Hepatocellular carcinoma (HCC) is the most

Transcription

1 Original Article / Biliary Hepatobiliary & Pancreatic Diseases International Prognosis of hepatocellular carcinoma with bile duct tumor thrombus after R0 resection: a matched study Ding-Ding Wang, Li-Qun Wu and Zu-Sen Wang Qingdao, China BACKGROUND: Hepatocellular carcinoma (HCC) with bile duct tumor thrombus (BDTT) is rare. The present study aimed to determine post-surgical prognoses in HCC patients with BDTT, as outcomes are currently unclear. METHODS: We compared the prognoses of 110 HCC patients without BDTT (group A) to 22 cases with BDTT (group B). The two groups were matched in age, gender, tumor etiology, size, number, portal vascular invasion, and TNM stage. Additionally, 28 HCC patients with BDTT were analyzed to identify prognostic risk factors. RESULTS: The 1-, 3-, and 5-year overall survival rates were 90.9%, 66.9%, and 55.9% for group A and 81.8%, 50.0%, and 37.5% for group B, respectively. The median survival time in groups A and B was 68.8 and 31.4 months, respectively (P=0.043). The patients for group B showed higher levels of serum total bilirubin, alanine aminotransferase and gammaglutamyl transferase, a larger hepatectomy range, and a higher rate of anatomical resection. In subgroup analyses of patients with BDTT who underwent R0 resection, TNM stage III- IV was an independent risk factor for overall survival; these patients had worse prognoses than those with TNM stage I-II after R0 resection (hazard ratio=6.056, P=0.014). Besides, univariate and multivariate analyses revealed that non-r0 resection and TNM stage III-IV were independent risk factors for both disease-free survival and overall survival of 28 HCC patients with BDTT. The median overall survival time of patients with BDTT who underwent R0 resection was longer than that of patients who did not undergo R0 resection (31.0 vs 4.0 months, P=0.007). Author Affiliations: Division of Hepatobiliary Pancreatic Surgery, The Affiliated Hospital of Qingdao University, Qingdao , China (Wang DD, Wu LQ and Wang ZS) Corresponding Author: Li-Qun Wu, MD, PhD, Division of Hepatobiliary Pancreatic Surgery, The Affiliated Hospital of Qingdao University, Qingdao , China (Tel: ; wulq5810@126.com) 2016, Hepatobiliary Pancreat Dis Int. All rights reserved. doi: /S (16) Published online September 29, CONCLUSIONS: R0 resection prolonged survival time in HCC patients with BDTT, although prognosis remains poor. For such patients, R0 resection is an important treatment that determines long-term survival. 626 Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15, (Hepatobiliary Pancreat Dis Int 2016;15: ) KEY WORDS: hepatocellular carcinoma; bile duct tumor thrombus; prognosis; recurrence; R0 resection Introduction Hepatocellular carcinoma (HCC) is the most common type of liver cancer with a high mortality rate worldwide. Long-term prognosis is poor even among HCC patients who undergo radical surgery or liver transplantation. Liver tumor can invade the portal vein and bile duct, and cause vascular tumor thrombus or bile duct tumor thrombus (BDTT). Compared to the frequency of vascular invasion, HCC with BDTT is rare, with an incidence ranging from 0.5% to 12.9% according to autopsy and pathological specimens of HCCs. [1-5] Obstructive jaundice is a special symptom caused by bile duct invasion, which often results in misdiagnosis and poorer liver function. Despite the paucity of HCC with BDTT, several reports [6-12] showed that HCC with BDTT has a higher recurrence rate and poorer prognosis after surgery. However, the post-surgical prognosis remains controversial because of the absence of a unified treatment procedure. Surgical resection is the only radical cure for these patients. Moreover, there is little information about the prognosis of patients with BDTT after R0 resection in particular. This study aimed to identify the post-surgical prognosis, including the risk factors analysis, in HCC patients with BDTT. To that end, we analyzed the clinicopathological characteristics of HCC patients who received R0 and non-r0 resection at

