PURPOSE: The objective of the present study was to rank the most common liver staging systems according to prognostic accuracy in patients with unresectable hepatocellular carcinoma (HCC) treated with transarterial chemoembolization (TACE). MATERIALS AND METHODS: Survival of 172 consecutive patients with unresectable HCC treated with TACE was correlated with the pretreatment Child-Pugh (categoric and nominal), Okuda, Cancer of the Liver Italian Program, Barcelona Clinic Liver Cancer, Model for End-stage Liver Disease, Chinese University Prognostic Index (CUPI), Japanese Integrated Staging, Tumor/Node/Metastasis, Group d'Etude de Traitement du Carcinoma Hepatocellulaire, Liver Cancer Study Group of Japan, and Tokyo staging systems. The systems were ranked according to error reduction in predicting median survival (Kaplan-Meier survival curve and Cox regression analysis). The error reduction was normalized to the error in predicting survival in the absence of a staging system. RESULTS: Median survival was 80 weeks. The error in predicting survival of an unstaged population was 51 weeks. The Child-Pugh nominal, CUPI, and Tokyo scores yielded the largest reduction in survival prediction error, at 20.8%, 21.6%, and 19.6%, respectively. Their actual error measurements in predicting survival were 40.4, 40.0, and 41.0 weeks, respectively. CONCLUSIONS: Child-Pugh nominal, CUPI, and Tokyo scores provide the best prognostic accuracy among the systems studied. However, those of the Tokyo and CUPI methods are artificially enhanced because of their greater number of staging levels. The Child-Pugh nominal liver staging system is the most accurate in predicting survival of patients with unresectable HCC treated with TACE, and it is recommended that it be adopted as the standard for HCC staging in such patients.