Simultaneous Ocular Adverse Event and Treatment Discontinuation Model of Pimasertib
P. Girard (1), B. Brockhaus (2), G. Massimini (1), E. Asatiani (1), N. Rejeb (1), R.A. Rajeswaran (1), C. Lüpfert (2), O. von Richter (2), A. Munafo (1)
(1) Merck Serono S.A., Geneva*, (2) Merck KGaA, Darmstadt, Germany
Objectives: Pimasertib is an oral inhibitor of MAPK/ERK Kinase, currently developed for treatment of pancreatic cancer and melanoma, that was given to patients with solid tumors and hematological malignancies in 2 phase I dose escalation studies. All patients were monitored for tolerability (diarrhea, skin rash, ophthalmology, etc). Since ocular adverse events (OAE) were the main dose-limiting events, the objective of present analysis was to develop a joint model for exposure, OAE and treatment discontinuation (TD).
Methods: Population PK yielded estimates of individual exposure. A proportional odds model, with Markov components, was build for OAE weekly highest grade. The logits of the different cumulative probabilities (P) were a non linear function of AUC accumulating in a KPD compartment . In addition, Cmax was tested as acting either continuously or on 1st month. Other covariates were schedule (QD/BID), hypertension history, comedications and demographic covariates. Time to TD and OAE were jointly modeled using Weibull hazard and completely at random, random and informative TD assumptions were tested . All models were built using NONMEM (FOCEI for PK, Laplacian for OAE and TD).
Results: 199 patients, receiving total daily dose ranging from 1 to 255 mg contributed to 4766 PK, OAEs or TD observations. A 3 compartment PK model with 1st order absorption and lag time provided individual Cmax and weekly AUC estimates. OAEs were fitted to the odds model. Markov parameters and AUC mediated through a KPD Emax model were highly significant. BID regimen was associated with a reduction in P(OAE), while higher Cmax was significantly but transiently increasing them on 1st month. For TD, Kaplan Meier curves showed that patients with either OAEs or highest doses were more likely to stay on treatment. Neither the random nor the informative TD models were successful and a simpler TD model with daily dose was found significant.
Conclusion: This analysis showed that OAE are related to higher exposure (AUC, Cmax) and to QD administration. Presumably, BID regimen reduces P(OAE) by reducing peak concentrations, as indicated by higher P(OAE) linked to higher Cmax during 1st month of treatment. High TD rate was found to be dose related, with higher doses leading to less TD. This suggests a potential treatment benefit but needs to be tested further using efficacy data. Model results and simulations are being used to support the choice of dosing regimen for future studies.
 Hénin E, You B, VanCutsem E, Hoff PM, Cassidy J, Twelves C, Zuideveld KP, Sirzen F, Dartois C, Freyer G, Tod M, Girard P. A dynamic model of hand-and-foot syndrome in patients receiving capecitabine. Clin Pharmacol Ther. 2009;85:418-25.
 Hu C, Szapary PO, Yeilding N, Zhou H. Informative dropout modeling of longitudinal ordered categorical data and model validation: application to exposure-response modeling of physician's global assessment score for ustekinumab in patients with psoriasis. J Pharmacokinet Pharmacodyn. 2011;38:237-60.
This study was funded by Merck Serono S.A. – Geneva, Switzerland.*
All authors are employees of MerckSerono S.A.*
*An affiliate of Merck KGaA, Darmstadt, Germany.