Prediction of occurrence of thrombocytopenia to select Phase 1b dose and dosing regimen for a selective inhibitor of p53-MDM2 in patients with solid tumors
S. Retout (1), C. Meille (2), G. Nichols (3), S. Middleton (3), D. Bottino (4), N. Frey (1)
(1) F. Hoffmann-La Roche Ltd., Pharma Research and Early Development, Translational Research Sciences, Modeling and Simulation, Basel, Switzerland; (2) F. Hoffmann-La Roche Ltd., Non Clinical Safety, DMPK, Modeling and Simulation, Basel, Switzerland; (3) Hoffmann-La Roche Inc., Pharmaceutical Research and Early Development, Oncology Translational Medicine Group; (4) Hoffmann-La Roche Inc., Pharma Research and Early Development, Translational Research Sciences, Modeling and Simulation, Nutley NJ, USA
Objectives: RG7112 is a selective inhibitor of p53-MDM2 binding that frees p53 from negative control, activating the p53 pathway in cancer cells leading to cell cycle arrest and apoptosis. Data collected in a Phase Ia program reported some occurrence of delayed thrombocytopenia (TCP) for patients receiving doses of RG7112 for 10 days quo die every 28 days generating exposures at or near the anticipated expected therapeutic range. The aim of this study was to develop a PK/PD model describing the time course of platelets after RG7112 and to identify doses and dosing regimens that would maintain the exposure into the therapeutic range while limiting the occurrence of TCP in Phase Ib.
Methods: The analysis included 1141 platelets observations from patients receiving RG7112 doses on a 28 day cycle basis, with administrations every day during either 10 days (N=70, 30 mg to 3900 mg) or during 5 days (N=22, 2500 mg). Individual concentration-time profiles were predicted from a previous population PK analysis and a semi-mechanistic model was used to describe the platelet time course. This model was developed pre-clinically to analyze platelet changes in different species receiving RG7112; it involves 5 compartments to link the progenitors in bone marrow to platelets in circulation via a maturation process . The parameters were estimated using NONMEM 7.1 (FOCE INTER). Due to sustained thrombocytopenia observed in some patients, an effect compartment was used to account for a potential accumulation of the drug at the site of action. Model performance was assessed using both goodness of fit plots and posterior predictive check. Then, for different daily doses (1500 mg to 5000 mg) and regimens (28 day cycles with daily doses during 3, 5, or 10 days), we investigated by simulation the tradeoff between the percentage of patients with TCP in the first cycle and the percentage of patients above a given threshold of exposure for anti-tumor activity.
Results: The model well characterizes the platelets time course including the nadir and it shows good predictive performance of occurrence of TCP. The simulations show that, for less than 20% of patients with TCP, 4500 mg daily of RG7112 during 3 days would provide the best benefit/risk ratio.
Conclusions: A robust model of the relationship between the PK of RG7112 and its effect on platelet time course has been developed. It provides a robust quantitative tool to support the dose and dosing regimen selection for future studies.
 Friberg L. et al. J Clin Oncol. 2002;20(24):4713-21.