J Hepatobiliary Pancreat Surg (2001) 8:81–86

Original article

Prognostic factors after recurrence of resected hepatocellularcarcinoma associated with hepatitis C virus

Kazuhiro Hirohashi, Taichi Shuto, Shoji Kubo, Hiromu Tanaka, Takatsugu Yamamoto, Takashi Ikebe,Junya Murase, and Hiroaki Kinoshita

Second Department of Surgery, Osaka City University Medical School, 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan

The clinicopathological features of HCC differ accord-ing to the associated hepatitis virus; hepatitis C virus(HCV)-associated HCC is predominant in Japan.6–8 Inthis retrospective study, we attempted to clarify theprognostic factors for survival of HCC patients afterrecurrence associated with HCV infection.

Methods

Two hundred and seventy patients underwent hepaticresection from April 1990 to March 1997. Thirteen pa-tients died in hospital prior to discharge. The remaining257 patients were followed closely in our outpatientclinic or at affiliated institutions, with liver functiontests, including alpha-fetoprotein (AFP) and/or proteininduced by vitamin K absence-II (PIVKA-II), ultra-sonography (US), and computed tomography (CT),carried out at least every 3 months. When tumor recur-rence was suspected, based on increased AFP and/orPIVKA-II levels, and US and CT findings, angiographyor fine-needle biopsy under US guidance were per-formed to obtain a diagnosis of recurrence.

As of March 1997, 130 (51%) of the 257 patients hadclinically confirmed intrahepatic recurrences. Ninety-nine of these patients, with detectable HCV antibodiesand absence of hepatitis B surface antigen, were in-cluded in this study. Their survival rate after recurrencewas examined in relation to various clinicopathologicalvariables. The following variables were considered atfirst resection: maximum tumor diameter, pathologictumor differentiation, cirrhosis, and active hepatitis.The following variables were considered at recurrence:sex, age, interval to recurrence, platelet count, albumin,bilirubin, alanine aminotransferase (ALT), AFP, andPIVKA-II levels, number of recurrent hepatic lesions,and distant metastasis. Treatment variables were alsoconsidered: type of operation at first resection, ability toundergo a second hepatic resection (SHR), and treat-

Abstract To clarify the variables related to survival afterrecurrence of resected hepatocellular carcinoma (HCC) asso-ciated with hepatitis C virus (HCV), we studied 17 clinico-pathological factors in 99 patients with recurrence of HCCassociated with HCV infection after hepatic resection. The1-, 3-, and 5-year survival rates after first resection in thesepatients were 91%, 81%, and 49%, while after recurrence theywere 81%, 51%, and 29%, respectively. Multivariate analysisshowed that the following six variables were independentprognostic factors after recurrence: platelet count, albuminlevel, bilirubin level, number of hepatic lesions, distantmetastasis, and any treatment at recurrence. A correlationbetween second hepatic resection (SHR) and liver functiontests was seen in regard to albumin and total bilirubin valuesat recurrence. Indeed, hepatic function and progression ofintrahepatic tumors at recurrence were significant prognosticfactors after recurrence of HCC associated with HCV infec-tion, while any treatment at recurrence was also a significantprognostic factor. Therefore, in order to improve prognosisafter recurrence, we should actively treat the recurrent he-patic lesions whenever possible.

Key words Hepatocellular carcinoma · Hepatic resection ·Recurrence · Prognosis · Hepatitis C virus

Introduction

The long-term results of hepatic resection for hepato-cellular carcinoma (HCC) in cirrhotic patients havebeen disappointing, primarily because of the high rateof recurrence after surgery.1–5 In order to achieve a bet-ter overall prognosis, it is important not only to improvethe tumor-free survival rate after the initial resectionbut also to improve the survival rate after recurrence.1

Offprint requests to: K. HirohashiReceived: August 22, 2000 / Accepted: November 20,2000

82 K. Hirohashi et al.: Prognosis after recurrence of HCC

ment at recurrence. All the significant variables forunivariate analysis were then examined in a Cox’s pro-portional hazards model to identify any independentvariables correlated with survival after recurrence.

