Gang Huang, MD; Wan Yee Lau, MD, FRCS; Wei-ping Zhou, MD, PhD; Feng Shen, MD, PhD; Ze-ya Pan, MD; Sheng-xian Yuan, MD; Meng-chao Wu, MD

Research Original Investigation Prediction of Hepatocellular Carcinoma Recurrence in Patients With Low Hepatitis B Virus DNA Levels and High Preoperative Hepatitis B Surface Antigen Levels Gang Huang, MD; Wan Yee Lau, MD, FRCS; Wei-ping Zhou, MD, PhD; Feng Shen, MD, PhD; Ze-ya Pan, MD; Sheng-xian Yuan, MD; Meng-chao Wu, MD IMPORTANCE In patients with low viral loads, high levels of hepatitis B surface antigen (HBsAg) have been shown to predict development of hepatocellular carcinoma (HCC). Whether high levels of HBsAg increase the risk for HCC recurrence after hepatic resection remains unknown. OBJECTIVE To investigate the association between levels of HBsAg and the risk for tumor recurrence after curative resection in HCC patients with low levels of hepatitis B virus (HBV) DNA. DESIGN, SETTING, AND PARTICIPANTS We performed a retrospective analysis of the clinical data of 1062 patients with low HBV DNA levels (<200 IU/mL) who underwent partial hepatectomy for HCC. In particular, we investigated the association between levels of HBsAg and recurrence of HCC. EXPOSURE Partial hepatectomy for HCC. MAIN OUTCOMES AND MEASURES The risk for first tumor recurrence between patients with high and low HBsAg levels. We calculated cumulative incidences and hazard ratios after adjusting for competing mortality. RESULTS The risk for tumor recurrence increased with HBsAg levels of 1000 IU/mL or greater. When we compared the groups with low (<1000 IU/mL) and high ( 1000 IU/mL) HBsAg levels, the 5-year disease-free survival rate (46.1% vs 34.1% [P =.002]) and the overall survival rate (57.5% vs 48.8% [P =.004]) were better in the group with low HBsAg levels. On multivariate analysis, hepatitis B e antigen seropositivity, HBsAg level of 1000 IU/mL or greater, tumor size of greater than 5 cm, blood transfusion, surgical margin of less than 1.0 cm, the presence of satellite nodules, and the presence of portal vein tumor thrombus were independent risk factors for HCC recurrence. When compared with hepatitis B e antigen status, HBsAg level better predicted recurrence of HCC. CONCLUSIONS AND RELEVANCE A preoperative HBsAg level of 1000 IU/mL or greater is an independent risk factor for HCC recurrence in patients with low HBV DNA levels. JAMA Surg. 2014;149(6):519-527. doi:10.1001/jamasurg.2013.4648 Published online April 2, 2014. Author Affiliations: Department of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital, Shanghai, China (Huang, Lau, Zhou, Shen, Pan, Yuan, Wu); Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong Special Administrative Region (Lau). Corresponding Author: Wei-ping Zhou, MD, PhD, Department of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital, Shanghai 200438, China (ehphwp @126.com). 519

Research Original Investigation HCC Recurrence in Patients With Low HBsAg Levels Hepatocellular carcinoma (HCC) is the fifth most common cancer and the third most common cause of cancer-related death in the world. 1 Hepatic resection is a well-accepted therapy for HCC, but high postoperative tumor recurrence remains a major obstacle for long-term survival. 2 A 2008 study 3 found a high serum level of hepatitis B virus (HBV) DNA to be an important predictor of recurrence. Although patients with HBV DNA levels of less than 2000 IU/mL are usually designated as inactive or low-risk HBV carriers, 4,5 the postoperative tumor recurrence rate among these patients is still high. 6 Since 2010, hepatitis B surface antigen (HBsAg) quantification has become recognized increasingly as a marker for evaluating viral replication and possible host immune control over HBV infection. 7-9 However, few studies have evaluated HBsAg level as a predictor of postoperative recurrence of HCC, especially in patients with low HBV DNA levels. The primary aim of this study was to assess the significance of HBsAg level in predicting postoperative HCC recurrence in patients with low HBV load. Methods This retrospective study was conducted in patients with HBVrelated HCC who underwent partial hepatectomy at the liver unit of the Eastern Hepatobiliary Surgery Hospital from January 1, 2006, through December 31, 2007. Patients who were seropositive for HBsAg with a low preoperative HBV DNA level were included in this study. Hepatic resection was the first choice of curative treatment for patients with resectable HCC and good liver function. Liver transplant was considered in patients meeting the Milan criteria 10 who had poor liver function. However, owing to the shortage of donor organs, transplant was not readily available for these patients. The 97 patients who received a liver transplant in our hospital were excluded from this study. For patients with small HCC (<3 cm and <3 lesions), radiofrequency ablation (RFA) was the procedure of choice if they were not candidates for surgery or had poor liver function when a donor liver was not available. The study was approved by the Ethics Committee of the Eastern Hepatobiliary Hospital; informed