Pretreatment eosinophil count predicts response to atezolizumab plus bevacizumab therapy in patients with hepatocellular carcinoma.

ABSTRACT

AIM: Pretreatment peripheral blood markers have value in predicting the treatment outcome of various cancers. In particular, the eosinophil count has recently gained attention. However, no study has reported the influence of the pretreatment eosinophil count on the outcomes of atezolizumab plus bevacizumab (ATZ/BEV), which is the recommended first-line systemic therapy for unresectable hepatocellular carcinoma (u-HCC). METHODS: We enrolled 114 patients with u-HCC treated with ATZ/BEV (n = 48) or lenvatinib (n = 66). The patients receiving ATZ/BEV or lenvatinib were divided into two groups by calculating the cutoff value of the pretreatment eosinophil count. The groups were compared regarding the clinicopathological characteristics, outcomes, and incidence of adverse events (AEs). RESULTS: Twenty-three of 48 patients (47.9%) who received ATZ/BEV therapy were categorized as the ATZ/BEV-eosinophil-high group, which had better responses than the ATZ/BEV-eosinophil-low group (P = 0.0090). Kaplan-Meier curves revealed a trend toward significantly better progression-free survival (PFS) in the ATZ/BEV-eosinophil-high group than the ATZ/BEV-eosinophil-low group (the median PFS 4.7 months in the ATZ/BEV-eosinophil-low group vs 12.6 months in the ATZ/BEV-eosinophil-high group; P = 0.0064). Multivariate analysis showed that a low eosinophil count was an independent risk factor for worse PFS after ATZ/BEV therapy (P = 0.0424, hazard ratio 2.24, 95% confidence interval 1.02-4.89). AEs (≥ grade 3) were significantly more likely to occur in the ATZ/BEV-eosinophil-high group (P = 0.0285). The outcomes did not significantly differ between the LEN-eosinophil-high group and the LEN-eosinophil-low group. CONCLUSION: A high pretreatment eosinophil count predicted a better response to ATZ/BEV therapy for u-HCC and was associated with the incidence of AEs (≥ grade 3).