Virus markers

Serum samples obtained before surgery from all pa-tients were assayed for hepatitis B and C viruses. Anenzyme immunoassay (International Reagents, Kobe,Japan) was used to examine serum for hepatitis B sur-face antigen. Samples were also examined for HCVantibodies, by second- or third-generation enzyme-linked immunosorbent assay (ELISA) (Ortho Diagnos-tic Systems, Tokyo, Japan).

Treatment strategy for recurrent HCC

For intrahepatic recurrence, we selected one or moreof the following treatments according to liver functiontest findings and tumor condition: SHR, transcatheterhepatic artery infusion of anticancer agents (lipiodoli-zation), transcatheter arterial embolization (TAE), per-cutaneous ethanol injection therapy (PEIT), andmicrowave coagulation therapy (PMCT) under USguidance.

As our primary concept, because SHR was consid-ered safe and effective, and offered an improved prog-nosis for recurrent HCC, we always selected SHRwhenever possible. Although the indications for SHRwere almost the same as those for the first resection, weperformed SHR only when the following conditionswere met: (1) the patient must have good liver function,similar to findings before the first resection. (2) Therecurrent tumor must be a localized lesion, so that itcan be completely extirpated. For exceptional patients

with multiple recurrent tumors in the remnant liverwho had not responded to other suitable treatments butwho had adequate liver function for undergoing SHR,we performed SHR as part of a multimodal treatmentplan.

Lipiodolization was used primarily in patients forwhom SHR was not an option, either because of theirrefusal, or because of an excessive number of hepatictumors or poor liver function. At the time thatlipiodolization was planned, if staining of the tumor onangiography was obvious, we usually performed TAErather than lipiodolization. However, in some patientswith fewer than three recurrent lesions that were 3cmor less in diameter, PEIT or PMCT, was performed,while PEIT or PMCT was also performed if lipiodo-lization or TAE was not assessed as effective.

Statistics

Survival rates, calculated using the Kaplan-Meiermethod, were compared with the log-rank test. The Coxregression test was used for multivariate analysis. TheMann-Whitney U-test was used to evaluate the signifi-cance of laboratory test values. A P value of less than0.05 was considered statistically significant.

Results

The 1-, 3-, and 5-year survival rates after first resectionin the 99 patients were 91%, 81%, and 49%, while afterrecurrence these rates were 81%, 51%, and 29%, re-spectively (Fig. 1). Univariate analysis was used toidentify significant factors related to survival afterrecurrence of resected HCC associated with HCV infec-tion (Tables 1, 2). There was one significant pathologi-

Fig. 1. Survival rates after first resectionand recurrence of resected hepatocellularcarcinoma associated with hepatitis C viralinfection. Open circles show survival ratesafter first resection, and closed circles showrates after recurrence

K. Hirohashi et al.: Prognosis after recurrence of HCC 83

Table 1. Pathological variables in patients with hepatitis C virus-associatedhepatocellular carcinoma at first resection

3-Year survival rateVariable after recurrence (%) P value*

Tumor size (cm)#5.0 (n 5 84) 50.0 0.0700.5.0 (n 5 15) 38.3

Tumor differentiationWell or moderately (n 5 66) 69.2 0.0002Poorly (n 5 18) 16.2

CirrhosisWithout (n 5 24) 51.5 0.9308With (n 5 74) 41.0

Active hepatitisWithout (n 5 43) 58.7 0.1994With (n 5 52) 42.7

*These results were compared by the log-rank test

Table 2. Clinical variables in patients with hepatitis C virus-associated hepatocellularcarcinoma at recurrence

3-Year survival rateVariable after recurrence (%) P value*

SexMale (n 5 82) 40.2 0.6800Female (n 5 17) 50.2

Age (years)#60 (n 5 28) 75.8 0.0107.60 (n 5 71) 36.4

Interval to recurrence.1 Year (n 5 60) 48.0 0.9184#1 Year (n 5 39) 48.9

Platelets (/mm3).50000 (n 5 85) 50.6 0.0144#50000 (n 5 3) 0

Albumin (g/dl).3.0 (n 5 82) 50.6 0.0010#3.0 (n 5 8) 0

Bilirubin (mg/dl)#1.0 (n 5 56) 56.8 0.0049.1.0 (n 5 34) 32.3

ALT (IU/dl)#40 (n 5 10) 46.7 0.3550.40 (n 5 80) 47.3

AFP (ng/ml)#20 (n 5 48) 51.3 0.0338.20 (n 5 41) 43.7

PIVKA-II (AU/ml),0.06 (n 5 45) 46.3 0.1740$0.06 (n 5 31) 38.4

Hepatic lesionsCountable (n 5 65) 64.7 ,0.0001Too numerous to count (n 5 34) 19.5