2 R0 resection for HCC with BDTT our center, including patients with and without BDTT. We also discuss the optimal therapeutics for this type of HCC. Methods Patients and definitions Between January 1998 and December 2012, 1084 patients underwent surgery for HCC at the Division of Hepatobiliary Pancreatic Surgery, the Affiliated Hospital of Qingdao University. All surgical specimens were identified by pathological analysis, including 882 cases of R0 resection, 105 of R1 resection, and 97 of R2 resection; BDTT was found in 28 patients, 22 of them underwent R0 resection while 6 underwent R2 resection. The clinical and pathological specimens of these patients were retrospectively reviewed. In order to reduce the influence of non-radical resection, only patients who underwent R0 resection were selected. Propensity score matching was used to minimize selection bias and confounding factors. Group A included 110 patients chosen from 860 HCC patients without BDTT who underwent R0 resection. Group A patients were matched at a 5:1 ratio with 22 patients with BDTT according to age, gender, tumor etiology, tumor size, tumor number, portal vascular invasion, and American Joint Committee on Cancer (AJCC) TNM staging; these 22 patients comprised of group B. Additionally, the features of all 28 HCC patients with BDTT were analyzed to identify prognostic risk factors. This study was approved by the Ethics Committee of our hospital, and all patients or their guardians provided informed consent. The histological grade was evaluated in accordance with the standards of the Edmondson Steiner classification. The tumor stage was evaluated by operation and pathological analysis according to the 7th Edition of the AJCC Cancer Staging Manual. BDTT was divided into two types: microscopic BDTT, which can be observed beyond the second-order branch of the biliary tree, and macroscopic BDTT, which is restricted to the secondorder branch. [13] Liver function was assessed according to the Child-Pugh classification and the model for endstage liver disease (MELD) scoring system. Child-Pugh class A and a MELD score <8 were required for inclusion in the study, although some patients had Child-Pugh class B due to obstruction of the biliary tract. The liver residual volume was also required to be more than 50% of the total liver volume according to computed tomography (CT)-assisted three-dimensional reconstruction. Hepatectomies were classified into anatomical resection, en bloc resection of liver tumor according to the Couinaud segment, non-anatomical resection, partial resection, or tumor enucleation. The surgical procedure was determined on the basis of the liver function reserve, cirrhosis, and hepatectomy range of the patients. Hepatectomy was considered R0 resection when the bile duct and specimen margins were tumor-negative on histological and pathological analysis and the disease-free survival (DFS) time exceeded 2 months. Resection was considered R1 if the margin was histologically positive, and R2 if the margin was macroscopically positive. Follow-up Follow-up surveys were scheduled every month during the first 3 months post-surgery and every 3 months thereafter. Serum alpha-fetoprotein (AFP) assessments and abdominal ultrasonography (US) or chest and abdominal CT images were obtained at each follow-up. Patients with a surgical margin of <0.5 cm, multiple tumors, or vascular invasion were treated with preventive transcatheter arterial chemoembolization (TACE) at 3-4 weeks after surgery. All patients were followed-up until death or the study end date of December Statistical analysis All statistical analyses were performed using SPSS (version 13.0, SPSS Inc., Chicago, IL, USA). The Chi-square test was used to compare categorical variables. Continuous numerical variables were presented as mean ± standard deviation (SD) or median with interquartile range. The Kaplan-Meier method was used for survival analyses. Multivariate analysis was performed using the Cox multivariate regression model. A P value <0.05 was considered statistically significant. Results Patient profiles and pathological features Baseline information on the 28 HCC patients with BDTT is shown in Table 1. Among the 28 patients, 16 underwent choledochotomy