An Albumin-Bilirubin (ALBI) Grade-based Prognostic Model For Patients With Hepatocellular Carcinoma Within Milan Criteria.

Abstract

The Milan criteria are recommended as the major reference for liver transplantation in patients with small hepatocellular carcinoma (HCC). However, alternative anticancer treatments are often utilized due to severe donor organ shortage. This study aimed to develop and validate an albumin-bilirubin (ALBI) grade-based prognostic model to stratify survival in patients within Milan criteria undergoing nontransplant therapy. A total of 1655 patients were assigned into the derivation and validation cohort according to treatment modalities. Multivariate analysis was used to identify independent predictors of survival in the derivation cohort. An ALBI-based model was evaluated in the validation cohort. In the Cox multivariate model, age 65 years or older (hazard ratio [HR]=1.576, P<0.001), serum α-fetoprotein (AFP) level >100 ng/mL (HR=1.671, P<0.001), ascites (HR=1.808, P<0.001), performance status 1 to 4 (HR=1.738, P<0.001), ALBI grade 2 (HR=1.827, P<0.001), and ALBI grade 3 (HR=3.589, P<0.001) were independent predictors of poor survival in the derivation cohort. An ALBI-based prognostic model with a total of 0 to 6 points was derived with the sum of 5 variables: 1 point each for age 65 years or older, AFP >100 ng/mL, presence of ascites, performance status 1 to 4, and ALBI grade 2, and 2 points for ALBI grade 3. This model can accurately predict long-term outcome in the validation cohort (P<0.001) and discriminate survival in patients stratified by curative and noncurative treatments (both P<0.001). The proposed ALBI grade-based model is feasible in predicting survival in HCC patients within the Milan criteria, and helps identify high-risk patients who need timely liver transplantation.