consent was not required. Preoperative Investigations All patients underwent chest radiography, ultrasonography, and intravenous contrast enhanced computed tomography or magnetic resonance imaging of abdomen. Laboratory blood tests including measurement of HBsAg, hepatitis C virus antibody, serum α-fetoprotein, carcinoembryonic antigen, carbohydrate antigen 19-9, serum albumin, serum total bilirubin, aspartate aminotransferase, and alanine aminotransferase levels and prothrombin time were performed. Tests for viral replication, including hepatitis B e antigen (HBeAg) and its antibody, were performed. The serum HBV DNA level was quantified by polymerase chain reaction assay (ABI 7300; Applied Biosystems) with a linear range of quantification of 200 to 2 000 000 IU/mL. The lower limit of detection was 200 IU/mL. The HBsAg level was quantified using a commercially available assay (Architect; Abbott Laboratories) with a detection range of 0.05 to 250 IU/mL. If the HBsAg level was greater than 250 IU/mL, samples were diluted from 1:100 to 1:1000 to obtain a reading within the calibration curve range. The preoperative diagnosis of HCC was based on the diagnostic criteria for HCC used by the European Association for the Study of the Liver. 11 Further investigations were performed only when clinical evidence suggested the presence of extrahepatic metastases. Surgical Procedure Surgery was performed through a bilateral subcostal incision. The abdominal cavity was searched for extent of local disease, extrahepatic metastases, and peritoneal seeding. Intraoperative ultrasonography assessed the number and the size of the lesions and the relationship of the tumor to vascular structures. The Pringle maneuver was applied to occlude the blood inflow of the liver with cycles of 15 minutes clamped and 5 minutes unclamped. Liver transection was performed using a clamp-crushing method. Assessment of Outcome The resected tumor and surrounding liver were examined macroscopically and microscopically for histopathologic features. The histologic grade of cirrhosis and inflammation in the nontumorous liver were scored according to the classification of Ishak et al. 12 Curative resection was defined as complete macroscopic and microscopic removal of the tumor (R0 resection). The presence of satellite formation and vascular permeation in the tumor was determined. Follow-up Patients were followed up at monthly intervals within the first postoperative year and then every 3 months. At each follow-up visit, adverse events were documented and blood samples were taken for complete blood cell counts, coagulation profile, renal and liver function tests, and measurement of serum α-fetoprotein, HBsAg, HBeAg and its antibody, and HBV DNA levels. Ultrasonography was performed once every month. Chest radiography and computed tomography or magnetic resonance imaging were performed once every 3 months in the first year after surgery and thereafter at 6-month intervals. Recurrent lesions were managed aggressively using a multidisciplinary approach that included repeated resection, transarterial chemoembolization (TACE), percutaneous RFA, and percutaneous ethanol injection. The treatment was decided by the pattern of recurrence, liver functional reserve, and general condition of the patient at the time of recurrence. The primary end point of this study was the time from surgery to tumor recurrence. This study was censored on December 31, 2012. Statistical Analysis Virologic data were analyzed with conventional clinical variables at the time of partial hepatectomy to identify factors that influenced recurrence using the Cox proportional hazards model. Risk factors contributing to recurrence were investigated by stratified Cox regression analysis. Cumulative recur- 520 JAMA Surgery June 2014 Volume 149, Number 6 jamasurgery.com

HCC Recurrence in Patients With Low HBsAg Levels Original Investigation Research Table 1. Characteristics of HCC Patients Who Underwent Partial Hepatectomy a Characteristic Group 1 (n = 280) Group 2 (n = 342) Group 3 (n = 440) P Value Age, y 52.4 (9.6) 52.0 (10.6) 51.8 (10.2).67 Sex, No. M/F 242/38 297/45 392/48.49 HBeAg positivity, No. (%) 26 (9.3) 42 (12.3) 84 (19.1).001 AFP level, median (IQR), ng/ml 80.5 (1.5-1210.0) 147.9 (1.5-1210.0) 82.9 (1.2-1210.0).39 Preoperative laboratory test results Hemoglobin level, g/dl 14.4 (1.9) 14.2 (1.8) 14.4 (1.8).24 Platelet count, 10 3 /μl 145.3 (74.7) 148.2 (69.8) 152.2 (70.3).44 Prothrombin time, s 12.2 (1.1) 12.2 (1.2) 12.1 (1.1).18 Bilirubin level, mg/dl 0.9 (0.2) 0.9 (0.2) 0.9 (0.2).66 Albumin level, g/dl 4.0 (0.4) 4.0 (0.4) 4.0 (0.4).86 ALT level, U/L 37.1 (16.1) 35.4 (12.1) 38.6 (16.4).01 Creatinine level, mg/dl 0.8 (0.1) 0.8 (0.2) 0.8 (0.1).85 Child-Pugh grade, No. A/B 279/1 339/3 435/5.54 Tumor diameter, cm 5.4 (2.4) 5.6 (2.8) 5.3 (2.5).33 Presence of satellite nodules, No. (%) 16 (5.7) 31 (9.1) 38 (8.6).25 Presence of vascular invasion, No. (%) 18 (6.4) 24 (7.0) 34 (7.7).80 Inflammation score b 4.7 (2.1) 4.6 (1.8) 4.9 (2.2).10 Fibrosis score b 3.2 (1.1) 3.2 (1.1) 3.2 (1.0).95 Resection amount, No. (%) Minor 225 (80.4) 267 (78.1) 345 (78.4) Major 55 (19.6) 75 (21.9) 95 (21.6).76 Operating time, min 181.4 (38.9) 180.0 (47.0) 178.2 (38.1).59 Clamp time, min 25.9 (8.4) 25.4 (9.0) 26.0 (8.3).61 Blood loss, median (range), ml 400 (100-3600) 400 (100-5400) 400 (100-4000).94 Transfusion, No. (%) 49 (17.5) 47 (13.7) 72 (16.4).41 Preoperative TACE, No. (%) 19 (6.8) 31 (9.1) 46 (10.5).25 Abbreviations: AFP, α-fetoprotein; ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; HCC, hepatocellular carcinoma; TACE, transarterial chemoembolization. SI conversion factors: To convert AFP to micrograms per liter, multiply by 1; albumin to grams per liter, multiply by 10; ALT to microkatals per liter, multiply by 0.0167; bilirubin to micromoles per liter, multiply by 17.104; creatinine to micromoles per liter, multiply by 88.4; hemoglobin to grams per liter, multiply by 10; and platelets to 10 9 cells per liter, multiply by 1. a Groups classified as having hepatitis B surface antigen levels of less than 100 IU/mL (group 1), 100 to 999 IU/mL (group 2), and 1000 IU/mL or greater (group 3). Unless otherwise indicated, data are expressed as mean (SD). b Based on the classification of Ishak et al. 12 rence rates were calculated by the Kaplan-Meier method, and differences were compared by the log-rank test. Multivariate analysis was performed by the Cox proportional hazards regression model. Statistical significance was defined as P <.05. Statistical analysis was performed using commercially available software (SPSS, version 18.0, for Windows; SPSS, Inc). Results During the study period, partial hepatectomy was performed in 2860 patients with a preoperative diagnosis of HBV-related HCC in our center. We excluded patients who were seropositive for HBsAg and hepatitis C virus antibodies (45 [1.6%]), who did not have a diagnosis of HCC on results of histopathologic examination of the resected specimen (202 [7.1%]), who had early tumor recurrence or metastasis within 1 month of surgery (81 [2.8%]), who died within 1 month of surgery (27 [0.9%]), or who were lost to or unavailable for follow-up within 6 months of surgery (69 [2.4%]). Compared with a previous study, 13 we chose a tougher standard with an HBV DNA level of less than 200 IU/mL as the low level in this study. Of the remaining 2436 patients who were seropositive for HBsAg, 1267 had a low preoperative HBV DNA level. Antiviral therapy was given to 205 of these 1267 patients (16.2%) before HCC was diagnosed or during the study period. Because antiviral treatment can modify the risk for recurrence of HCC, these patients were also excluded from this study. We included 1062 patients who received no adjuvant therapy. The patients were divided into the following 3 groups according to the following preoperative HBsAg levels: less than 100 IU/mL (group 1), 100 to 999 IU/mL (group 2), and 1000 IU/mL or greater (group 3). The baseline characteristics of these patients are presented in Table 1. Except for the HBeAg seropositivity rate and preoperative serum alanine aminotransferase level, we found no significant difference in the variables among the 3 groups of patients. During a median follow-up of 60 (range, 4-84) months, 49 patients (4.6%) were lost to or unavailable for follow-up, 605 (57.0%) developed tumor recurrence, and 508 (47.8%) died. For those patients with tumor recurrence, 89 (14.7%) underwent a repeated resection; 332 (54.9%), percutaneous RFA; 15 (2.5%), percutaneous ethanol injection; 43 (7.1%), TACE; 79 (13.1%), percutaneous RFA combined with TACE; and 34 (5.6%), percutaneous ethanol injection combined with TACE. The remaining 13 patients (2.1%) were not given any anticancer treatment because of deteriorating liver function or poor general condition. The prognosis of patients who received no anticancer treatment was extremely poor, probably because of selection of patients with poor conditions. The prognosis of patients who received TACE was significantly worse than that of the patients who re- jamasurgery.com JAMA Surgery June 2014 Volume 149, Number 6 521

Research Original Investigation HCC Recurrence in Patients With Low HBsAg Levels Figure 1. Comparison of Disease-Free Survival Among 3 Subgroups Figure 3. Comparison of Overall Survival Between 2 Subgroups 100 90 80 Group 1 Group 2 Group 3 100 90 80 Low HBsAg level High HBsAg level Disease-Free Survival, % 70 60 50 40 30 Overall Survival, % 70 60 50 40 30 20 20 10 10 0 0 12 24 36 48 60 72 84 Follow-up, mo 0 0 12 24 36 48 60 72 84 Follow-up, mo Patients are divided into groups by level of hepatitis B surface antigen as follows: less than 100 IU/mL (group 1 [n = 280]), 100 to 999 IU/mL (group 2 [n = 342]), and 1000 IU/mL or greater (group 3 [n = 440]). For comparison of survival curves, P =.006, log-rank test. Figure 2. Comparison of Disease-Free Survival Between 2 Subgroups Disease-Free Survival, % 100 90 80 70 60 50 40 30 20 10 0 0 12 24 36 48 60 72 84 Follow-up, mo Low HBsAg level High HBsAg level Patientgroupsareclassifiedashavinglow(<1000IU/mL[n = 622])andhigh( 1000 IU/mL [n = 440]) hepatitisbsurfaceantigen (HBsAg) levels. P =.002, log-ranktest. ceived