Comparison of new prognostic systems for patients with resectable hepatocellular carcinoma: Albumin-Bilirubin grade and Albumin-Indocyanine Green Evaluation grade

Aim We aimed to compare the prognostic abilities of two novel liver function-estimating models, Albumin-Bilirubin (ALBI) and Albumin-Indocyanine Green Evaluation (ALICE) grades, in patients with hepatocellular carcinoma. Methods Data of 1270 patients who underwent initial hepatectomy for hepatocellular carcinoma between 1986 and 2016 were retrospectively collected from a multi-institutional database. The prognostic impact of each system was analyzed according to the results of the area under the receiver operating characteristic curve, the Cox regression model and the linear trend chi(2)-test. Results The ALBI and ALICE scores, which were obtained before grading status, were significantly correlated (correlation coefficient 0.930; P < 0.001). Both ALBI and ALICE grades stratified well in terms of overall survival, and were found to be independent prognostic factors on multivariate analysis (P < 0.05). The area under the receiver operating characteristic curves for 5-year survival in both groups were equivalent (0.602 vs. 0.614, P = 0.402); however, homogeneity, discriminatory ability, and the Akaike information criterion were superior for the ALICE grade than for the ALBI grade (73.8 vs. 65.7, 43.4 vs. 34.9, and 7204.1 vs. 7212.2, respectively). Conclusions Both grading systems could estimate the liver function of patients with hepatocellular carcinoma. Regarding hepatectomy patients, the ALICE grade was a more suitable model than the ALBI grade.