Integrated clinicopathological model versus TNM for predicting survival in resected hepatocellular carcinoma: a retrospective cohort study.

Objective: To explore prognostic factors of postoperative hepatocellular carcinoma (HCC) and compare the prognostic performance of the tumor-node-metastasis (TNM) system with an integrated clinicopathological model.

Methods: We conducted a cohort study evaluating 249 patients with HCC who underwent radical liver resection between January 2015 and December 2024. Overall survival (OS) was the primary endpoint. Independent prognostic factors were determined using the Cox proportional hazards model. Model performance was assessed using the correlation index (C-index) with bootstrap internal validation. Time-dependent receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration plots were used to assess discrimination, clinical utility, and agreement between predicted and observed outcomes.

Results: In multivariable analysis, TNM stage, microvascular invasion (MVI), resection margin status, and albumin-bilirubin (ALBI) grade were independent predictors of OS. For recurrence-free survival (RFS), independent predictors included hepatitis status, ALBI grade, TNM stage, and MVI. The integrated model demonstrated improved discrimination compared with TNM alone, with a higher optimism-corrected C-index (0.676 vs. 0.647). Time-dependent ROC analysis showed significantly higher AUC values for the integrated model at 1 year (0.725 vs. 0.669, p = 0.046), 3 years (0.703 vs. 0.657, p = 0.024), and 5 years (0.684 vs. 0.649, p = 0.031). DCA indicated greater net benefit across clinically relevant thresholds, and calibration plots showed good agreement between predicted and observed survival probabilities.

Conclusion: An integrated clinicopathological model incorporating TNM stage, MVI, ALBI grade, and resection margin status improves prognostic performance compared with TNM alone in patients undergoing curative resection for HCC. This model may support more accurate postoperative risk stratification.