Prediction model of hepatocellular carcinoma risk in Asian patients with chronic hepatitis B treated with entecavir

Abstract

Background: Until now, no risk score could predict hepatocellular carcinoma (HCC) in nucleos(t)ide analog (NA)-treated Asian patients.

Methods: We enrolled 1325 NA-naïve chronic hepatitis B patients with entecavir monotherapy for >12 months, with 883 and 442 patients randomly assigned to the development and validation groups, respectively, in the risk model.

Results: The cumulative probabilities of HCC were 2.4%, 4.1%, and 9.9% after 2, 3, and 5 years of treatment, respectively. In the development group, age, platelet counts, and alpha-fetoprotein levels after 12 months of treatment were the independent predictors of HCC. We converted the Cox proportional hazards regression coefficients for these predictors into risk scores and developed the APA-B model, with the total risk scores ranging from 0 to 15. The risk scores accurately categorized patients with low (0–5), medium (6–9), and high (10–15) risks in the validation group (P <0.001). The areas under the receiver operating characteristic curve for predicting HCC risk after 2, 3, and 5 years were 0.877, 0.842, and 0.827, respectively, in the development group and 0.939, 0.892, and 0.862, respectively, in the validation group.

Conclusion: The proposed HCC risk prediction model exhibited excellent predictive accuracy in NA-naïve Asian patients receiving entecavir therapy.