other types of treatment (P <.05), probably because TACE was used in patients with multiple recurrent HCC. No significant difference was observed in survival among the groups of patients who were treated with repeated resection, percutaneous RFA, or percutaneous ethanol injection (P =.09). The 1-, 3-, and 5-year cumulative disease-free survival (DFS) rates were 81.6%, 47.6%, and 41.1%, respectively. The corresponding cumulative overall survival (OS) rates were 94.2%, 67.6%, and 53.9%, respectively. Of the 508 patients who died, 461 (90.7%) died of tumor recurrence; 40 (7.9%), liver failure; 5 (1.0%), complications of portal hypertension; and 2 (0.4%), other causes. Patient groups are classified as having low (<1000 IU/mL [n = 622]) and high ( 1000 IU/mL [n = 440]) hepatitis B surface antigen (HBsAg) levels. P =.004, log-rank test. Comparison of DFS and OS by Preoperative HBsAg Level When the patients were divided into the 3 groups according to the preoperative HBsAg level, the 5-year DFS rates were 48.9%, 43.8%, and 34.1%, respectively; the 5-year OS rates were 59.3%, 56.0%, and 48.8%, respectively. The results suggested that lower HBsAg levels resulted in better prognosis (Figure 1). The DFS for groups 1 and 2 was significantly better than that for group 3 (hazard ratio [HR], 0.74 [95% CI, 0.60-0.90; P =.003] and 0.82 [0.68-0.99; P =.03], respectively), but differences between groups 1 and 2 were not significant (0.90 [0.72-1.12; P =.35]). Thus, the risk for HCC recurrence was similar in patients with HBsAg levels of less than 1000 IU/mL (low level), but it increased when the level was 1000 IU/mL or greater (high level). When the patients were divided only into those 2 subgroups, the 5-year DFS rates for the subgroup with low HBsAg levels was significantly better than that for the subgroup with high HBsAg levels (46.1% vs 34.1% [P =.002]) (Figure 2). The 5-year OS rates for the subgroup with low HBsAg levels was also significantly better than that for the subgroup with high HBsAg levels (57.5% vs 48.8% [P =.004]) (Figure 3). Risk of HCC Recurrence in Patients With Low Preoperative HBV DNA Levels The factors that were significantly associated with a high cumulative risk for tumor recurrence on univariate analysis were entered into multivariate analysis (Table 2). Seropositivity for HBeAg (HR, 1.30 [95% CI, 1.04-1.62; P =.02]), HBsAg levels of 1000 IU/mL or greater (1.23 [1.04-1.44; P =.01]), tumor size of greater than 5 cm (1.34 [1.12-1.59; P =.001]), blood transfusion (1.45 [1.04-2.03; P =.03]), surgical margin of less than 1.0 cm (1.22 [1.03-1.44; P =.02]), the presence of satellite nodules (2.64 [2.04-3.41; P <.001]), and the presence of portal vein tumor thrombus (6.06 [4.57-8.03; P <.001]) were significant factors associated with a high risk for HCC recurrence after partial hepatectomy. 522 JAMA Surgery June 2014 Volume 149, Number 6 jamasurgery.com

HCC Recurrence in Patients With Low HBsAg Levels Original Investigation Research Table 2. Univariate and Multivariate Analyses for Recurrence of HCC in Patients With Low HBV DNA Levels a No. (%) of Factor Patients Sex Male 931 (87.7) Female 131 (12.3) Age, y >65 127 (12.0) 65 935 (88.0) HBeAg seropositivity Yes 152 (14.3) No 910 (85.7) HBsAg level, IU/mL 1000 440 (41.4) <1000 622 (58.6) ALT level, U/L 40 403 (37.9) <40 659 (62.1) Total bilirubin level, mg/dl 1.0 350 (33.0) <1.0 712 (67.0) Prothrombin time, s 13 239 (22.5) <13 823 (77.5) Albumin level, g/dl <3.5 154 (14.5) 3.5 908 (85.5) AFP level, ng/ml >100 540 (50.8) 100 522 (49.2) Preoperative TACE Yes 96 (9.0) No 966 (91.0) Liver resection Major 225 (21.2) Minor 837 (78.8) Tumor diameter, cm >5 417 (39.3) 5 645 (60.7) Clamp time, min >30 283 (26.6) 30 779 (73.4) Operation time, min >180 624 (58.8) 180 438 (41.2) Blood loss, ml >1000 109 (10.3) 1000 953 (89.7) Transfusion Yes 168 (15.8) No 894 (84.2) Surgical margin, cm <1.0 555 (52.3) 1.0 507 (47.7) Univariate Analysis Multivariate Analysis HR (95% CI) P Value HR (95% CI) P Value 1.17 (0.90-1.50).24 0.86 (0.66-1.11).23 1.25 (1.01-1.55).04 1.30 (1.04-1.62).02 1.28 (1.09-1.50).003 1.23 (1.04-1.44).01 1.17 (0.99-1.37).06 0.93 (0.79-1.11).43 1.04 (0.86-1.25).70 0.83 (0.65-1.06).13 1.14 (0.97-1.33).12 1.31 (1.01-1.70).04 1.15 (0.88-1.50).31 1.30 (1.08-1.57).06 0.89 (0.73-1.09).25 1.64 (1.40-1.93) <.001 1.34 (1.12-1.59).001 1.46 (1.23-1.74) <.001 1.03 (0.84-1.26).78 1.35 (1.14-1.59) <.001 1.13 (0.94-1.35).18 2.10 (1.66-2.65) <.001 1.15 (0.78-1.68).48 2.07 (1.70-2.53) <.001 1.45 (1.04-2.03).03 1.43 (1.21-1.68) <.001 1.22 (1.03-1.44).02 (continued) jamasurgery.com JAMA Surgery June 2014 Volume 149, Number 6 523

