Establishment and validation of a risk prediction model in patients with hepatocellular carcinoma treated with transarterial radioembolization

Background/aims Few studies have reported the treatment outcomes of transarterial radioembolization (TARE) using yttrium-90 (⁹⁰Y) for hepatocellular carcinoma (HCC). We established and validated a new risk prediction model for patients with HCC treated with TARE. Methods Between 2010 and 2017, 113 and 35 patients with intrahepatic HCC treated with TARE were selected for the training and validation cohorts, respectively. The modified response evaluation criteria in solid tumors (mRECIST) were used for response evaluation. Results In the training cohort, the median age was 64.1 years (92 males and 21 females) and the mean survival after TARE was 50.3 months. The cumulative survival rates at six and 12 months were 92.0 and 84.0%, respectively. A new risk prediction model for patients with HCC treated with TARE (Y-scoring system) was established from the training cohort using five independent baseline variables [serum albumin < 3.5 g/dL, hazard ratio = 5.446; alpha-fetoprotein > 200 ng/mL (hazard ratio = 5.071); tumor number ≥ 3 (hazard ratio = 2.933); portal vein thrombosis (hazard ratio = 4.915); and hepatic vein invasion (hazard ratio = 8.500)] and two on-treatment variables [no des-gamma-carboxy prothrombin response (hazard ratio = 15.346) and progressive disease at three months (hazard ratio = 4.154)] for mortality (all P < 0.05). The predictive accuracy of the Y-scoring system was acceptable to predict six [area under the curve (AUC) = 0.845], nine (AUC = 0.868), and 12-month mortality (AUC = 0.886) (all P < 0.05). The predictive accuracy of the system was similarly maintained in the validation cohort (AUC 0.737-0.901 at 6-12 months). Conclusion Our new risk prediction model can be used to stratify different prognoses in patients with HCC treated with TARE. Validation studies are required.