Dosing Regimen selection supported by population PKPD model of thrombocytopenia
Christophe Meille, Nelson Guerreiro, Claire Fabre and Astrid Jullion
Novartis, Basel, Switzerland
Objectives: The aim of this work was to develop a mechanism-based pharmacokinetic-pharmacodynamic (PK/PD) model that describes the longitudinal time-course of platelet changes in patients treated with the p53-HDM2 protein-protein interaction inhibitor HDM201, and to apply it to support the identification of an optimal dosing regimen that would be tolerated for at least six treatment cycles.
Methods: Plasma concentrations, platelet and platelet transfusions data were obtained from an on-going phase I study on patients with p53 wild-type solid tumors. In this first in human study, four oral regimens are being investigated: Q3W, day 1 and day 8 of a 4-week cycle, QD 2 week on/2 weeks off in a 4-week cycle, and QD 1 week on/3 weeks off in a 4-week cycle. Delayed thrombocytopenia is the primary reported dose limiting toxicity, often having an impact on long-term treatment sustainability with patients needing dose reductions and/or interruptions. Plasma drug concentrations and platelet data were fitted in a two-step approach using non-linear mixed-effects modeling implemented in Monolix 2016R1. The PK/PD model for thrombocytopenia was modified from Friberg et al. (2002) [1].
In a first step, the PK/PD model has been simulated using a Shiny application developed on R-3.2.3 [2,3] to graphically explore the platelet profiles under different regimens, and to derive key metrics such as time to thrombocytopenia event, recovery time and percentage of patients developing grade 3/4 thrombocytopenia during the 6 first treatment cycles. In a second step, optimization of the drug dosing regimen has been performed using the methodology described in Meille C et al, 2016 [4] extended to take into account inter-patient variability. Optimization criterion was selected as the maximum total drug per cycle (assuming efficacy schedule independency), while complying with predefined clinical constraints on platelet counts. For that purpose, a set of possible dosing regimens has been defined with daily dose amount ranging from 1mg to 500mg and number of days of administration ranging from 1 to the cycle length resulting in a simulations matrix. Then, platelet profiles of 500 virtual patients have been simulated over 6 cycles. In a final step, the proportion of patients with grade 4 thombocytopenia during the first 6 treatment cycles and the total dose amount per cycle have been derived and the most favorable dosing regimens identified.
Results: The concentration-time course of HDM201 was best described as a one-compartment model with a delayed zero- and first-order absorption process, and linear clearance (Cl/F). Body weight was identified as a continuous covariate showing significant influence on the central volume (V/F). The modifications from the model of Friberg et al. (2002) [1] were (1) to include a drug action decoupled from systemic feedback, and (2) to introduce an indirect drug effect on early proliferative cells and on systemic regulation through an effect compartment. In addition, platelet transfusion events were implemented as platelet infusions. Simulations matrix suggest that distributing the total dose in a concentrated period allows reaching a larger total dose per cycle while respecting the constraints on safety.
Conclusions: This work suggests a methodology where a population PKPD model of safety endpoint can be used to support dosing regimen selection, satisfying clinical constraints while maximizing the dose. Possible extension could be to optimize efficacy through a PKPD model on tumor growth in clinics.
References:
[1] Friberg et al, 2002. “Model of chemotherapy-induced myelosuppression with parameter consistency across drugs.” J. Clin. Oncol. 20:4713–4721.
[2] https://CRAN.R-project.org/package=shiny
[3] https://CRAN.R-project.org/package=mlxR
[4] Hénin E, Meille C, Barbolosi D, You B, Guitton J, Iliadis A, Freyer G. Revisiting dosing regimen using PK/PD modeling: the MODEL1 phase I/II trial of docetaxel plus epirubicin in metastatic breast cancer patients. Breast Cancer Res Treat. 2016 Apr;156(2):331-41. doi: 10.1007/s10549-016-3760-9. Epub 2016 Mar 22.