Analysis of risk factors and prediction of prognosis in patients with primary liver cancer undergoing transarterial chemoembolization.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND: Transarterial chemoembolization (TACE) is a key locoregional therapy for unresectable primary liver cancer, yet post-TACE survival remains heterogeneous. We evaluated prognostic determinants and developed a prediction model for overall survival. METHODS: This retrospective cohort study included consecutive adults undergoing index TACE between January 2022 and December 2023. Patients were followed until death or November 30, 2025. Overall survival (OS) and progression-free survival (PFS) were estimated using Kaplan-Meier methods. Candidate predictors were screened using univariable Cox regression and entered into multivariable Cox modeling. A prognostic risk score was derived from model coefficients; discrimination was assessed by Harrell's concordance index (C-index) and time-dependent receiver operating characteristic (ROC) analyses with area under the ROC curve (AUC), with internal validation. RESULTS: Among 185 included patients (median follow-up 30.6 months), 85 deaths (45.9%) and 135 progression/death events (73.0%) occurred. Median OS was 28.0 months (12- and 24-month OS: 76.4% and 55.7%), and median PFS was 10.9 months (12- and 24-month PFS: 45.0% and 25.9%). Modified Response Evaluation Criteria in Solid Tumors (mRECIST) responses yielded an objective response rate (ORR) of 48.1% and disease control rate (DCR) of 73.5%. Independent predictors of worse OS included maximum tumor diameter (hazard ratio [HR] 1.137 per 1 cm), portal vein tumor thrombosis (PVTT; HR 2.616), extrahepatic metastasis (HR 1.839), Child-Pugh class B (HR 1.991), international normalized ratio (INR; HR 1.529 per 0.1), and alpha-fetoprotein (AFP; HR 1.946 per 1 log10). The model showed good discrimination (C-index 0.812 training; 0.791 validation) and outperformed established staging/scores. CONCLUSIONS: A parsimonious Cox-based model integrating tumor burden, disease extent, hepatic reserve, coagulation status, and tumor biology provides individualized OS risk stratification after TACE.