Research Original Investigation HCC Recurrence in Patients With Low HBsAg Levels Table 2. Univariate and Multivariate Analyses for Recurrence of HCC in Patients With Low HBV DNA Levels a (continued) No. (%) of Factor Patients Presence of satellite nodules Yes 85 (8.0) No 977 (92.0) Presence of portal vein tumor thrombus Yes 76 (7.2) No 986 (92.8) Inflammation score b >6 188 (17.7) 6 874 (82.3) Fibrosis score b 4 410 (38.6) <4 652 (61.4) Univariate Analysis Multivariate Analysis HR (95% CI) P Value HR (95% CI) P Value 3.27 (2.56-4.18) <.001 2.64 (2.04-3.41) <.001 8.30 (6.40-10.76) <.001 6.06 (4.57-8.03) <.001 1.42 (1.17-1.73) <.001 1.213 (0.988-1.489).07 1.05 (0.89-1.24).56 Abbreviations: AFP, α-fetoprotein; ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HR, hazard ratio; TACE, transarterial chemoembolization. SI conversion factors: To convert AFP to micrograms per liter, multiply by 1; albumin to grams per liter, multiply by 10; ALT to microkatals per liter, multiply by 0.0167; and bilirubin to micromoles per liter, multiply by 17.104. a The factors that were significantly associated with a high cumulative risk for of tumor recurrence on univariate analysis were entered into multivariate analysis. b Based on the classification of Ishak et al. 12 Risk of Death in Patients With Low Preoperative HBV DNA Levels The factors that were significantly associated with a high cumulative risk for death on univariate analysis were entered into multivariate analysis (Table 3). Seropositivity for HBeAg (HR, 1.29 [95% CI, 1.02-1.63; P =.03]), HBsAg level of 1000 IU/mL or greater (1.26 [1.06-1.51; P =.01]), tumor size of greater than 5 cm (1.37 [1.13-1.65; P =.001]), surgical margin of less than 1.0 cm (1.24 [1.04-1.49; P =.02]), the presence of satellite nodules (2.08 [1.58-2.75; P <.001]), the presence of portal vein tumor thrombus (3.36 [2.53-4.45; P <.001]), and based on the classification of Ishak et al, 12 inflammation score of greater than 6 (1.32 [1.06-1.64; P =.01]) and fibrosis score of 4 or greater (1.47 [1.23-1.75; P <.001]) were significant factors associated with a high risk for death after partial hepatectomy. Comparison of HBeAg Status and HBsAg Levels for Predicting HCC Recurrence Because the preoperative HBeAg status and HBsAg levels were associated with HCC recurrence, we used the area under the receiver operating characteristic (ROC) curve to compare these 2 predictors. The area under the ROC curve was 0.53 (compared with 0.49-0.56) for HBeAg seropositivity and 0.56 (compared with 0.52-0.59) for HBsAg level. The ROC curve analysis showed HBsAg level to be superior to HBeAg seropositivity in predicting recurrence of HCC in patients who had a low preoperative HBV DNA level. Multivariate Analysis for Early and Late Recurrences Of the 1062 patients, 457 developed early HCC recurrence (ie, <2 years after liver resection). On multivariate analysis, tumor size (HR, 1.40 [95% CI, 1.14-1.71; P =.001]), operation time (1.25 [95% CI, 1.01-1.54; P =.04]), blood transfusion (1.58 [95% CI, 1.11-2.26; P =.01]), surgical margin of less than 1.0 cm (1.39 [1.14-1.69; P =.001]), the presence of satellite nodules (2.80 [2.14-3.68; P <.001]), and the presence of portal vein tumor thrombus (5.51 [4.13-7.34; P <.001]) weresignificantlyassociatedwithearly tumor recurrence. A high HBsAg level (HR, 1.08 [95% CI, 0.90-1.30; P =.41]) was not associated with a higher risk for early tumor recurrence. However, on analysis of the risk factors associated with late recurrence in the remaining 552 patients (excluding 457 patients with early tumor recurrence and 53 patients without early tumor recurrence who had died or were lost to or unavailable for follow-up within 2 years), a high HBsAg level (HR, 1.93 [95% CI, 1.38-2.70; P <.001]) was an independent risk factor of late tumor recurrence. Discussion Even after curative partial hepatectomy, the risk for tumor recurrence is extremely high. 14 Previous studies 3,15-17 have shown that high levels of HBV DNA and seropositivity for HBeAg are closely associated with HCC development and recurrence. The challenge is to find a good marker that can predict postoperative prognosis in HCC patients with a low HBV DNA load. Recent advances in HBsAg assay using a quantitative measure have led to the exploration of its potential role in indicating active HBV infection and predicting treatment outcome. 18 A recent study 19 showed a high HBsAg level ( 1000 IU/mL) can be used as an independent risk factor for HCC development, suggesting that HBsAg level may be used to complement HBV DNA level in predicting HCC recurrence, especially in patients with low HBV DNA levels. During the 10 years since 2004, HBsAg level has been shown to be a steady, reliable but unspectacular marker of active HBV infection. The HBsAg level was positively correlated with HBV DNA level and with the rate of HBeAg seropositivity. 9 However, the correlations have been shown to be lowest at the lowly replicative phase, 20 which was consistent with our results. This discrepancy between the levels of HBsAg and HBV DNA at the lowly replicative phase might be caused by an accumulation of integrated viral envelope sequences in infected hepatocytes. Hepatitis B surface antigen is derived mainly from the integrated form of HBV DNA rather than the episomal form. 18 In other words, patients with low viremia who have a high HBsAg level might harbor more hepatocytes with HBV integration than those who have a low HBsAg level. 524 JAMA Surgery June 2014 Volume 149, Number 6 jamasurgery.com

