Low Hepatitis C Viral Load Predicts Better Long-Term Outcomes in Patients Undergoing Resection of Hepatocellular Carcinoma Irrespective of Serologic Eradication of Hepatitis C Virus

Purpose Hepatitis C virus (HCV) infection has been recognized as a potent risk factor for the postoperative recurrence of hepatocellular carcinoma (HCC). However, little is known about the impact of HCV viral load on surgical outcomes. The study objective was to investigate clinical significance of HCV viral load on long-term outcomes of HCC. Patients and Methods Three hundred seventy patients who were classified as Child-Pugh class A and underwent curative liver resections for HCV-related HCC were divided into low and high viral load groups ( or 5.3 log10IU/mL) based on the results of a minimum P value approach to predict moderate to severe activity of hepatitis; the clinical outcomes were then compared. Results The 5-year recurrence-free survival rate was 36.1% in the low viral load group and 12.4% in the high viral load group (P .001). The 5-year overall survival rate was 76.6% in the low viral load group and 57.7% in the high viral load group (P .001). Multivariate analysis confirmed significant correlation between high viral load and tumor recurrence with a hazard ratio of 1.87 (95% CI, 1.41 to 2.48; P .001). Subanalysis revealed that the favorable results in the low viral load group were not attributed to whether or not serologic eradication of HCV was obtained both in primary and recurrent lesions. Conclusion Low HCV viral load predicts better long-term surgical outcomes in patients with HCC regardless of the serologic eradication of HCV. J Clin Oncol 30. © 2012 by American Society of Clinical Oncology


INTRODUCTION
][4][5][6][7] Hepatitis C virus (HCV), a major cause of chronic hepatitis and liver cirrhosis, has been recognized as a potent risk factor of carcinogenesis 8 and/or the recurrence of HCC. 9,10][13] Conventionally, the eradication of HCV and a sustained status of undetectable HCV-RNA have been regarded as the most important factors for obtaining better clinical results after IFN therapy.In recent studies, however, possibly favorable effects of a reduced viral load on longterm outcomes have been suggested in patients with chronic hepatitis. 14,15ur hypothesis was that a correlation existed between the HCV viral load and long-term surgical outcomes.Simply labeling patient as having viremia did not sufficiently stratify those who had low viral load versus those who had high viral load.In this study, we tested this hypothesis by examining patients who had undergone curative liver resection for HCV-related HCC and analyzed the impact of the HCV viral load on postoperative outcomes.

Study Population
This study was performed in accordance with the ethical guidelines for clinical studies at the University of Tokyo Hospital (Tokyo, Japan).The subject pool consisted of 508 consecutive patients who underwent curative liver resection for HCV-related HCC between January 2002 and December 2011.Patients classified as Child-Pugh class B (n ϭ 49) or patients missing preoperative viral load data (n ϭ 89) were excluded because the goal of this study was to reveal the prognostic impact of the HCV viral load in patients who were considered to be capable of tolerating antiviral therapies.The remaining 370 patients were included in the analysis.

Serum HCV-RNA Quantification
Serum HCV-RNA was quantified within 4 weeks before surgery using a conventional reverse transcriptase polymerase chain reaction (PCR) assay before 2007 and a new commercially available real-time PCR assay (TaqMan PCR; Roche Molecular Systems, Pleasanton, CA) in 2007 and thereafter.In this study, the viral load unit was standardized to a logarithm style (log 10 IU/ mL) for the statistical analysis according to the following equation: Y (log 10 IU/ mL) ϭ log 10 [X (kIU/mL) ϫ 10 3 ].

