IL‐5, a key mediator of type 2 inflammation, underlies various diseases, including severe asthma, CRSwNP, EGPA, and HES. Reduction in blood eosinophil count (BEC), a biomarker of IL‐5 activity, is commonly used to evaluate the efficacy of anti‐IL‐5 biologic therapies. Model‐informed drug development (MIDD) and quantitative decision making (QDM) were used to shorten the clinical development of depemokimab (an ultra‐long‐acting anti‐IL‐5 biologic). A Bayesian nonlinear mixed effects dose–time response model predicted the depemokimab dose in severe asthma achieving comparable BEC reductions to those observed in mepolizumab (an approved anti‐IL‐5 biologic) Phase III MUSCA and MENSA trials. Prespecified QDM go/no‐go criteria were applied to assess success probability. Phase IIb efficacy‐based trial simulations were conducted using negative binomial distribution to simulate individual annualized exacerbation rate. A depemokimab PK/PD (BEC) model predicted Phase III trial doses in CRSwNP/EGPA/HES. Single depemokimab doses were well‐described by the Bayesian model; a single depemokimab dose ≥ 60 mg had probability ≥ 80% of exceeding Minimum (78%; MUSCA) and ≥ 10% probability of exceeding Target (84%; MENSA) values for trough BEC reduction from baseline vs. placebo. Clinical trial simulations demonstrated < 3% probability of more precise estimation of the Phase III dosing regimen with a conventional efficacy‐based dose‐ranging study. Depemokimab 100 mg for severe asthma/CRSwNP and 200 mg for EGPA/HES, administered subcutaneously every 26 weeks, were selected for Phase III trials. MIDD and QDM shortened the depemokimab development program by 2–3 years, emphasizing the potential of this approach for progressing new therapies from Phase I directly to Phase III.
Zecchin et al. (Tue,) studied this question.