Prognostic Prediction and Risk Stratification of Transarterial Chemoembolization Combined with Targeted Therapy and Immunotherapy for Unresectable Hepatocellular Carcinoma: A Dual-Center Study

Purpose: The combination of transarterial chemoembolization, molecular targeted therapy, and immunotherapy (triple therapy) has shown promising outcomes in the treatment of unresectable hepatocellular carcinoma (HCC). This study aimed to build a prognostic model to identify patients who could benefit from triple therapy. Patients and Methods: This retrospective study encompassed 242 patients with HCC who underwent triple therapy from two centers (Training cohort: 158 patients from the Center 1; External validation cohort: 84 patients from the Center 2). Independent predictors of overall survival (OS) and progression-free survival (PFS) were identified through Cox regression analyses, and prognostic models based on Cox proportional hazards models were developed. Prognosis was assessed using Kaplan – Meier curves. Results: In the training cohort, independent predictors of PFS included vascular invasion and the C-reactive protein and alpha-fetoprotein in immunotherapy (CRAFITY) score. Independent predictors of OS were the CRAFITY score, extrahepatic metastasis, and the neutrophil-to-lymphocyte ratio. Prognostic prediction models were constructed based on these variables. The prognostic model for OS demonstrated a C-index of 0.715 (95% confidence interval (CI), 0.662– 0.768) in the training cohort and 0.701 (95% CI, 0.628– 0.774) in the validation cohort. Patients were divided into low- and high-risk categories using the predictive model (P< 0.001). These findings were corroborated by the external validation cohort. Conclusion: The developed prognostic model serves as a reliable and convenient tool to predict outcomes in patients with unresectable HCC undergoing triple therapy. It aids clinicians in making informed treatment decisions. Keywords: hepatocellular carcinoma, combined regimen, transarterial chemoembolization, prognostic model, risk stratification, immunotherapy