BACKGROUND Prognosis prediction in patient with hepatocellular carcinoma (HCC) after transarterial radioembolization (TARE) remains difficult. The aim of this study was to develop a prognostic model to aid in the decision to use TARE.

METHODS A total of 174 patients in Korea who underwent TARE for HCC as the initial treatment were included. We developed a prediction model for overall survival (OS) based on independent risk factors for OS and validated the model by bootstrap method.

RESULTS The median maximal size of the tumors was 8.2 cm, the median number of tumors was 2, and the median albumin level was 4.0 g/dL. Portal vein tumor thrombosis was found in 46.0% (Vp1-3 [39.7%] and Vp4 [6.3%]). Four independent risk factors associated with OS (maximal tumor size, tumor number, albumin, and portal vein tumor thrombosis) were used to develop the SNAP-HCC score. Bootstrap validation of the scoring index determined that the Harrell's c-index for OS was 0.756 (95% confidence interval: 0.729-0.783). Patients grouped based on their SNAP-HCC (scores 0-5) were well discriminated, with significant differences between the groups (all P < 0.05). Patients with SNAP-HCC < 3 showed significantly longer OS than patients with SNAP-HCC ≥ 3 (P < 0.001). The respective survival probabilities at years 1 and 3 were 0.81 and 0.73 in the low-risk (SNAP-HCC < 3) and 0.32 and 0.14 in the high-risk (SNAP-HCC ≥ 3) patients.

CONCLUSIONS The SNAP-HCC scoring system predicted the outcome of HCC patients undergoing TARE as an initial treatment. This model could be helpful for initial planning the treatment of HCC patients.