Yan Feng (1), Williams Daphne (1), David Berman, Eric Masson (1), and Amit Roy* (1)
(1) Bristol-Myers Squibb, Lawrenceville, NJ 08648, USA
Objectives: Ipilimumab is a fully human monoclonal antibody directed against cytotoxic T-lymphocyte antigen-4 (CTLA-4) that is being developed as a novel immunotherapeutic agent against melanoma, and other solid tumors. The main safety concern with ipilimumab is irAEs that may occur as a consequence of an increase in activated T-cells. The objective of this analysis was to describe the ipilimumab exposure-response (E-R) relationship of the occurrence, severity, and resolution of irAEs, incorporating the modulating effect of steroids that may have been administered to manage severe irAEs.
Methods: The E-R relationship of irAEs was described with data pooled from 498 patients who participated in four phase 2 studies of ipilimumab (CA184004, CA184007, CA184008 and CA184022). A Markov model was employed to describe the probability of transitioning between two of the following three possible irAE states: Grade less than 1 (including no irAE), Grade 2, and Grade 3+ irAE. The transition probabilities were described as a function of ipilimumab serum concentration-time (C-T) profile (predicted by a previously developed population pharmacokinetc [PPK] model) and prior steroid doses. The PPK and E-R models were applied to predict the time-course of irAEs for alternative ipilimumab dosing regimens by clinical trial simulation, assuming no drop-out (to obtain a conservatively high prediction of % irAEs).
Results: Model predicted maximal % of subjects with Grade 2+ irAEs on a given day (mpirAE) were 16.9% and 24.4% for 3 mg/kg and 10 mg/kg respectively, which were in good agreement with observed data (19.0% and 24.3%, respectively). The model predicted time to mpirAE was 78-84 days, which was slightly longer than observed data (43-73 days), indicating that the % of subjects with Grade 2+ irAE may reach a plateau earlier than predicted. The clinical trial simulation results predict that the prevalence of irAEs appears to reach a plateau by the end of 12 weeks, and suggest that continuation of induction doses beyond 12 weeks does not lead to markedly higher rate of irAEs.
Conclusions: The Markov E-R model was predictive of the observed irAE rate. This model suggests that the risk of irAE is near maximal by end of 12 weeks, and that extending the induction dosing regimen (every 3 weeks) beyond 12 weeks does not markedly increase the risk of irAEs. Prospective studies would be required to test this hypothesis and evaluate the benefit/risk of continuous dosing.
References:
[1] Feng Y, Masson E, Berman D, Parker SM, Hoos A, Chen T-T, Pfister M, and Roy A. Characterization of Ipilimumab Exposure-Efficacy/Safety Response Relationships in Advanced Melanoma Patients, American Conference on Pharmacometrics, San Diego, CA (April 2011).
Reference: PAGE 21 (2012) Abstr 2433 [www.page-meeting.org/?abstract=2433]
Poster: Safety (e.g. QT prolongation)