and exploration, 11 underwent choledochotomy and embolectomy, and 1 underwent bile duct resection. The pathological features of the 28 HCC patients with BDTT are shown in Table 2. According to TNM staging, there were 15 stage I, 3 stage II, 9 stage III, and 1 stage IV. Sixteen cases were microscopic BDTT and 12 were macroscopic BDTT. Additionally, 5 cases had BDTT combined with bile duct wall invasion, which were proven by postoperative pathological analysis of surgical specimens, including 1 invasion to extrahepatic bile duct wall. Prognosis of HCC patients with BDTT after R0 resection The characteristics of groups A and B were compared Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,

3 Hepatobiliary & Pancreatic Diseases International and analyzed in Table 3. There were no significant intergroup differences in age, gender, cirrhosis, portal hypertension, preoperative TACE, surgical margin, amount of bleeding, blood transfusion, serum AFP level, tumor size, tumor number, portal vascular invasion, TNM stage, or differentiation status between the two groups. Levels of serum total bilirubin, serum alanine aminotransferase (ALT), and serum gamma-glutamyl transferase (GGT) were higher in patients in group B. These patients also showed a larger hepatectomy range and higher rate of anatomical resection. Fourteen patients from group B had recurrence; 7 Table 1. Preoperative demographic and clinical features of patients with HCC and BDTT (n, %) Values Age (yr) 54.8±11.3 Gender (male/female) 22/6 Primary background liver status HBsAg 23 (82.1) Anti-HCV 0 Non-viral liver disease 5 (17.9) Child-Pugh classification Class A 16 (57.1) Class B 12 (42.9) Total bilirubin (µmol/l) 34.6 ( ) ALT level (U/L) 69 (14-481) AFP (ng/ml) > (39.3) (60.7) Preoperative HCC treatment None 25 (89.3) TACE 3 (10.7) Hepatectomy range 1 4 (14.3) 2 9 (32.1) 3 15 (53.6) Surgical margin (mm) <5 18 (64.3) 5 10 (35.7) Hepatic vascular occlusion Yes 18 (64.3) No 10 (35.7) Bleeding (ml) < (71.4) (28.6) Blood transfusion Yes 15 (53.6) No 13 (46.4) Anatomical resection Yes 14 (50.0) No 14 (50.0) R0 resection Yes 22 (78.6) No 6 (21.4) Table 2. Pathological findings of patients with HCC and BDTT (n, %) Values Background liver status Normal or chronic hepatitis 4 (14.3) Cirrhosis 24 (85.7) Tumor number 1 20 (71.4) 2 3 (10.7) 3 5 (17.9) Maximal tumor size (cm) 6.8±3.5 Portal vascular invasion Yes 18 (64.3) No 10 (35.7) Histologic differentiation Well 1 (3.6) Moderate 15 (53.5) Poor 12 (42.9) TNM staging I 15 (53.6) II 3 (10.7) III 9 (32.1) IV 1 (3.6) Bile duct thrombus type Microscopic 16 (57.1) Macroscopic 12 (42.9) Bile duct wall invasion Yes 5 (17.9) No 23 (82.1) Table 3. Comparison of clinicopathological features between groups A and B Group A (n=110) Group B (n=22) P value Age ( 65/>65 yr) 86/24 20/ Gender (male/female) 90/20 18/ Child-Pugh classification (A/B) 105/5 21/ Cirrhosis 88/22 18/ Portal hypertension 14/96 3/ Total bilirubin ( 17.1/>17.1 µmol/l) 74/36 6/ ALT ( 50/>50 U/L) 80/30 10/ GGT ( 64/>64 U/L) 72/38 2/ Preoperative TACE 20/90 2/ Hepatectomy range >3 23/87 11/ Anatomical resection 37/73 14/ Surgical margin (<0.5/ 0.5 cm) 71/39 17/ Blood transfusion 36/74 12/ Bleeding 1000 ml 13/97 5/ AFP ( 400/>400 ng/ml) 74/36 12/ Multiple tumor 18/92 4/ Tumor size ( 5.0/>5.0 cm) 51/59 8/ Portal vascular invasion 21/89 5/ TNM stage (I-II/ III-IV) 78/32 16/ Differentiation (well/poor) 75/35 13/ Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,2016

4 R0 resection for HCC with BDTT were within 1 year. The median survival time was significantly different between the two groups after R0 resection (68.8 vs 31.4 months for group A vs B, respectively, P=0.043) (Fig. 1). The 1-, 3-, and 5-year overall survival (OS) rates were 90.9%, 66.9%, and 55.9% for group A and 81.8%, 50.0%, and 37.5% for group B, respectively. Subgroup analysis of R0 resection In subgroup analysis of group B, univariate and multivariate analyses revealed that TNM stage III-IV was an independent risk factor for OS of R0 resection patients (Table 4). However, these analyses did not reveal any risk factors for DFS in the R0 resection group. The median survival time of the TNM