Continuous-Time Markov Population PK/PD Modeling of Longitudinal EASI Categorical Score in Atopic Dermatitis Treated With Rocatinlimab, an Anti-OX40 Monoclonal Antibody.

Abstract

Rocatinlimab (AMG 451/KHK4083) is a potential T-cell rebalancing therapy that inhibits and reduces the number of pathogenic T cells by targeting the OX40 receptor; it is associated with a progressive and sustained response for patients with moderate-to-severe atopic dermatitis (AD). This population pharmacokinetic-pharmacodynamic (PPK-PD) analysis of rocatinlimab analyzed the longitudinal relationship between drug exposure and response using a four-state continuous-time Markov model. The model was developed using a two-compartment PPK model for exposure based on five clinical studies and the categorical scores of Eczema Area and Severity Index (EASI) response from two clinical studies in AD (NCT03096223/NCT03703102). EASI categorical score was classified by percentage reduction from baseline (EASI₀: < 50%; EASI₅₀: ≥ 50% to < 75%; EASI₇₅: ≥ 75% to < 90%; EASI₉₀: ≥ 90%) following rocatinlimab exposure. Overall, 413 patients/healthy subjects were included in the full PPK dataset. The observed serum concentration-time data following rocatinlimab exposure was well described using the two-compartment PPK model. For PPK-PD analysis, EC₅₀ and E_max estimates were 24.0 μg/mL (% relative standard error [RSE]: 32.4) and 1.26 (%RSE: 25.4), respectively, with dropout probability for EASI₀ (0.0291) greater than for other EASI categories. Race was identified as a covariate but was not clinically meaningful. The PPK-PD model adequately described the time course and drug-related change in EASI score and was used to simulate various dosage regimens to predict the time course of EASI response. The identified parameters and drug-exposure relationship were considered to potentially support dose revisions and optimization for phase 3 trials in patients with moderate-to-severe AD that are utilizing a subcutaneous regimen with dosing every 4 weeks.