Abstract
<title>Abstract</title> <p> Background Hepatocellular carcinoma(HCC)shows substantial prognostic heterogeneity even within the same clinical stage. We aimed to compare survival across multimodal treatment patterns, identify independent predictors of overall survival༈OS༉, and develop a Cox-based risk model to support individualized management. Methods We retrospectively collected clinical data from 96 patients with primary HCC who received transarterial interventional therapy from January 2019 to January 2024. Diagnosis and staging followed the 2022 guideline for primary liver cancer. All patients received TACE-based therapy combined with targeted agents and immunotherapy, and some underwent sequential microwave ablation. OS was defined as time from the first interventional procedure to death or last follow-up(January 2024). Kaplan–Meier methods estimated survival. Prognostic factors were screened using univariable and multivariable Cox regression. A risk score and nomogram were constructed, and model performance was evaluated using 10-fold cross-validation, calibration curves, and time-dependent ROC at 1 year, 652 days, 2 years, and 3 years. Results The median age was 59 years, and 74% of patients were Child–Pugh A. Univariable Cox analysis showed that maximum tumor diameter, portal vein tumor thrombus, BCLC stage, ascites score, and ablation were significantly associated with OS. Multivariable Cox analysis identified sex, hepatitis B infection, tumor capsule, portal vein tumor thrombus, ascites score, bilirubin score, albumin score, prothrombin score, ablation, and maximum ablation diameter as independent prognostic factors. The optimized model achieved a C-index of 0.68 and an R² of 0.35, with a calibration slope close to 1.1. AUCs at 1 year, 652 days, 2 years, and 3 years were 0.805, 0.733, 0.707, and 0.681. The risk score effectively stratified patients into low, intermediate, and high-risk groups(log-rank <italic>P</italic> <0.001), with median OS of 808.5, 568.5, and 270.5 days, respectively. Conclusions Survival in intermediate-to-advanced HCC treated with TACE-based multimodal therapy is influenced by tumor characteristics, liver function, and ablation-related factors. The Cox-based model demonstrated acceptable discrimination, calibration, and stability, enabling clinically meaningful risk stratification and supporting individualized treatment planning. </p>