HCC Recurrence in Patients With Low HBsAg Levels Original Investigation Research Table 3. Univariate and Multivariate Analyses for Overall Survival of HCC in Patients With Low HBV DNA Levels a No. (%) of Factor Patients Sex Male 931 (87.7) Female 131 (12.3) Age, y >65 127 (12.0) 65 935 (88.0) HBeAg positivity Yes 152 (14.3) No 910 (85.7) HBsAg level, IU/mL 1000 440 (41.4) <1000 622 (58.6) ALT level, U/L 40 403 (37.9) <40 659 (62.1) Total bilirubin level, mg/dl 1.0 350 (33.0) <1.0 712 (67.0) Prothrombin time, s 13 239 (22.5) <13 823 (77.5) Albumin level, g/dl <3.5 154 (14.5) 3.5 908 (85.5) AFP level, ng/ml >100 540 (50.8) 100 522 (49.2) Preoperative TACE Yes 96 (9.0) No 966 (91.0) Liver resection Major 225 (21.2) Minor 837 (78.8) Tumor diameter, cm >5 417 (39.3) 5 645 (60.7) Clamp time, min >30 283 (26.6) 30 779 (73.4) Operation time, min >180 624 (58.8) 180 438 (41.2) Blood loss, ml >1000 109 (10.3) 1000 953 (89.7) Transfusion Yes 168 (15.8) No 894 (84.2) Surgical margin, cm <1.0 555 (52.3) 1.0 507 (47.7) Univariate Analysis Multivariate Analysis HR (95% CI) P Value HR (95% CI) P Value 1.19 (0.90-1.58).21 0.90 (0.68-1.18).44 1.32 (1.05-1.66).02 1.29 (1.02-1.63).03 1.29 (1.08-1.53).005 1.26 (1.06-1.51).01 1.16 (0.97-1.39).10 1.12 (0.93-1.34).24 1.21 (0.99-1.47).07 1.06 (0.83-1.35).67 0.99 (0.83-1.18).92 1.16 (0.87-1.55).31 1.11 (0.90-1.37).33 1.68 (1.41-2.00) <.001 1.37 (1.13-1.65).001 1.50 (1.25-1.81) <.001 1.08 (0.87-1.34).51 1.28 (1.07-1.53).007 1.02 (0.84-1.24).86 2.01 (1.56-2.57) <.001 1.41 (0.93-2.14).11 1.81 (1.46-2.25) <.001 1.06 (0.73-1.52).77 1.41 (1.18-1.68) <.001 1.24 (1.04-1.49).02 (continued) jamasurgery.com JAMA Surgery June 2014 Volume 149, Number 6 525

Research Original Investigation HCC Recurrence in Patients With Low HBsAg Levels Table 3. Univariate and Multivariate Analyses for Overall Survival of HCC in Patients With Low HBV DNA Levels a (continued) No. (%) of Factor Patients Presence of satellite nodules Yes 85 (8.0) No 977 (92.0) Presence of portal vein tumor thrombus Yes 76 (7.2) No 986 (92.8) Inflammation score b >6 188 (17.7) 6 874 (82.3) Fibrosis score b 4 410 (38.6) <4 652 (61.4) Univariate Analysis Multivariate Analysis HR (95% CI) P Value HR (95% CI) P Value 2.47 (1.89-3.24) <.001 2.08 (1.58-2.75) <.001 4.69 (3.62-6.08) <.001 3.36 (2.53-4.45) <.001 1.46 (1.18-1.80).001 1.32 (1.06-1.64).01 1.37 (1.15-1.63) <.001 1.47 (1.23-1.75) <.001 Abbreviations: See Table 2. SI conversion factors: See Table 2. a The factors that were significantly associated with a high cumulative risk for of tumor recurrence on univariate analysis were entered into multivariate analysis. b Based on the classification of Ishak et al. 12 In our study, the factors that might have contributed to HCC recurrence in patients with a low HBV DNA load of less than 200 IU/mL, in addition to tumor factors and surgical factors, include viral factors such as HBeAg seropositivity and HBsAg levels of 1000 IU/mL or greater. On ROC curve analysis, HBsAg level was superior to HBeAg seropositivity in predicting HCC recurrence. Although these observations need to be confirmed in further studies, they can be explained. First, sustained viremia and subsequent active viral replication may contribute to the carcinogenic process. Patients with low HBV DNA levels are not true inactive carriers, and they are still at a considerable risk of developing HCC. 13,18,20,21 Previous studies 7,22,23 have also indicated that a low HBsAg level was associated with better clinical outcomes, including a higher likelihood of HBsAg loss, a lower risk of HBeAg-seronegative hepatitis, and a sustained viral suppression. As our findings showed, even in patients with an HBV DNA level of less than 200 IU/mL, a high HBsAg level was also a risk factor for tumor recurrence. Therefore, a combination of low HBsAg and HBV DNA levels might allow a more accurate identification of the true inactive and minimal-risk HBV carriers. 24,25 The usefulness of the cutoff HBsAg level as reported in our study has been confirmed by a study of patients from Korea and Hong Kong. 26 These findings strongly supported the indication of adequate host immune control against HBV infection by a low HBsAg (<1000 IU/ ml) level, thus leading to a decrease in the risk for HCC recurrence over time. Second, the HBsAg level is positively correlated not just with the serum HBV DNA and HBeAg levels, 27 but with the HBV covalently closed circular (ccc) DNA level, 28-30 which is the viral template for HBV replication in the maintenance of chronic infection and the predominant form of intrahepatic HBV DNA in the late stage. 31 A previous study 32 found that tumor tissues had a significantly higher level of cccdna when compared with the nontumor counterpart, and most of the tumor tissues had intrahepatic HBV DNA solely in the form of cccdna. Thus, cccdna might play a direct role in hepatocarcinogenesis. A decline in serum HBV DNA level reflects a reduction in viral replication. In contrast, a decline in serum HBsAg level represents a reduction in the translation of messenger RNAs produced from transcriptionally active cccdna or integrated sequences. 