Prognostic model for predicting recurrence-free survival in HBV-related hepatocellular carcinoma patients after...

Hepatocellular carcinoma (HCC) is a lethal malignancy, with hepatitis B virus (HBV) infection as its leading cause in China. Transarterial chemoembolization (TACE) combined with radiofrequency ablation (RFA) treatment was gradually applied in clinic, but the lack of targeted prognostic tools for this cohort remains a critical issue. This multicenter study aimed to develop a prognostic model for recurrence-free survival (RFS) in HBV-related HCC patients after the combined treatment.

A total of 604 patients from two hospitals were enrolled; 502 (Beijing You'an Hospital) formed the training/internal validation cohort (7:3 split), and 102 (Beijing Ditan Hospital) served as the external validation cohort. Baseline clinical, tumor, and laboratory data were collected. LASSO regression and random survival forest were used for variable screening, followed by multivariate Cox regression to identify independent predictors. The model was visualized as a nomogram, evaluated via Kaplan-Meier survival curves, receiver operating characteristic (ROC) curves, and decision curve analysis (DCA).

Age, tumor number, tumor size, gamma-glutamyl transferase (GGT), and total bilirubin (TBIL) were identified as independent RFS factors. The nomogram developed based on the five factors exhibited good discriminative ability: AUC values for 1-, 3-, 5-year RFS were 0.759, 0.777, 0.783 (training cohort), 0.708, 0.751, 0.714 (internal validation cohort), and 0.701, 0.708, 0.751 (external validation cohort). Kaplan-Meier curves confirmed significant RFS differences between high/low-risk groups (all P<0.001), and DCA demonstrated positive net clinical benefit.

This study successfully developed a well-performed nomogram model to predict 1-, 3-, and 5-year RFS in HBV-related HCC patients following TACE combined with RFA.