In Silico Re‐Optimization of Atezolizumab Dosing Using Population Pharmacokinetic Simulation and Exposure–Response Simulation

Atezolizumab, a humanized monoclonal antibody against programmed cell death ligand 1 (PD-L1), was initially approved in 2016, around the same time that the sponsor published the minimum serum concentration to maintain the saturation of receptor occupancy (6 μg/mL). The initially approved dose regimen of 1200 mg every 3 weeks (q3w) was subsequently modified to 840 mg q2w or 1680 mg q4w through pharmacokinetic simulations. Yet, each standard regimen yields steady-state trough concentrations (CMIN,SS ) far exceeding (≈ 40-fold) the stated target concentration. Additionally, the steady-state area under the plasma drug concentration-time curve (AUCSS ) at 1200 mg q3w was significantly (P = .027) correlated with the probability of adverse events of special interest (AESIs) in patients with non-small cell lung cancer (NSCLC) and, coupled with excess exposure, this provides incentive to explore alternative dose regimens to lower the exposure burden while maintaining an effective CMIN,SS . In this study, we first identified 840 mg q6w as an extended-interval regimen that could robustly maintain a serum concentration of 6 μg/mL (≥99% of virtual patients simulated, n = 1000), then applied this regimen to an approach that administers 2 "loading doses" of standard-interval regimens for a future clinical trial aiming to personalize dose regimens. Each standard dose was simulated for 2 loading doses, then 840 mg q6w thereafter; all yielded cycle-7 CMIN,SS values of >6 μg/mL in >99% of virtual patients. Further, the AUCSS from 840 mg q6w resulted in a flattening (P = .63) of the exposure-response relationship with adverse events of special interest (AESIs). We next aim to verify this in a clinical trial seeking to validate extended-interval dosing in a personalized approach using therapeutic drug monitoring.