stage I-II group was longer than that of the TNM stage III-IV group after R0 resection (31.4 vs 10.0 months, P=0.031). The 22 patients with BDTT had a 5-year OS rate of 37.5%; of those, the 16 patients with TNM stage I-II had a 5-year OS rate of 46.9%. Survival analysis of 28 HCC patients with BDTT The 1-, 3-, and 5-year OS rates were 90.9%, 39.3%, and 29.9% for 28 HCC patients with BDTT. We performed univariate Kaplan-Meier analysis to identify the factors influencing the prognosis of 28 HCC patients with BDTT (Table 5). Cox multivariate regression analysis was also performed on factors with P<0.10 on univariate analysis for both DFS and OS (Table 6). The analyses Fig. 1. Comparison of survival time curves between groups A and B after R0 resection (P=0.043). Table 4. Univariate and multivariate analysis of risk factors for overall survival of R0 resection patients Univariate analysis Median (mon) Age ( 65/>65 yr) 61.0/ Gender (male/female) 31.4/ Cirrhosis 31.0/ Portal hypertension 31.0/ BDTT type (micro/macro) 21.0/ TNM stage (I-II/ III-IV) Tumor size ( 5.0/>5.0 cm) Surgical margin (<0.5/ 0.5 cm) Blood transfusion Bleeding 1000 ml Hepatectomy range >3 Anatomical resection Multivariate analysis P value HR 95% CI P value 31.4/ / / / / / / HR: hazard risk; CI: confidence interval. Table 5. Univariate analysis of risk factors for patient survival and recurrence DFS Median (mon) OS P value Median (mon) P value Age ( 65/>65 yr) 25.0/ / Gender (male/female) 15.0/ / Cirrhosis 9.9/ / Portal hypertension 7.0/ / Differentiation (well/poor) 25.0/ / BDTT type (micro/macro) 8.0/ / AFP ( 400/>400 ng/ml) 25.4/ / TNM stage (I-II/III-IV) 25.4/ / Multiple tumor 9.9/ / Portal vascular invasion 13.0/ / Tumor size ( 5.0/>5.0 cm) 37.8/ / Surgical margin (<0.5/ 0.5 cm) 7.0/ / Blood transfusion 15.0/ / Bleeding 1000 ml 13.0/ / Hepatectomy range >3 7.0/ / Anatomical resection 50.2/ / R0 resection 25.0/ / DFS: disease-free survival; OS: overall survival. Table 6. Multivariate analysis of risk factors for patient survival and recurrence HR 95% CI P value Disease-free survival TNM stage (III-IV) Non-R0 resection Overall survival TNM stage (III-IV) Non-R0 resection HR: hazard risk; CI: confidence interval. Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,

5 Hepatobiliary & Pancreatic Diseases International Fig. 2. Comparison of survival time of HCC patients with BDTT treated with invasive procedures and those treated conservatively after recurrence (P=0.002). revealed that non-r0 resection and TNM stage III-IV were independent risk factors of DFS and OS for HCC patients with BDTT. Additionally, the median overall survival time of patients with BDTT who underwent R0 resection was longer than that of patients who underwent non-r0 resection (31.0 vs 4.0 months, P=0.007). Treatment for tumor recurrence Of the 28 HCC patients with BDTT, 20 (71.4%) experienced intrahepatic recurrence after surgery, including 4 (20.0%) BDTT recurrences (within 1 year) and 2 (10.0%) who also had extrahepatic recurrence. Among patients with recurrence, 13 (65.0%) underwent active treatment, including TACE (n=8), percutaneous transhepatic cholangiography (n=2), endoscopic biliary metal stent drainage (n=2), and radiofrequency ablation (n=1). The remaining 7 patients with recurrence underwent conservative treatment alone because of poor liver function or for other reasons. Notably, 4 patients with BDTT recurrence exhibited macroscopic BDTT, and this difference was significant (likelihood ratio test, odds ratio=6.556, P=0.010). The DFS of each of these 4 patients, who underwent choledochotomy, was 3.0, 6.2, 7.0, and 9.9 months, respectively. The median survival time after recurrence of patients with active treatment was 9.2 months, which was longer than that of patients with conservative treatment (3.0 months; P=0.002) (Fig. 2). Discussion The incidence of HCC with BDTT in this study was 2.58% (28/1084), which is within the 0.5% to 12.9% range reported previously. [1-5] However, studies have not conclusively shown whether HCC patients with BDTT have a poorer long-term prognosis after surgery. Shiomi et al [1] reported that there are no significant differences in