Distant metastasisWithout (n 5 84) 52.0 0.0371With (n 5 15) 31.6

*These results were compared by the long-rank testALT, Alanine aminotransferase; AFP, alpha-fetoprotein; PIVKA-II, protein induced by vitaminK absence-II

84 K. Hirohashi et al.: Prognosis after recurrence of HCC

cal factor indicating poor prognosis at initial resection;poorly differentiated resected tumors. Significantfactors indicating poor prognosis at recurrence wereage more than 60 years, platelet count, 50000/mm3 orless, albumin, 3.0 g/dl or less, total bilirubin more than1.0mg/dl, AFP more than 20 ng/ml, recurrent hepaticlesions too numerous to count, and the presence ofdistant metastasis.

At first resection, major hepatic resection was per-formed in 45 patients, and minor hepatic resection in54 patients (see Table 3 for definitions of major andminor). At recurrent, 83 patients received treatment(of any type), while 16 received no treatment. Of these83 patients, a SHR was done in 17. For the other 66patients, lipiodolization, TAE, PEIT, PMCT, and some

Table 3. Treatment of patients with hepatitis C virus-associated hepatocellularcarcinoma

3-Year survival rateVariable after recurrence (%) P valuea

Hepatic resection at first resectionb

Major (n 5 45) 53.9 0.1395Minor (n 5 54) 44.7

Treatment at recurrencec

Second hepatic resection (n 5 17) 85.1 ,0.0001Other treatment (n 5 66) 59.5No treatment (n 5 16) 0

a These results were compared by the log-rank testb Major resection was defined as a bisegmentectomy; a segmentectomy (resection of more thanone segment but less than a hemihepatic lobe); or a subsegmentectomy (anatomic resection ofhepatic portions smaller than a segment). All other nonanatomic resections were defined as minorresectionsc For intrahepatic recurrence, according to our criteria, one or more of the following treatmentswere performed: lipiodolization, transcatheter arterial embolization, percutaneous ethanolinjection therapy or microwave coagulation therapy under ultranonographic (US) guidance, andsecond hepatic resection

combination of these, was performed. SHR and anytreatment for hepatic recurrence were good prognosticfactors (Table 3). In particular, the prognosis of thepatients who had SHR was excellent (Fig. 2). Multivari-ate analysis using Cox’s proportional hazard model wasalso done (Table 4). Any treatment for hepatic recur-rence, platelet count, bilirubin and albumin levels atrecurrence, number of recurrent hepatic lesions, anddistant metastasis at recurrence were independent prog-nostic factors. No significant correlation between theperformance of treatment and liver function test resultswas seen at recurrence (Table 5). On the other hand, asignificant correlation between SHR and liver functiontest results was seen in terms of values for albumin andtotal bilirubin at recurrence (Table 6).

Fig. 2. Survival rates after recurrenceof resected hepatocellular carcinomaassociated with hepatitis C viral infection.Closed circles show survival rates inpatients with second hepatic resection(n 5 17), open triangles show rates inpatients with other treatment (n 5 66),and open squares show rates in thosewithout treatment (n 5 16). There weresignificant differences between the threegroups (P , 0.0001)

K. Hirohashi et al.: Prognosis after recurrence of HCC 85

Table 4. Multivariate analysis to assess survival rates after recurrence

Variable Coefficient SE Relative risk P valuea

No treatment 4.800 0.912 121.482 ,0.0001Platelets .50000/mm3 24.139 1.008 0.016 ,0.0001Albumin #3.0 g/dl 2.481 0.706 11.950 0.0004Uncountable hepatic lesions 1.916 0.564 6.791 0.0007Bilirubin #1.0mg/dl 21.744 0.595 0.175 0.0034No distant metastasis 21.441 0.686 0.237 0.0356a These results were calculated by Cox’s proportional hazards modelAll variables were determed at recurrence

