Dynamic changes of serum α-fetoprotein predict the prognosis of bevacizumab plus immunotherapy in hepatocellular carcinoma.

Abstract

Alpha-fetoprotein (AFP) has been established as a biomarker for hepatocellular carcinoma (HCC); however, whether its dynamic changes could predict the response to systemic therapy remains elusive. This study explored the AFP trajectory and the association with survival in patients received bevacizumab plus immunotherapy. We retrospectively enrolled 536 HCC patients received bevacizumab plus immunotherapy between February 2021 and February 2023. Patients were divided into two groups according to AFP values before treatment (400ng/ml). Dynamic changes of AFP were fitted using a latent class model to generate the AFP trajectories. Multivariable Cox models were utilized to compute hazard ratios (HRs) for survival. Inverse-probability-of-treatment weighted analyses were conducted to mitigate the influence of unmeasured confounding variables. The primary endpoint is progression free survival (PFS). The second endpoint is overall survival (OS). Three distinct trajectories were identified for AFP-low and AFP-high patients, respectively. In AFP-low group, compared with the high-rising class (25%; n=69), HRs of PFS were 0.39 and 0.2 for the low-stable class (59.1%; n=163) and sharp-falling class (15.9%; n=44), after adjusting by tumor diameter, tumor number, and extra-hepatic metastasis. In AFP-high group, compared with the high-stable class (18.5%; n=48), HRs of PFS were 0.3 and 0.04 for the middle-stable class (56.5%; n=147) and sharp-falling class (25%; n=65), after adjusting by tumor diameter, tumor number, and extra-hepatic metastasis. Furthermore, the AFP trajectories exhibited the utmost relative importance among all covariates regarding PFS and OS in the multivariable regression models. The AFP trajectories in HCC patients receiving bevacizumab and immunotherapy, constituted an independent biomarker indicative of clinical outcomes. Findings from this study hold potential clinical utility in dynamically forecasting the prognosis of systemic therapy in HCC patients and facilitating clinical decision-making. Rapid reduction of AFP post-treatment can lead to favorable patient prognoses.