prognosis between patients with and without BDTT after appropriate preoperative management. Likewise, Satoh et al [9] showed that the prognosis of HCC patients with BDTT is similar to that of patients without BDTT. However, Ikenaga et al [6] reported that HCC with BDTT has a poorer prognosis than HCC without BDTT after resection. Yeh et al [8] found that HCC patients with BDTT have significantly worse OS than those without BDTT, especially those with concomitant vascular invasion. Noda et al [12] also found that HCC patients with BDTT have lower OS rates compared to those without BDTT, and that patients with BDTT have a lower curative resection rate. Previous studies [6-12] have only selected patients who underwent hepatectomy, but did not distinguish between curative or non-curative resection. Therefore, we only chose patients who underwent R0 resection to avoid the influence of non-curative resection on our analysis. The median survival time for patients with BDTT was still significantly shorter than that for patients without BDTT after R0 resection which implied that patients with BDTT have worse prognoses even after R0 resection. Peng et al [14] reported that the poor prognosis of HCC patients with BDTT is attributed to low resectability and poor liver function. In our study, tumor progression and portal vascular invasion were matched between patients with or without BDTT. The preoperative levels of serum total bilirubin, ALT, and GGT were higher in patients with BDTT; this indicates poor liver function owing to bile duct obstruction. Obstructive jaundice occurred in up to 12 patients with BDTT diagnosed by preoperative laboratory examination. Meanwhile, the larger hepatectomy range and higher rate of anatomical resection were due to tumor invasion of the bile duct, which necessitates anatomical resection of this part of the Glisson s sheath in order to reduce the residual tumor. Additionally, Shao et al [7] and Zeng et al [3] reported that 25.9% and 52.3% of patients with BDTT, respectively, showed BDTT recurrence after surgery, and speculated that bile duct invasion might be another route of intrahepatic metastasis that causes BDTT recurrence. In the current study, 4 patients experienced early local BDTT recurrence (<1 year). Although the tumor thrombus easily separated from the bile duct during surgery, we speculated that BDTT recurrence in these patients was due to the duct wall invasion of the tumor thrombus. Compared to patients with microscopic BDTT, those with macroscopic BDTT exhibit a higher possibility of bile duct wall invasion. Therefore, bile duct invasion may be a route of metastasis that leads to early local recurrence and poor prognosis in HCC patients with BDTT. These findings indicate that HCC with BDTT can result 630 Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,2016

6 R0 resection for HCC with BDTT in worse prognosis for poor liver function and potential metastasis through the bile duct even after R0 resection, and that such cases should receive more attention. Notably, BDTT is considered in the Japanese TNM staging system; however, intrahepatic bile duct invasion is no longer included in the chapter on HCC staging in the 7th Edition of the AJCC Cancer Staging Manual. Minagawa et al [15] showed that the prognostic stratification ability of the Japanese TNM staging system is more effective than that of the AJCC TNM staging system. According to the 7th Edition of the AJCC Cancer Staging Manual, TNM staging of HCC considers most tumor characteristics, including tumor size, number, vascular invasion, lymph node involvement, and distant metastasis; hence, TNM stage is greatly relevant to tumor prognosis. Notably, we found that TNM stage III-IV was the only independent risk factor for OS in this subgroup. The 22 patients with BDTT who underwent R0 resection had a 5-year OS rate of 37.5%; of those, the 16 patients with TNM stage I-II had a 5-year OS rate of 46.9%. Moon et al [16] and Zeng et al [3] found that surgical curability is an independent risk factor for both OS and DFS in HCC patients with BDTT. In our study, non-r0 resection and TNM stage III-IV were independent risk factors for