Abstract
OBJECTIVE: Hepatocellular carcinoma (HCC) can still occur in patients with chronic hepatitis C after achieving a sustained virologic response (SVR) with direct-acting antiviral (DAA) therapy. Therefore, we aimed to identify and validate predictors and HCC risk models using longitudinal data. METHOD: This retrospective cohort study included patients who achieved SVR after DAA therapy. Model performance was assessed using Harrell's C-index and the area under the time-dependent receiver operating characteristic curve (AUROC). RESULTS: After excluding patients who developed HCC or died within 1 year after achieving SVR, a total of 2778 patients were finally included in this study. 154 (5.5%) patients developed HCC at a median of 3.3 years. LSM value showed significant improvement after antiviral therapy (all p < 0.05). In the multivariable analysis, LSM at baseline and 1 year after SVR were significantly associated with the occurrence of HCC (all p < 0.05). LSM at 1 year after SVR could better predict HCC risk than baseline LSM (C-index: 0.760, 0.788 at baseline, 1 year after SVR, respectively). Low-risk patients (LSM < 10 kPa) using LSM values at 1 year after SVR (76.4%) showed an incidence risk of 0.482 incidence rate/100 patient-years. In patients with advanced chronic liver disease, LSM at 1 year after SVR demonstrated superior predictive performance for HCC development compared to baseline LSM and other prediction models assessed at various time points (C-index: 0.744, 95% CI: 0.660-0.829; 3-year follow-up AUROC: 0.794, 95% CI: 0.754-0.831). CONCLUSIONS: LSM after achieving SVR effectively predicts HCC risk, especially in patients with advanced chronic liver disease.