Predictors of Recurrence in Hepatitis C Virus Related Hepatocellular Carcinoma after Hepatic Resection : A Retrospective Cohort Study

Objective: Egypt is one of the hot spots in the international map of Hepatocellular carcinoma (HCC), which is where hepatitis C virus (HCV) infection is the major risk factor in development of HCC (80%). Due to low organ donation rates and lack of deceased liver transplantation, hepatic resection is the main line of treatment for HCC patients with sufficient liver reserve. We introduce our experience with patients who had HCV related HCC who underwent hepatic resection to determine various predictors of tumour recurrence in this group. This is the first study to come from a country where chronic HCV hepatitis is endemic. Materials and Methods: This is a retrospective cohort study of 208 cases of HCC in hepatitis C virus positive patients with cirrhotic livers who underwent first-time liver resection, in Gastroenterology Surgical Centre, Mansoura University, Egypt during the period from January 2002 to December 2011. Shapiro-Wilk test was used to assess normality of data. Predictors of HCC recurrence were assessed by bivariate correlation tests, univariate analysis using the chi-square and t-test and binary logistic regression analysis. A P value <0.05 was considered statistically significant. Results: Tumour recurrence occurred in 88 patients (42.3%). Most of the recurrences occurred within the first year 55 patients (62.5%). The most common site for recurrence was the liver (n=68, 77.3%). Based on the univariate analysis; significant variables predicting tumor recurrence were alpha feto-protein (AFP), blood transfusion, multifocality, cut margin, microvascular invasion, lack of capsule, tumour grade and stage. Based on multivariate analysis, the main variables predicting tumor recurrence were blood transfusion, cut margin, tumour capsule and microvascular invasion. Conclusion: Although the predictors of recurrence are the same for both HBV and HCV related HCC, the rate and aggressiveness of recurrence are higher in HCV related HCC.


Introduction
Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide, and the third most common aetiology of cancer-associated mortality [1].The incidence of HCC in the Europe is 8.29/100 000, while in Asia and Africa it is 120/100 000 owing to high rates of viral hepatitis (B and C) [2].Nowadays, Egypt is considered as one of the hot spots in the international map of HCC, that is where hepatitis C virus (HCV) infection (genotype 4) is the major risk factor in development of HCC (approximately 80% of the cases) [3].
Hepatic resection and liver transplantation are the main lines of curative treatment for HCC [1].In countries (such as Egypt) with a high incidence of HCC owing to endemic viral hepatitis, low organ donation rates and lack of deceased liver transplantation, primary hepatic resection is the first line of treatment for HCC patients with sufficient liver reserve [4].
In this study, we introduce our experience after long term follow-up of patients who underwent hepatic resection for HCV related HCC in cirrhotic liver.

Materials and Methods
This is a retrospective cohort study of 208 cases of HCC in hepatitis C virus positive patients with cirrhotic livers who underwent first-time liver resection, in the Gastroenterology Surgical Centre, Mansoura University, Egypt during the period from January 2002 to December 2011.Patient data was retrieved from internal web-based registry system supplemented by paper records included in the medical archive.
Patient selection criteria were Child-Pugh class A or B, performance status 0-2, and positive markers for HCV.Other patients with positive serum markers for hepatitis B virus (HBV) or negative for both HCV and HBV were excluded.
The extent of the hepatic resection was based on the International Hepato-Pancreato-Biliary Association classification.Major hepatectomy was defined by resection of three or more hepatic segments according to Couinaud's classification, and segmentectomy was defined by resection of less than three hepatic segments [5].
All patients were followed up every month in the first 3 months and every 3-6 months thereafter.The visit consisted of physical examination, liver function tests, serum alpha fetoprotein (AFP) level, abdominal ultrasound (US), and triphasic computed tomography (CT) when recurrence is suspected.
Shapiro-Wilk test was used to assess normality of data.Numerical data is presented as means and standard deviations or as medians with ranges.A P value <0.05 was considered statistically significant.Bivariate correlation tests were done to estimate the correlation between different variables and recurrence.Univariate analysis then was done for all the correlated factors (independent variables) using the chisquare and t-test.The variables that were significant by univariate analysis were subsequently analysed using the binary logistic regression analysis.Statistical analysis was done with the help of IBM SPSS v. 20.

Results
The clinical characteristics of the patients and operative data is shown in (Table 1).Postoperative pathological data is shown in (Table 2).The mean hospital stay was 9.04 days Based on the univariate analysis; significant variables predicting tumour recurrence were AFP, blood transfusion, multifocality, cut margin, microvascular invasion, lack of capsule, tumour grade and stage (Table 4).Based on multivariate analysis, the main variables predicting tumour recurrence were blood transfusion, cut margin, tumour capsule and microvascular invasion (Table 5).