Surgical Treatment and Histopathologic Assessments
The indications for hepatic resection and the types of operative procedures were determined as previously described. 16Briefly, operative decisions were based on an algorithm consisting of the presence of ascites, the serum total bilirubin level, and the results of an indocyanine green tolerance test. 17ecause HCC has a high propensity to invade the portal veins and because intrahepatic metastasis via vascular invasion is one of the major forms of recurrence, tumor-bearing portal regions (ie, the segment or subsegment of the liver) were systematically removed (ie, an anatomic resection) to reduce the risk of local recurrence as long as such resections were feasible given the functional reserve of the liver. 18he histologic classifications of the tumor and background liver were described based on the system of the Liver Cancer Study Group of Japan. 19The histologic differentiation of HCC (well, moderate, or poor) was determined according to the Edmondson grade. 19,20Both the fibrotic stage and the activity of the hepatitis in the background liver were also recorded according the classification proposed by Desmet et al. 21

Postoperative Antiviral Therapy
Postoperative adjuvant IFN therapy was performed only in patients who had a good performance status and were capable of tolerating a standard high-dose combination therapy with ribavirin.Specifically, patients who were younger than 65 years of age, had no evidence of cirrhosis, and had a sufficient platelet count (Ͼ 9.0 ϫ 10 4 /L) were considered good candidates for postoperative antiviral therapy.

Patient Follow-Up
All patients were regularly screened for recurrences through the evaluation of the HCC-specific tumor markers ␣-fetoprotein (AFP) and des-␥-carboxyprothrombin every 1 to 2 months, with ultrasonography every 2 months, and with dynamic computed tomography every 4 months, as previously reported. 22The HCV viral load was re-examined after surgery in possible candidates for adjuvant antiviral therapy.The function of the background liver was monitored using the serum ALT levels.If the ALT levels increased beyond 100 IU/L, an appropriate dose of ursodeoxycholic acid and/or monoammonium glycyrrhizinate was administered expecting their liver protective effects. 23,24ecurrence was defined as the appearance of a new lesion with radiologic features compatible with HCC, as confirmed using at least two imaging modalities.When a recurrence was detected, the patient received further treatment using a repeated hepatectomy, radiofrequency ablation, transcatheter arterial chemoembolization (TACE), or other treatment options, as indicated.
In the present study, recurrence-free survival (RFS) was defined as the interval between the operation and the date of the diagnosis of the first recurrence or the last follow-up examination, and overall survival (OS) was calculated based on the time from surgery to death or last follow-up.

Data Analysis
Statistical analysis was performed using SAS software, version 9.3 (SAS Institute, Cary, NC).Medians and ranges of continuous data were compared using the Mann-Whitney U test.Categorical data were compared using Pearson's 2 test or Fisher's exact test as appropriate.P Ͻ .05 was considered statistically significant.
High viral load was defined as HCV viral load to predict moderate to severe activity of hepatitis (grade 2 or 3 in Desmet classification 21 ).The cutoff value was determined using the minimum P value approach, and clinical outcomes were compared between the patients with a high viral load and those with a low viral load.In addition, the low viral load group was further subclassified according to whether or not HCV-RNA was detectable, and clinical outcomes between these subgroups were also compared.
Survival curves for OS and RFS were generated using the Kaplan-Meier method and were compared using the log-rank test.To identify risk factors for tumor recurrence, multivariate regression analysis was performed with the Cox proportional hazards model using a backward elimination procedure.To prevent overfitting, only factors that showed statistically significant association with tumor recurrence with P Ͻ .10 were included in the final model.Prognostic value of HCV viral load was

B A
Odds Ratio (95% CI) HCV Viral Load (log 10 IU/mL) quantified by comparing Harrell's concordance statistics of prognostic models based on the results of the multivariate analysis.

