Exposure–response analysis using time-to-event data for bevacizumab biosimilar SB8 and the reference bevacizumab

Abstract

Purpose: This analysis aimed to characterize the exposure–response relationship of bevacizumab in non-small-cell lung cancer (NSCLC) and evaluate the efficacy of SB8, a bevacizumab biosimilar, and Avastin®, the reference bevacizumab sourced from the European Union (EU), based on the exposure reported in a comparative phase III efficacy and safety study (EudraCT, 2015-004026-34; NCT 02754882).Materials and methods: The overall survival (OS) and progression-free survival (PFS) data from 224 patients with steady-state trough concentrations (Css,trough) were analyzed. A parametric time-to-event (TTE) model was developed using NONMEM®, and the effects of treatments (SB8 and bevacizumab-EU) and patient demographic and clinical covariates on OS and PFS were evaluated. Simulations of median OS and PFS by bevacizumab Css,trough were conducted, and concentrations required to achieve 50% and 90% of the maximum median TTE were computed.Results: A log-logistics model with Css,trough best described the OS and PFS data. Treatment was not a predictor of the hazard for OS or PFS. Simulations revealed steep exposure–response curves with a phase of rapid rise before saturating to a plateau. The median Css,trough values of SB8 and bevacizumab-EU reported from the clinical study were on the plateaus of the exposure–response curves. The concentrations required to achieve 50% and 90% of the maximum effect were 82.4 and 92.2 μg/mL, respectively, for OS and 79.7 and 89.1 μg/mL, respectively, for PFS.Conclusion: Simulations based on the constructed TTE models for OS and PFS have well described the exposure–response relationship of bevacizumab in advanced NSCLC. The analysis demonstrated comparable efficacy between SB8 and bevacizumab-EU in terms of OS and PFS based on their exposure levels.