Discussion
With advancement of surgical techniques and postoperative care, hepatic resection in cirrhotic patients with HCC became a safe procedure and the gold standard treatment for HCC patients.The long term outcome of hepatic resection remains unsatisfactory due to tumour recurrence [6].The incidence is extremely high, with 40-100% 5-year cumulative recurrence rates and 80-95% of recurrences occur in the remnant liver [7,8].In our study tumour recurrence occurred in 88 patients (42.3%) and most recurrences occurred in the liver (77.3%).The cumulative recurrence rates of HCV related HCCs are higher than HBV related HCCs [9][10][11].This could be explained by the high viral replication and hepatic inflammation in HCV related HCCs.Also, HCV related HCCs have a higher incidence of tumour multicentricity [9,10].Recently, it is found that HCV related HCCs are associated with expression of Twist (a regulator of mesenchymal cells transition), which plays an important role in invasiveness and metastasis [11].The prognosis of patients with a single tumour nodule is better than those with multiple nodules [12].Tumour multifocality is due to either intrahepatic metastasis or multicentric occurrence.Both of them could cause recurrence in the remaining liver [7,8].In our study, tumour multifocality was a significant predictor for tumour recurrence.This is corroborated by other studies for both HCV related HCCs [3,9] and HBV related HCCs [6,10,13].
A wide resection margin to ensure R0 resection is a general rule in oncological surgery.Despite multiple studies evaluating the importance of wide resection margin for HCC, its importance remained a matter of debate.In Egyptians, the associated liver cirrhosis in HCC patients, due to chronic HCV, limits the extent of hepatic resection.In those patients, if a major hepatic resection is performed, they may die from liver cell failure, as occurred in 10 of our cases.In our study, infiltrated safety margin was a significant variable predicting tumour recurrence.In comparison to other studies evaluating the recurrence in HBV related HCCs; the role of the resection margin is also controversial.Some studies supported a wide safety margin (more than 1 cm) and found it a significant predictor of tumour recurrence [13,14].Other studies found no significant association between safety margin and tumour recurrence [15][16][17].
Alpha feto-protein (AFP) has been suggested as a strong predictor of survival and tumour recurrence after hepatic resection [18].This arises from the association between high AFP levels and tumour size, multifocality and microvascular invasion, which are all recognized predictors of HCC recurrence [19].HCCs associated with high AFP level had a higher cell proliferative activity and more aggressive behaviour [20].
In our study, AFP was a significant predictor of HCC recurrence; similar to other studies evaluating HCV related HCCs [3,21] and hepatitis B virus (HBV) related HCCs [13,22].Hepatocellular carcinoma (HCC) is characterized by its high affinity for vascular invasion (microvascular or microvascular invasion), which indicates aggressive biological manner of the tumour and is currently one of the most grave predictors of HCC recurrence [23].The presence of vascular invasion is a reported risk factor for HCC recurrence after hepatic resection for both HCV related HCCs [3,24] and HBV related HCCs [10,13,19].This was similar to findings in our study.
Hepatic resection for cirrhotics is associated with a high incidence of blood transfusions.Numerous studies reported that blood transfusion causes nonspecific immunosuppression, and affects postoperative complications and prognosis of HCC [25,26].It is reported that blood transfusion is associated with increased incidence of HCC recurrence especially in early stages (I or II) and absence of vascular invasion [27].In our study, perioperative blood transfusion was a significant predictor of tumour recurrence.Several studies reported that perioperative blood transfusions are significantly associated with increased the incidence of tumour recurrence irrespective to its viral aetiology [13,[27][28][29].
The modified tumor-lymph node-metastasis system (pTNM) includes tumour size, number, and vascular invasion in its tumour (T) classification.Therefore, it should be a significant predictor to HCC recurrence.However it has been widely evaluated and showed low prognostic significance regarding HCC recurrence after hepatic resection for both HCV and HBV related HCCs [10,13,16,30].Few studies had shown that pTNM staging provides a significant predictor for recurrence in HCV related HCCs [3] and HBV related HCCs [9,11,31].This significant correlation was similar to the findings in our study.
The effects of tumour encapsulation and histologic differentiation of HCC on recurrence risk are less convincing.The prognostic significance of tumour encapsulation on recurrence risk had been debated for both HCV and HBV related HCCs.Absence of tumour capsule has been associated with a higher incidence of recurrence in some studies [3,32], although not in other studies [9,10].Both viewpoints have a theoretical foundation.Encapsulated tumours displace, rather than invade, the surrounding normal parenchyma and vasculature which comprise a better prognosis [3].Conversely, the presence of tumour capsule is a predictor of portal venous invasion attributed to tumour invasion of blood vessels in the capsule [33].In our study, lack of tumour capsule was a significant predictor of tumour recurrence.
Also, the prognostic significance of tumour differentiation on recurrence risk has also been debated for both HCV and HBV related HCCs.Some studies found that the tumour grade was a significant predictor for tumour recurrence [19,34].This was similar to findings in our study.However, tumour grade was not a significant predictor on the recurrence risk in other studies [9,10,13].
In our experience from a tertiary high volume centre for hepatic surgery, significant variables predicting tumour recurrence were AFP, blood transfusion, multifocality, cut margin, microvascular invasion, lack of capsule, tumour grade and stage.In Egypt with high incidence of HCV related HCCs, although the predictors of recurrence are the same for both HBV and HCV related HCC, the rate and aggressiveness of recurrence are higher in HCV related HCC.

Table 3 )
. Various clinical, laboratory, operative and pathological variables were analysed to determine its relation to tumour recurrence.Clinical variables include age, sex, symptoms and Child-Pugh classification.Laboratory variables include preoperative serum albumin, bilirubin, aspartate transaminase (AST), alanine transaminase (ALT) and AFP.Operative variables include type of resection, perioperative blood transfusion, use of Pringle's manoeuvre and operative time.Pathological variables include site of tumour, size of tumour, multifocality, portal vein invasion, cut margin infiltration, positive lymph nodes, lack of capsule, tumour grade and stage.

Table 4 . Univariate analysis for predictors of tumour recurrence: AFP
Predictors of Recurrence in HCV Cirrhosis Related HCC AFP: Alpha feto-proteinWahab et al.