Table 5. Correlation between any treatment and liver function at recurrence

Any treatmenta

Variable With (n 5 83) Without (n 5 16) P valueb

Platelets (3104/mm3) 12 (4.4; 4.3–26) 12 (4.6; 6.0–24) 0.5699Total bilirubin (mg/dl) 1.0 (0.46; 0.2–2.7) 1.0 (0.41; 0.4–1.8) 0.9608Albumin (g/dl) 3.7 (0.42; 2.8–4.6) 3.6 (0.36; 2.7–4.1) 0.2970ALT (IU/l) 100 (41; 14–714) 70 (33; 19–142) 0.1560

All values are expressed as means, with SDs and ranges given in parenthesesALT, Alanine aminotransferasea One or more of the following treatments were used: transcatheter hepatic arterial infusion ofanticancer agents, transcatheter arterial embolization, percutaneous ethanol injection therapy, ormicrowave coagulation therapy under US guidance, and second hepatic resectionb These values were compared with the Mann-Whitney U-test

Discussion

To achieve better overall survival after hepatic re-section for HCC,1 both prevention of recurrenceand improved survival after recurrence are necessary.The time until recurrence after surgery,1,4,9 tumor size,5

and poor tumor differentiation,10 have been shown to beindependent prognostic factors relevant to both recur-rence and survival. Although these factors were be-lieved to be strongly related to metastatic recurrence,rather than to recurrence of multicentric origin,11

the time until recurrence after surgery and tumor sizewere not a significant prognostic factor in this study.This may be because our study population was limitedto those with HCV-associated HCC.2

It is well known that the clinicopathological featuresof HCC associated with HBV and HCV are different.7,8

High rates of histological cirrhosis and active hepatitiswith HCV may not be the result, but rather, the cause ofHCC in patients with HCV.7 Because cirrhosis itself hasbeen considered a premalignant condition,12 several in-vestigators have speculated that recurrences, especiallywith HCV-associated HCC, result from the multicentricorigin of the tumor.3,12 Thus, when we consider onlypatients with HCV-associated HCC, many prognosticfactors at first resection are strongly related to meta-static recurrence, but may not be significant after multi-centric recurrence.

Conditions at recurrence, such as the number ofrecurrent hepatic lesions,3 distant metastasis, and liver

Table 6. Correlation between second hepatic resection and liver function at recurrence

Second hepatic resection

Variable With (n 5 17) Without (n 5 82) P valuea

Platelets (3104/mm3) 13 (4.6; 6.8–22) 11 (4.3; 4.3–26) 0.1112Total bilirubin (mg/dl) 0.79 (0.28; 0.4–1.5) 1.1 (0.46; 0.2–2.7) 0.0043Albumin (g/dl) 3.9 (0.36; 3.3–4.6) 3.6 (0.40; 2.7–4.5) 0.0086ALT (IU/l) 87 (41; 29–174) 98 (89; 14–714) 0.9519

All values are expressed as means, with SDs and ranges given in parenthesesALT, Alanine aminotransferasea These values were compared with the Mann-Whitney U-test

86 K. Hirohashi et al.: Prognosis after recurrence of HCC

function, were significant for prognosis after recurrenceof HCC associated with HCV. In this study, the type offirst resection was not a significant factors in the progno-sis after recurrence. Thus, the aim of the first resectionfor HCC associated with HCV should be the improve-ment of tumor-free survival rates.

As for treatment at recurrence, the most importantconcept is that of continuing to try to treat recurrentlesions, because the performance of any treatment wasa significant prognostic factor in this study. Indeed,SHR is now widely accepted as an effective treatmentfor recurrent HCC,1,2,9 but because of altered anatomyand the presence of perihepatic adhesions,1,11 it can beperformed only in a few selected patients. In particular,because we found a significant correlation betweenSHR and liver function test results at recurrence, thetreatment decision to resect was dependent on the se-verity of the liver disease. When SHR is not indicated,combined treatment with PEIT and lipiodolization isoften recommended.1 Although the effectiveness ofPEIT for recurrent HCC has been shown,13 PMCT hasalso been reported recently to be safe and effective.14

There is also some evidence that surgical resection of anisolated extrahepatic recurrence can be beneficial.1,2

In conclusion, hepatic function and progression ofintrahepatic tumors at recurrence were significant prog-nostic factors after recurrence of HCC associated withHCV infection, while the performance of any treatmentat recurrence was also a significant prognostic factor.Therefore, in order to improve prognosis after recur-rence, we should actively treat the hepatic lesions when-ever possible.

