Verification of hepatitis B–related hepatocellular carcinoma predictive models to evaluate the risk of HCC in patients with liver cirrhosis under antiviral treatment

Background Hepatocellular carcinoma (HCC) is the fifth most common cancer in the world. The development of chronic hepatitis B (CHB)-related HCC can be attributed to continuous hepatitis B virus (HBV) infection. It is crucial to identify and monitor patients with CHB at high risk of HCC occurrence so that HCC can be detected early and the patients are able to receive effective treatment promptly to increase the survival rate and improve prognosis. Aim This study aimed to verify hepatitis B–related hepatocellular carcinoma predictive models to evaluate the risk of HCC in patients with liver cirrhosis under antiviral treatment. Methods Patients with HBV-related compensated cirrhosis were treated with lamivudine and adefovir randomly, and then with a combination of two drugs at different time points based on the virologic response. Patients with HCC occurrence during follow-up were categorized as HCC group, whereas others as control group. They were further divided into 2-year HCC, 2-year control, 5-year HCC, and 5-year control groups according to the observation time. The operating curves of the patients were used to verify models before and after antiviral treatment. Results Using the baseline as a parameter, the area under the curve (AUC) of risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B) after 2 and 5 years was 0.863 and 0.797, respectively. The AUC after 2 and 5 years was 0.839 and 0.747, respectively, for guide with age, gender, HBV DNA, core promoter mutations and cirrhosis (GAG-HCC) and 0.741 and 0.748, respectively, for Taiwanese HBV cohort (TW1). Using 48 weeks as the parameter, it has an optimal critical value of 8 points. The AUC of REACH-B after 2 and 5 years was 0.738 and 0.721, respectively. Conclusion REACH-B can predict the risk of HCC occurrence in patients with compensated liver cirrhosis before and after antiviral treatment. GAG-HCC and TW1 could predict the risk before antiviral treatment.