FIVE-YEAR ON-TREATMENT PARAMETERS-BASED MODEL PREDICTS SUBSEQUENT HEPATOCELLULAR CARCINOMA IN ENTECAVIR/TENOFOVIR-TREATED PATIENTS

Abstract

Background: Considering the lower risk of hepatocellular carcinoma (HCC) in chronic hepatitis B (CHB) patients receiving long-term potent antiviral therapy, models predicting HCC after 5 years of therapy are needed.

Methods: We conducted a multicenter retrospective cohort study to construct and validate a model predicting HCC after 5 years of entecavir (ETV) or tenofovir (TFV) therapy for CHB. The endpoint was HCC after 5 years of ETV/TFV therapy. Risk factors for HCC were identified using a random forest model and multivariable Cox regression in the train set. The final prediction model was validated using the test set.

Results: Among 7,542 patients, 345 (4.6%) developed HCC after 5 years of ETV/TFV therapy. Clinical characteristics at baseline and at 5 years of treatment were comparable between the train and test sets. HCC risk after 5 years of ETV/TFV therapy was increased by 4.1-fold in patients with liver cirrhosis than in those without cirrhosis at baseline. Furthermore, Platelet counts and Prothrombin time at 5 years, Age, and Sex were associated with increased risk of HCC and were incorporated to construct a prediction model, PPACS. PPACS had the best performance when compared to other prediction models in terms of a time-dependent area under the curve of 0.78 (95% confidence interval, 0.72–0.83) at 10-year of ETV/TFV therapy, a Brier score of 0.037, and an integrated Brier score of 0.013 in the test set.

Conclusion: The established PPACS model enables easy assessment of HCC risk after 5 years of ETV/TFV therapy (https://ppacs.shinyapps.io/shiny_app_up/).