18 Thus, HBsAg quantitation provides different but complementary information that may help to characterize an individual s infection status. Recurrence of HCC in patients with a high HBsAg level might be attributed to increased genomic instability as a result of integrated viral sequences, which play an important role in hepatocarcinogenesis. 33 Third, the serum HBV DNA level may fluctuate during the course of chronic infection, 34 but the HBsAg level is relatively stable and endurable. 35,36 This supposition is supported by results of natural history studies that showed increased survival, lower rates of hepatic decompensation, and decreased frequency of HCC in patients who have cleared their HBsAg. 37,38 In addition, in patients with a low HBV DNA level, the dynamic range of HBsAg level is wider than the HBV DNA level. Thus, the HBsAg level is more powerful than the HBV DNA level or HBeAg status to differentiate patients at different risks of HCC recurrence. Our DFS and OS results are somewhat disappointing. The reason may be that the difference in early tumor recurrence was not significant (P =.41), but the late recurrence rate was significantly different (P <.001) between the groups with low and high HBsAg levels. Therefore, the HBsAg level is primarily associated with late tumor recurrence, and perhaps we will observe an exciting result after the longer follow-up. Conclusions The most important finding of this study is that a preoperative HBsAg load of 1000 IU/mL or greater was independently associated with tumor recurrence in patients with a low HBV DNA load. Quantitative measurement of HBsAg level can be used as a new prognostic factor of HCC recurrence after partial hepatectomy in patients with a low HBV DNA level. A future study should examine whether patients with low HBV DNA but high HBsAg levels should receive antiviral treatment. 526 JAMA Surgery June 2014 Volume 149, Number 6 jamasurgery.com

HCC Recurrence in Patients With Low HBsAg Levels Original Investigation Research ARTICLE INFORMATION Accepted for Publication: July 29, 2013. Published Online: April 2, 2014. doi:10.1001/jamasurg.2013.4648. Author Contributions: Drs Huang and Zhou had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Huang, Lau, Zhou, Shen. Acquisition, analysis, or interpretation of data: Huang, Shen, Pan, Yuan, Wu. Drafting of the manuscript: Huang, Lau, Pan. Critical revision of the manuscript for important intellectual content: Lau, Zhou, Shen, Yuan, Wu. Statistical analysis: Lau, Shen, Pan. Obtained funding: Huang, Zhou, Shen. Administrative, technical, or material support: Zhou, Shen. Study supervision: Huang, Lau, Shen. Conflict of Interest Disclosures: None reported. Funding/Support: This study was supported by grants 2012ZX10002010 and 2012ZX10002016 from the State Key Project on Infectious Diseases of China; by contract 81201940 from the Science Fund for Creative Research Groups, National Natural Science Foundation of China (NSFC); and by contract 81201555 from the NSFC. Role of the Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. REFERENCES 1. Llovet JM, Burroughs A, Bruix J. Hepatocellular carcinoma.lancet. 2003;362(9399):1907-1917. 2. Tung-Ping Poon R, Fan ST, Wong J. Risk factors, prevention, and management of postoperative recurrence after resection of hepatocellular carcinoma.ann Surg. 2000;232(1):10-24. 3. Hung IF, Poon RT, Lai CL, Fung J, Fan ST, Yuen MF. Recurrence of hepatitis B-related hepatocellular carcinoma is associated with high viral load at the time of resection. Am J Gastroenterol. 2008;103(7):1663-1673. 4. Liaw YF, Chu CM. Hepatitis B virus infection. Lancet. 2009;373(9663):582-592. 5. Villa E, Fattovich G, Mauro A, Pasino M. Natural history of chronic HBV infection: special emphasis on the prognostic implications of the inactive carrier state versus chronic hepatitis. Dig Liver Dis. 2011;43(suppl 1):S8-S14. 6. Chuma M, Hige S, Kamiyama T, et al. The influence of hepatitis B DNA level and antiviral therapy on recurrence after initial curative treatment in patients with hepatocellular carcinoma.jgastroenterol. 2009;44(9):991-999. 7. Brunetto MR, Oliveri F, Colombatto P, et al. Hepatitis B surface antigen serum levels help to distinguish active from inactive hepatitis B virus genotype D carriers. Gastroenterology. 2010;139 (2):483-490. 8. Nguyen T, Thompson AJ, Bowden S, et al. Hepatitis B surface antigen levels during the natural history of chronic hepatitis B: a perspective on Asia. J Hepatol. 2010;52(4):508-513. 9. Su TH, Hsu CS, Chen CL, et al. Serum hepatitis B surface antigen concentration correlates with HBV DNA level in patients with chronic hepatitis B. Antivir Ther. 2010;15(8):1133-1139. 10. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med. 1996;334(11):693-699. 11. Bruix J, Sherman M, Llovet JM, et al; EASL Panel of Experts on HCC; European Association for the Study of the Liver. Clinical management of hepatocellular carcinoma: conclusions of the Barcelona-2000 EASL conference. J Hepatol.2001; 35(3):421-430. 12. Ishak K, Baptista A, Bianchi L, et al. Histological grading and staging of chronic hepatitis. J Hepatol. 1995;22(6):696-699. 