorubicin-eluting beads and conventional transarterialchemoembolization for treatment of hepatocellular carcinoma. J Hepatol2012;57:1244–1250.

[26] Okusaka T, Kasugai H, Shioyama Y, Tanaka K, Kudo M, Saisho H, et al.Transarterial chemotherapy alone vs. transarterial chemoembolization forhepatocellular carcinoma: a randomized phase III trial. J Hepatol 2009;51:1030–1036.

[27] Hsieh MY, Chen SC, Lu SN, Wang LY, Tsai JF, Chuang WL, et al. Treatment ofhepatocellular carcinoma smaller than 5 cm by transcatheter arterialchemoembolization. Kaohsiung J Med Sci 1996;12:274–278.

[28] Zhao WC, Fan LF, Yang N, Zhang HB, Chen BD, Yang GS. Preoperativepredictors of microvascular invasion in multinodular hepatocellular carci-noma. Eur J Surg Oncol 2013;39:858–864.

[29] Kim BK, Han KH, Park YN, Park MS, Kim KS, Choi JS, et al. Prediction ofmicrovascular invasion before curative resection of hepatocellular carci-noma. J Surg Oncol 2008;97:246–252.

[30] El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemiology andmolecular carcinogenesis. Gastroenterology 2007;132:2557–2576.

[31] Yeh SH, Chen PJ. Gender disparity of hepatocellular carcinoma: the roles ofsex hormones. Oncology 2010;78:172–179.

[32] Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Bozzetti F, et al.Liver transplantation for the treatment of small hepatocellular carcinomas inpatients with cirrhosis. N Engl J Med 1996;334:693–699.

14 vol. 61 j 82–88