Characteristics of High and Low HCV Viral Load Groups
The best cutoff value of HCV viral load to predict moderate to severe activity of hepatitis was more than 5.3 log 10 IU/mL in both the plots of odds ratio and P value in the likelihood test (Fig 1).The background characteristics are compared between the high viral load (n ϭ 202) and low viral load (n ϭ 168) groups in Table 1.Female sex was more frequent in the high viral load group than in the low viral load group.The rates of coinfection with hepatitis B were not significantly different between the two groups.A history of IFN therapy was more common in the low viral load group.Number and maximum diameter of lesions were comparable between the two groups.The serum ALT and AST levels, indocyanine green retention rate at 15 minutes, and AFP levels were significantly higher in the high viral load group, whereas the platelet count was almost the same between the groups.
As for surgical factors (Table 2), the initial hepatectomy rates were 51.8% and 67.3% in the low and high viral load groups, respectively (P Ͻ .001).The remaining patients had repeat hepatectomies for recurrent lesions.Type of surgery (anatomic v nonanatomic), operating time, blood loss, and surgical margins were comparable between the groups.Histopathologically, no significant difference was observed in histologic grade of tumor or presence of vascular invasions.Fibrotic scores tended to be higher in the high viral load group.
Postoperative IFN therapy was performed in only six patients (3.6%) and seven patients (3.5%) in the low and high viral load groups, respectively.Because the median age of the patients in this study was 70 years and approximately 50% of the patients exhibited marked thrombocytopenia and/or cirrhotic changes in their background livers, the standard combination therapy of IFN with ribavirin was difficult to apply in most of the patients.
In 59 patients in whom postoperative viral load data were available, the HCV-RNA levels did not significantly change from baseline to 1 year after surgery (4.5 Ϯ 2.0 log 10 IU/mL before surgery v 4.6 Ϯ 1.9 log 10 IU/mL after surgery; P ϭ .78).HCV-RNA viral load at 1 year and postoperative mean ALT levels were higher in the high viral load group.Postoperative AFP levels were also higher in the high viral load group even after curative resection.

Patient Survival
The median follow-up time of the studied population was 38.4 months (range, 1 to 120 months), and no hospital deaths occurred.During the study period, recurrence was observed in 108 patients (60.7%) and 137 patients (71.4%) in the low and high viral load groups, respectively.
The median HCV viral load of the positive HCV-RNA subgroup in the patients with low viral load was 4.9 log 10 IU/mL (range, 2.3 to 5.3 log 10 IU/mL), and it was significantly lower than that in the high viral load group (P Ͻ .001).Clinicopathologic parameters were almost comparable between the two subgroups in the low viral group except that HCV-RNA titers and serum AST and ALT levels were significantly higher in positive HCV-RNA patients (P Ͻ .001).The 1-and 3-year RFS rates were similar between the two subgroups (65.6% and 38.8% for the negative HCV-RNA patients and 66.5% and 35.9% for the positive HCV-RNA patients, respectively; P ϭ .61;Fig 3A).The RFS rate among the low viral load group with positive HCV-RNA was superior to that of the high viral load group (P ϭ .009).A similar tendency was also observed in the OS rates.The positive HCV-RNA patients had relatively favorable results, similar to the negative HCV-RNA patients when the viral load was Յ 5.3 log 10 IU/mL.The 3-and 5-year OS rates were 85.8% and 78.1% for the negative HCV-RNA patients, respectively, and 89.0% and 75.8% for the positive HCV-RNA patients, respectively (P ϭ .94;Fig 3B).The OS rate of the low viral load group with positive HCV-RNA was superior to that of the high viral load group (P ϭ .005).These observations were constant when stratifying the study population according to hepatectomies for primary or recurrent lesions (Appendix Fig A1, online only).

Risk Factors for Postoperative Recurrence
Risk factors for postoperative recurrence were investigated in 357 patients without postoperative antiviral therapy.In the multivariate analysis, we chose 17 potential confounders considering their clinical significance and reported evidences, 4,[25][26][27][28][29][30][31] as indicated in Table 3.There were no specific combinations of factors suggesting multicollinearity in scatter plots.In multivariate analysis, high HCV viral load (Ͼ 5.3 log 10 IU/mL), macroscopic vascular invasion, repeat resection for recurrent tumor, tumor exposure, and tumor size greater than 2 cm were selected in the final model.The concordance statistic of the four-factor model (macroscopic vascular invasion ϩ repeat resection ϩ tumor exposure ϩ size Ͼ 2 cm) was 0.603 (95% CI, 0.559 to 0.647),

Impact of HCV Viral Load on Surgical Outcomes of HCC
and it improved to 0.627 (95% CI, 0.590 to 0.665) when including HCV viral load greater than 5.3 log 10 IU/mL in the prognostic model (Table 3).