References

1. Shimada M, Takenaka K, Gion T, Fujiwara Y, Kajiyama K,Maeda T, Shirabe K, Nishizaki T, Yanaga K, Sugimachi K(1996) Prognosis of recurrent hepatocellular carcinoma: a 10-

year surgical experience in Japan. Gastroenterology 111:720–726

2. Shimada M, Takenaka K, Gion T, Fujiwara Y, Kajiyama K,Maeda T, Shirabe K, Yanaga K, Sugimachi K (1998) Prognosticfactors after repeat hepatectomy for recurrent hepatocellular car-cinoma. Ann Surg 227:80–85

3. Belghiti J, Panis Y, Farges O, Benhamou JP, Fekete F (1991)Intrahepatic recurrence after resection of hepatocellular carci-noma complicating cirrhosis. Ann Surg 214:114–117

4. Matsumata T, Kanematsu T, Takenaka K, Yoshida Y, NishizakiT, Sugimachi K (1989) Patterns of intrahepatic recurrence aftercurative resection of hepatocellular carcinoma. Hepatology9:457–460

5. Adachi E, Maeda T, Matsumata T, Shirabe K, Kinukawa N,Sugimachi K, Tsuneyoshi M (1995) Risk factors for intrahepaticrecurrence in human small hepatocellular carcinoma. Gastroen-terology 108:768–775

6. Trevisani F, D’Intino PE, Caraceni P, Pizzo M, Stefanini GF,Mazziotti A, Grazi GL, Gozzetti G, Gasbarrini G, Bernardi M(1995) Etiologic factors and clinical presentation of hepato-cellular carcinoma. Cancer 75:2220–2232

7. Shiratori Y, Shiina S, Imamura M, Kato N, Kanai F, Okudaira T,Teratani T, Tohgo G, Toda N, Ohashi M, Ogura K, Niwa Y,Kawabe T, Omata M (1995) Characteristic difference of hepato-cellular carcinoma between hepatitis B- and C-viral infection inJapan. Hepatology 22:1027–1033

8. Shuto T, Hirohashi K, Kubo S, Tsukamoto T, Yamamoto T,Wakasa K, Kinoshita H (1998) Differences in clinicopathologicalcharacteristics of resected hepatocellular carcinoma associatedwith hepatitis B or C virus. Hepatogastroenterology 45:1722–1725

9. Shuto T, Kinoshita H, Hirohashi K, Kubo S, Tanaka H,Tsukamoto T, Okuda T (1996) Indications for, and effectivenessof, a second hepatic resection for recurrent hepatocellular carci-noma. Hepatogastroenterology 43:932–937

10. Jwo SC, Chin JH, Chau GY, Loong CC, Lui WY (1992) Riskfactors linked to tumor recurrence of human hepatocellular carci-noma after hepatic resection. Hepatology 16:1367–1371

11. Lee PH, Lin WJ, Tsang YM, Hu RH, Sheu JC, Lai MY, Hsu HC,May W, Lee CS (1995) Clinical management of recurrent hepato-cellular carcinoma. Ann Surg 222:670–676

12. Sakamoto M, Hirohashi S, Shimosato Y (1991) Early stages ofmultistep hepatocarcinogenesis. Hum Pathol 22:172–178

13. Tanikawa K, Majima Y (1993) Percutaneous ethanol injectiontherapy for recurrent hepatocellular carcinoma. Hepatogastroen-terology 40:324–327

14. Seki T, Wakabayashi M, Nakagawa T, Itho T, Shiro T, KuniedaK, Sato M, Uchiyama S, Inoue K (1994) Ultrasonically guidedpercutaneous microwave coagulation therapy for small hepato-cellular carcinoma. Cancer 74:817–825