tumor recurrence and prognosis in 28 HCC patients with BDTT. Although patients with BDTT have a poor prognosis and low curative resection rate, R0 resection is still an important treatment for improving long-term prognosis, and can significantly reduce recurrence in patients with BDTT. Meanwhile, patients with recurrence who obtained active treatment had longer OS than patients who received conservative management. Peng et al [14] also reported that repeated surgery resulted in longer survival for HCC patients with BDTT. Therefore, in addition to tumor progression, the prognosis of HCC patients with BDTT can be estimated by R0 resection, which yields a satisfactory outcome. Patients with TNM stage III-IV should undergo frequent follow-up checks after R0 resection. Although en bloc resection of the primary tumor with choledochotomy and embolectomy is considered an effective technique for curative resection of HCC with BDTT, this approach remains controversial. Shiomi et al [1] and Satoh et al [9] reported that hepatectomy with choledochotomy and embolectomy was as effective as bile duct resection, and that there were no differences in OS between patients who underwent choledochotomy and those who underwent bile duct resection. However, [17, 18] other studies reported that tumor embolectomy through a choledochotomy might cause intraoperative iatrogenic implantation and thereby cause tumor recurrence. In this study, 16 patients underwent choledochotomy and exploration, 11 patients underwent choledochotomy and embolectomy, and only 1 patient underwent bile duct resection for macroscopic bile duct wall invasion; this yielded a favorable 5-year OS rate of 29.9%. It is generally accepted that most of the BDTT can be easily separated from the bile duct because BDTT rarely invades the bile duct walls and usually only loosely attaches to the epithelium. [14] In our series, BDTT was easily separated in most cases, and only 1 patient underwent bile duct resection because a tumor thrombus was tightly lodged in the bile duct. However, 2 of the 27 patients who underwent choledochotomy and embolectomy experienced bile duct bleeding after embolectomy; both belonged to the R2 resection group, and the DFS was 1 and 2 months, respectively. Bile duct bleeding might have been due to BDTT invasion of the bile duct walls, or to the entry of the tumor into the bile duct after embolectomy, both of which might have caused early recurrence because of residual tumor in the bile duct wall. Furthermore, we speculate that the BDTT recurrence in 4 patients was due to duct wall invasion by the tumor thrombus. Zeng et al [3] reported that 66.7% of macroscopic BDTT-proven patients exhibited bile duct wall invasion according to postoperative or intraoperative pathological findings. In our study, 5 patients had proven bile duct thrombi combined with bile duct wall invasion on postoperative pathology. Therefore, bile duct resection might be necessary to reduce the recurrence rates in patients with HCC who have macroscopic BDTT and who are suspected of having bile duct wall invasion. Unfortunately, the diagnosis of bile duct wall invasion is difficult, whether by macroscopic testing during surgery or imaging prior to surgery. Therefore, we deduce that hepatectomy with choledochotomy for patients with microscopic BDTT ought to be recommended; bile duct resection is necessary to achieve R0 resection and to reduce the possibility of recurrence for patients with macroscopic BDTT suspected to have, or diagnosed with, bile duct wall invasion. This is especially true for those with bile duct bleeding after embolectomy or those with invasion confirmed by pathology. However, further direct evidence of this approach will be necessary before drawing a definitive conclusion. There are several limitations of this study. First, the number of patients is small. Second, the fact that this is a single-center study may have introduced bias. Additional clinical retrospective and prospective studies are required to confirm these results. Furthermore, the pathogenesis of BDTT in HCC patients requires further research. In conclusion, R0 resection is an important treatment for the long-term survival of HCC patients with BDTT, although the prognosis remains poor. However, Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,

