Population pharmacokinetic characteristics of cemiplimab in patients with advanced malignancies

Cemiplimab, a human monoclonal antibody targeting programmed cell death-1 (PD-1) receptor, demonstrated antitumor activity in patients with advanced malignancies and a safety profile comparable to other anti–PD-1 therapies. This population pharmacokinetics (PopPK) analysis of cemiplimab included 11,178 pharmacokinetics (PK) observations from 548 patients pooled from a first-in-human study (Study 1423; NCT02383212) in advanced malignancies and a Phase 2 study (Study 1540; NCT02760498) in advanced cutaneous squamous cell carcinoma (CSCC). Most patients (80.3%) received cemiplimab 3 mg/kg every 2 weeks (Q2W) intravenously (IV). A PopPK model was developed by evaluating two-compartment linear models with an empirical non-linear function describing time-varying change in cemiplimab clearance and covariates that improved goodness-of-fit. PopPK simulations were used to describe cemiplimab exposure generated by a fixed 350 mg every 3 weeks (Q3W) IV dose regimen. PopPK modeling showed that a two-compartment model with zero-order IV infusion rate and first-order elimination rate well described individual concentrations of cemiplimab. Although several covariates, including baseline body weight and albumin concentrations, had a modest impact on cemiplimab exposure, the magnitude of influence was within the typical observed PK variability of approximately 30%. Based on PopPK simulation results, the 350 mg Q3W dose regimen was selected for further studies in advanced malignancies, including advanced CSCC. Similarity in observed cemiplimab exposure at the fixed 350 mg Q3W and the weight-based 3 mg/kg Q2W dose regimens confirmed this fixed dose selection. A robust PopPK model was developed to describe cemiplimab concentrations and supported use of the fixed 350 mg Q3W IV dose regimen. SUPPLEMENTARY INFORMATION: The online version of this article (10.1007/s10928-021-09739-y) contains supplementary material, which is available to authorized users.