DISCUSSION
In this study, we analyzed 370 patients who underwent curative liver resection for HCV-related HCC.The current study indicates that a low viral load (Յ 5.3 log 10 IU/mL) is strongly associated with lower recurrence rate and better OS regardless of the serologic eradication of HCV.These observations were constant both in initial hepatectomy for primary lesions and repeat hepatectomy for recurrent lesions.Multivariate analysis confirms that a high HCV viral load (Ͼ 5.3 log 10 IU/mL) is an independent factor associated with a 1.84-fold greater risk of tumor recurrence after curative resection of HCC.
In patients with HCV-related HCC, virologic status of HCV has been thought to be a prognostic factor associated with high tumor recurrence rate. 9,10Adjuvant antiviral therapy is performed with the aim of preventing tumor recurrence by improving the fibrotic status and/or activity of inflammation in the background liver through the reduction of the viral load.Several studies have shown favorable longterm outcomes of adjuvant IFN therapy after locoregional treatments or surgical resections. 11,32-34However, the effectiveness of antiviral therapy has been discussed mainly from the view point of virus eradication, 11-13,34-36 and little is known about the significance of the viral load itself for tumor recurrence.
Akamatsu et al 37 previously reviewed 371 patients who had undergone locoregional treatments for HCV-related HCC and denied a correlation between the viral load and the recurrence rate of HCC.However, their study contained a heterogeneous population that underwent Recurrence-Free Survival (proportion) Time (months)   various types of treatments including surgery, ablation, and TACE.Therefore, the true clinical influence of HCV viral load on long-term outcomes of HCV-related HCC is still unclear.In the current study, we carefully reviewed patients who underwent curative surgical resections under a consistent treatment strategy in a single high-volume hepatobiliary center.Major prognostic improvements were observed both in recurrence and survival when a low viral load was obtained according to the cutoff value (5.3 log 10 IU/mL) that was determined by the minimum P value approach to predict moderate to severe activity of hepatitis.Comparison of clinicopathologic factors revealed that high viral load was associated with higher serum ALT and AST levels (both before and after surgery) and higher fibrotic status.These correlations are consistent with previous reports 38,39 and suggest the higher carcinogenic potential in the background liver in patients with high HCV viral load.Another noteworthy result is that the preferable outcomes in the low viral load group are not significantly influenced by whether or not the serologic eradication of HCV is obtained.As shown in Figure 3, when the survival curves were compared between the RNA-positive and RNA-negative patients, no significant difference was observed, although both curves represented apparently better outcomes than that for the high viral load group.We also confirmed a similar tendency both in initial hepatectomy and repeat hepatectomy in a subset analysis (Appendix Figure 1, online only).These results suggest that a lower viral load might be preferable even if the serologic eradication of HCV is not obtained, supporting the outcomes of previous studies 14,15 and a recent meta-analysis 40 studying the effectiveness of IFN therapy.
Recent introduction of combination therapy consisting of pegylated IFN and ribavirin has dramatically improved the sustained viral response rate in patients with HCV. 32,33However, the postoperative use of IFN remains a major concern because HCC usually emerges in the liver that has been damaged over the course of decades, and accordingly, patients tend to be elderly and to exhibit cirrhotic changes.Therefore, a high-dose standard combination therapy is not always applicable because of the issue of tolerability.2][43][44] In fact, the median age of the current population was 70 years, and 47.8% of the patients were clinically diagnosed with cirrhosis.The proportion of women was higher in the high viral load group, and 71.4% of the patients had genotype 1b.
Given the current results, a low HCV viral load can be a new clinical end point in adjuvant therapy for HCV-related HCC.In this context, a more tolerable antiviral therapy, including low-dose IFN therapy with prolonged therapeutic duration 45,46 or possibly a combination with protease inhibitors, 47-49 may be a therapeutic option for elderly patients or patients with liver cirrhosis.Given the fact that anatomic resection of the liver was also an independent predictor of recurrence in the multivariate analysis, combination of anatomic resection and adjuvant IFN therapy may enhance the postoperative outcomes in patients with HCV-related HCC by eradicating micrometastases and reducing the carcinogenic potential in the underlying liver.
Because this study was retrospective, prospective/randomized trials are needed to confirm the true influence of the HCV viral load and the effectiveness of adjuvant antiviral therapy on postoperative outcomes.In addition, the results of the Sorafenib as Adjuvant Treatment in the Prevention of Recurrence of Hepatocellular Carcinoma (STORM) trial, if sorafenib is found to be of benefit, will impact the selection of postoperative therapy in HCC in the near future.Given the possibility of drug interactions and competing toxicity between sorafenib and antiviral agents, further investigation on the selection of adjuvant treatment is needed, especially in patients with HCV-associated HCC.
In conclusion, a low viral load may predict lower recurrence and better survival in patients undergoing hepatic resection for HCV-related HCC irrespective of the serologic eradication of HCV.Postoperative antiviral therapy with individually adjusted intensity and incorporation of direct antiviral agents may warrant prospective study to characterize safety and impact on recurrence risk in patients undergoing surgical resection for HCV-associated HCC.