7 Hepatobiliary & Pancreatic Diseases International further studies with a greater number of subjects are necessary to verify these conclusions. Contributors: WLQ proposed the study. WDD and WZS performed the research and wrote the first draft. WDD collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. WLQ is the guarantor. Funding: None. Ethical approval: This study was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University. Competing interest: The authors do not choose to declare any conflict of interest related directly or indirectly to the subject of this article. References 1 Shiomi M, Kamiya J, Nagino M, Uesaka K, Sano T, Hayakawa N, et al. Hepatocellular carcinoma with biliary tumor thrombi: aggressive operative approach after appropriate preoperative management. Surgery 2001;129: Ueda M, Takeuchi T, Takayasu T, Takahashi K, Okamoto S, Tanaka A, et al. Classification and surgical treatment of hepatocellular carcinoma (HCC) with bile duct thrombi. Hepatogastroenterology 1994;41: Zeng H, Xu LB, Wen JM, Zhang R, Zhu MS, Shi XD, et al. Hepatocellular carcinoma with bile duct tumor thrombus: a clinicopathological analysis of factors predictive of recurrence and outcome after surgery. Medicine (Baltimore) 2015;94:e Huang JF, Wang LY, Lin ZY, Chen SC, Hsieh MY, Chuang WL, et al. Incidence and clinical outcome of icteric type hepatocellular carcinoma. J Gastroenterol Hepatol 2002;17: Qin LX, Ma ZC, Wu ZQ, Fan J, Zhou XD, Sun HC, et al. Diagnosis and surgical treatments of hepatocellular carcinoma with tumor thrombosis in bile duct: experience of 34 patients. World J Gastroenterol 2004;10: Ikenaga N, Chijiiwa K, Otani K, Ohuchida J, Uchiyama S, Kondo K. Clinicopathologic characteristics of hepatocellular carcinoma with bile duct invasion. J Gastrointest Surg 2009;13: Shao W, Sui C, Liu Z, Yang J, Zhou Y. Surgical outcome of hepatocellular carcinoma patients with biliary tumor thrombi. World J Surg Oncol 2011;9:2. 8 Yeh CN, Jan YY, Lee WC, Chen MF. Hepatic resection for hepatocellular carcinoma with obstructive jaundice due to biliary tumor thrombi. World J Surg 2004;28: Satoh S, Ikai I, Honda G, Okabe H, Takeyama O, Yamamoto Y, et al. Clinicopathologic evaluation of hepatocellular carcinoma with bile duct thrombi. Surgery 2000;128: Yu XH, Xu LB, Liu C, Zhang R, Wang J. Clinicopathological characteristics of 20 cases of hepatocellular carcinoma with bile duct tumor thrombi. Dig Dis Sci 2011;56: Kim JM, Kwon CH, Joh JW, Sinn DH, Park JB, Lee JH, et al. Incidental microscopic bile duct tumor thrombi in hepatocellular carcinoma after curative hepatectomy: a matched study. Medicine (Baltimore) 2015;94:e Noda T, Nagano H, Tomimaru Y, Murakami M, Wada H, Kobayashi S, et al. Prognosis of hepatocellular carcinoma with biliary tumor thrombi after liver surgery. Surgery 2011;149: Esaki M, Shimada K, Sano T, Sakamoto Y, Kosuge T, Ojima H. Surgical results for hepatocellular carcinoma with bile duct invasion: a clinicopathologic comparison between macroscopic and microscopic tumor thrombus. J Surg Oncol 2005;90: Peng SY, Wang JW, Liu YB, Cai XJ, Deng GL, Xu B, et al. Surgical intervention for obstructive jaundice due to biliary tumor thrombus in hepatocellular carcinoma. World J Surg 2004;28: Minagawa M, Ikai I, Matsuyama Y, Yamaoka Y, Makuuchi M. Staging of hepatocellular carcinoma: assessment of the Japanese TNM and AJCC/UICC TNM systems in a cohort of patients in Japan. Ann Surg 2007;245: Moon DB, Hwang S, Wang HJ, Yun SS, Kim KS, Lee YJ, et al. Surgical outcomes of hepatocellular carcinoma with bile duct tumor thrombus: a Korean multicenter study. World J Surg 2013;37: Mok KT, Chang HT, Liu SI, Jou NW, Tsai CC, Wang BW. Surgical treatment of hepatocellular carcinoma with biliary tumor thrombi. Int Surg 1996;81: Wang HJ, Kim JH, Kim JH, Kim WH, Kim MW. Hepatocellular carcinoma with tumor thrombi in the bile duct. Hepatogastroenterology 1999;46: Received October 13, 2015 Accepted after revision May 20, Hepatobiliary Pancreat Dis Int,Vol 15,No 6 December 15,2016