Introduction: To establish a predictive model to demonstrate that transcatheter arterial chemoembolization (TACE) prolonged survival time in patients with Barcelona Clinic Liver Cancer (BCLC) stage-C HCC.Material and methods: Patients with BCLC stage C HCC treated between January 2009 and April 2016 were included. The training group (n = 336) and control group (n = 141) underwent TACE as a first or supportive treatment. Factors related to survival time were retrospectively analyzed by multivariate logistic regression to establish a predictive model. Validation of the model was undertaken prospectively in a validation group (n = 159) that underwent TACE as first treatment between May 2016 and December 2017.Results: Classification of portal vein tumor thrombus, maximum tumor size, Child-Pugh score, and log alphafoetoprotein levels were independent risk factors included in the mathematical model. Receiver operating characteristic curves confirmed the model was helpful in predicting survival time. The area under the curve was 0.714 (95% confidence interval, 0.659-0.769). A cutoff value of 7.1 months had maximum Youden indeces of 0. 458, sensitivity 62.4%, and specificity 83.8%. The validation group supported the model.Conclusions: Portal vein tumor thrombus, maximum tumor size, Child-Pugh score, and log alphafoetoprotein levels helped predict survival time in patients with BCLC stage-C HCC.
Keywords: hepatocellular carcinoma; predictive model; survival analysis; transcatheter arterial chemoembolization.