Evaluation of prognostic scores in patients with HCC undergoing first-line immunotherapy with atezolizumab and bevacizumab

Abstract

Background & Aims: Immunotherapy with atezolizumab and bevacizumab (a + b) has improved the prognosis of patients with unresectable hepatocellular carcinoma (HCC). However, the outcome for individual patients is highly variable. This study aimed to (i) develop and validate a prognostic prediction model to estimate individual prognosis and (ii) compare it with established models. Methods: In this multicenter retrospective study, patients with HCC undergoing first-line immunotherapy with a + b from 24 centers (Europe, USA) were included. Statistical analysis and reporting followed the TRIPOD guidelines. The primary objective was overall survival (OS). A Cox model was developed and externally validated. Results: In total, 683 patients were included (training: 526, validation: 157). The C-reactive protein, albumin, bilirubin, lymphocytes, ECOG performance status, and extrahepatic spread (CABLE score) remained significantly associated with OS in Cox regression analysis. In the training set, the CABLE score had a higher discriminatory accuracy relative to ALBI, EZ-ALBI, mALBI, CRAFITY, PNI, NLR, PLR, and GPS (time-dependent AUC 0.79 and C-index 0.75 (95% CI 0.71–0.78) at 12 months). In the external validation set, the discriminatory performance of the CABLE score was comparable to ALBI, EZ-ALBI, and mALBI, but on average higher than PNI, CRAFITY, NLR, PLR, and GPS. In patients with Child-Pugh A, the CABLE score outperformed ALBI, EZ-ALBI, and mALBI in the first 9 months. We provide a web-based calculator for the CABLE score to allow estimation of individual prognosis for these patients (http://shiny.imbei.uni-mainz.de:3838/CABLE_Score/). Conclusions: The CABLE score shows good discriminatory performance in assessing the individual prognosis of patients undergoing first-line immunotherapy with a + b. Further validation studies are needed to investigate its performance compared with the ALBI score, in particular in subgroup analysis. Impact and implications: The CABLE score allows estimation of the prognosis of patients with unresectable hepatocellular carcinoma undergoing first-line immunotherapy with atezolizumab and bevacizumab at an individual level using our web-based calculator. This feature, as well as the evaluation of the score's added benefit through an extensive comparison with other established scores, can inform clinicians on their significance and may guide clinical decision-making in the context of a malignant disease where the prognosis has become highly variable. Further large validation studies are needed to investigate the incremental value of the CABLE score compared with the ALBI score, in particular in subgroups such as patients designated as Child-Pugh A.