Fig 2 .
Fig 2. (A) Cumulative recurrence rate and (B) cumulative overall survival curves of low and high viral load groups.

Fig 3 .
Fig 3. (A) Cumulative recurrence rate and (B) cumulative overall survival curves of low and high viral load groups stratified according to the results of hepatitis C virus RNA quantification.

Table 2 .
Surgical, Histopathologic, and Postoperative Factors Abbreviations: AFP, ␣-fetoprotein; DCP, des-␥-carboxyprothrombin; HCV, hepatitis C virus; HX, hepatectomy; IFN, interferon.‫ء‬Based on modification of the Edmondson grade.19†Based on the classification by Desmet et al.21‡Based on data from 31 and 28 patients with low viral load and high viral load, respectively.

Table 3 .
Factors Associated With Recurrence of Hepatocellular Carcinoma The concordance statistic for the full model (the four-factor model ϩ HCV-RNA Ͼ 5.3 log 10 IU/mL) was 0.627 (95% CI, 0.590 to 0.665).Multivariate Cox regression was applied with stepwise backward selection.Initially, all factors were included in the model.Then factors that showed no or limited statistically significant association (P Ͼ .01)withtumor recurrence adjusted for the remaining factors in the model were deleted from the model in a stepwise fashion.The 17 factors tested were as follows: sex, primary versus repeat resection, tumor size (Ͼ v Յ 2 cm), number of tumors (solitary v multiple), hepatitis B core antibody (yes v no), HCV viral load (Ͼ v Յ 5.3 log 10 IU/mL), fibrotic status of the underlying liver (F3-4 v F0-2), serum ALT level (Ͼ v Յ 40 IU/L), indocyanine green retention rate at 15 minutes (Ͼ v Յ 15%), serum ␣-fetoprotein level (Ͼ v Յ 20 ng/mL), plasma des-␥-carboxyprothrombin level (Ͼ v Յ 40 mAU/mL), type of hepatectomy (anatomic v nonanatomic), perioperative transfusion (yes v no), tumor exposure (yes v no), microvascular invasion (yes v no), macrovascular invasion (yes v no), and tumor differentiation (well/moderate v poor).Abbreviations: HCV, hepatitis C virus; HR, hazard ratio.test adjusted for the other factors in the final model.†Estimated coefficient for the variable and the associated SE.
‫ء‬Based on likelihood