13. Chen JD, Yang HI, Iloeje UH, et al; Risk Evaluation of Viral Load Elevation and Associated Liver Disease/Cancer in HBV (REVEAL-HBV) Study Group. Carriers of inactive hepatitis B virus are still at risk for hepatocellular carcinoma and liver-related death. Gastroenterology. 2010;138(5): 1747-1754. 14. Poon RT, Fan ST, Lo CM, Liu CL, Ng IO, Wong J. Long-term prognosis after resection of hepatocellular carcinoma associated with hepatitis B-related cirrhosis. J Clin Oncol.2000;18(5):1094-1101. 15. Yang HI, Lu SN, Liaw YF, et al; Taiwan Community-Based Cancer Screening Project Group. Hepatitis B e antigen and the risk of hepatocellular carcinoma. N Engl J Med. 2002;347(3):168-174. 16. Sun HC, Zhang W, Qin LX, et al. Positive serum hepatitis B e antigen is associated with higher risk of early recurrence and poorer survival in patients after curative resection of hepatitis B-related hepatocellular carcinoma. J Hepatol. 2007;47(5): 684-690. 17. Chan HL, Tse CH, Mo F, et al. High viral load and hepatitis B virus subgenotype ce are associated with increased risk of hepatocellular carcinoma. J Clin Oncol. 2008;26(2):177-182. 18. Brunetto MR. A new role for an old marker, HBsAg. J Hepatol. 2010;52(4):475-477. 19. Tseng TC, Liu CJ, Yang HC, et al. High levels of hepatitis B surface antigen increase risk of hepatocellular carcinoma in patients with low HBV load.gastroenterology. 2012;142(5):1140-1149.e3, e13-e4. doi:10.1053/j.gastro.2012.02.007. 20. Thompson AJ, Nguyen T, Iser D, et al. Serum hepatitis B surface antigen and hepatitis B e antigen titers: disease phase influences correlation with viral load and intrahepatic hepatitis B virus markers. Hepatology. 2010;51(6):1933-1944. 21. Fung J, Yuen MF, Yuen JC, Wong DK, Lai CL. Low serum HBV DNA levels and development of hepatocellular carcinoma in patients with chronic hepatitis B: a case-control study. Aliment Pharmacol Ther. 2007;26(3):377-382. 22. Chan HL, Wong GL, Tse CH, Chan HY, Wong VW. Viral determinants of hepatitis B surface antigen seroclearance in hepatitis B e antigen-negative chronic hepatitis B patients. J Infect Dis. 2011;204(3):408-414. 23. Tseng TC, Liu CJ, Su TH, et al. Serum hepatitis B surface antigen levels predict surface antigen loss in hepatitis B e antigen seroconverters. Gastroenterology. 2011;141(2):517-525.e2. doi:10.1053/j.gastro.2011.04.046. 24. Chen CH, Lee CM, Wang JH, Tung HD, Hung CH, Lu SN. Correlation of quantitative assay of hepatitis B surface antigen and HBV DNA levels in asymptomatic hepatitis B virus carriers. Eur J Gastroenterol Hepatol. 2004;16(11):1213-1218. 25. Tseng TC, Liu CJ, Yang HC, et al. Serum hepatitis B surface antigen levels help predict disease progression in patients with low hepatitis B virus loads.hepatology. 2013;57(2):441-450. 26. Yang HI, Yuen MF, Chan HL, et al; REACH-B Working Group. Risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B): development and validation of a predictive score. Lancet Oncol. 2011;12(6):568-574. 27. Ozaras R, Tabak F, Tahan V, et al. Correlation of quantitative assay of HBsAg and HBV DNA levels during chronic HBV treatment. Dig Dis Sci. 2008;53 (11):2995-2998. 28. Wursthorn K, Lutgehetmann M, Dandri M, et al. Peginterferon alpha-2b plus adefovir induce strong cccdna decline and HBsAg reduction in patients with chronic hepatitis B. Hepatology. 2006;44(3): 675-684. 29. Chan HL, Wong VW, Tse AM, et al. Serum hepatitis B surface antigen quantitation can reflect hepatitis B virus in the liver and predict treatment response. Clin Gastroenterol Hepatol.2007;5(12): 1462-1468. 30. Chan HL, Wong VW, Wong GL, Tse CH, Chan HY, Sung JJ. A longitudinal study on the natural history of serum hepatitis B surface antigen changes in chronic hepatitis B. Hepatology. 2010;52 (4):1232-1241. 31. Wong DK, Yuen MF, Yuan H, et al. Quantitation of covalently closed circular hepatitis B virus DNA in chronic hepatitis B patients. Hepatology. 2004;40 (3):727-737. 32. Wong DK, Yuen MF, Poon RT, Yuen JC, Fung J, Lai CL. Quantification of hepatitis B virus covalently closed circular DNA in patients with hepatocellular carcinoma. J Hepatol. 2006;45(4):553-559. 33. Kao JH, Chen PJ, Chen DS. Recent advances in the research of hepatitis B virus-related hepatocellular carcinoma: epidemiologic and molecular biological aspects. Adv Cancer Res.2010; 108:21-72. 34. Chu CJ, Hussain M, Lok AS. Quantitative serum HBV DNA levels during different stages of chronic hepatitis B infection. Hepatology. 2002;36(6): 1408-1415. 35. Ganem D, Prince AM. Hepatitis B virus infection: natural history and clinical consequences. N Engl J Med. 2004;350(11):1118-1129. 36. Liang TJ. Hepatitis B: the virus and disease. Hepatology. 2009;49(5)(suppl):S13-S21. 37. Manno M, Cammà C, Schepis F, et al. Natural history of chronic HBV carriers in northern Italy: morbidity and mortality after 30 years. Gastroenterology. 2004;127(3):756-763. 38. Fattovich G, Olivari N, Pasino M, D Onofrio M, Martone E, Donato F. Long-term outcome of chronic hepatitis B in Caucasian patients: mortality after 25 years. Gut. 2008;57(1):84-90. jamasurgery.com JAMA Surgery June 2014 Volume 149, Number 6 527