ABSTRACT: Introduction: We compared the efficacy of the ALBI (albumin-bilirubin) score to the established Child-Pugh (CP) grade in hepatocellular carcinoma (HCC) patients treated with yttrium-90 radioembolization (Y90). We further assessed the individual contributions of albumin and bilirubin to survival prediction. Methods: 1000 consecutive HCC patients treated with Y90 were included. Overall survival (OS) was assessed using Kaplan Meier analysis. Sub-stratification analyses were performed using CP and ALBI and in subgroups determined by United Network for Organ Sharing (UNOS) or Barcelona Clinic Liver Cancer (BCLC) staging. The independent impact (hazard ratio (HR)) of ALBI, CP, albumin, and bilirubin on survival was assessed using Cox proportional hazards analysis. Results: Median OS for ALBI 1, 2, and 3 grades was 46.7, 19.1, and 8.8 months, respectively. The HR for death for ALBI 2 vs. ALBI 1 was 3.39 (1.75-6.57); ALBI 3 vs. ALBI 1 was 7.58 (3.89-14.79); and the c-index was 0.623. Median OS for CP A, B, and C was 21.7, 11.3, and 6.0 months, respectively. The HR for death for CP B vs. CP A was 2.04 (1.71-2.43); CP C vs. CP A was 3.27 (2.08-5.14); and the c-index was 0.616. Stratified OS showed unique prognostic groups identified by ALBI within CP-B and CP-C. Median OS for albumin grades 1, 2, and 3 was 46.0, 17.1, and 9.1 months, respectively. Median OS for bilirubin grades 1, 2, and 3 was 15.6, 21.0, and 5.8 months, respectively. The HR for death for albumin 2 vs. 1 was 2.48 (1.81-3.41); albumin 3 vs. 1 was 4.74 (3.44-6.54); and the c-index was 0.640. The HR for death for bilirubin 2 vs. 1 was 1.09 (0.82-1.44); bilirubin 3 vs. 1 was 2.37 (1.66-3.40); and the c-index was 0.533. Conclusions: ALBI outperforms CP in survival prognosis in Y90 treated patients. On sub-analyses, serum albumin (not bilirubin) appears to be the main driver of survival prediction. Our study supports the prognostic ability of ALBI and may suggest a role of albumin alone as a biomarker for patients with HCC.
