Translational Modeling of Anticancer Efficacy to Predict Clinical Outcomes in a First-in-Human Phase 1 Study of MDM2 Inhibitor HDM201.

We report on a retrospective model-based assessment of the predictive value of translating antitumor drug activity from in vivo experiments to a phase I clinical study in cancer patients treated with the MDM2 inhibitor, HDM201. Tumor growth inhibition models were developed describing the longitudinal tumor size data in human-derived osteosarcoma xenograft rats and in 96 solid tumor patients under different HDM201 treatment schedules. The model structure describing both datasets captures the delayed drug effect on tumor growth via a series of signal transduction compartments, including a resistance component. The models assumed a drug-killing effect on both sensitive and resistant cells and parameterized to estimate two tumor static plasma drug concentrations for sensitive (TSC) and resistant cells (TSC). No change of TSC and TSC with schedule was observed, implying that antitumor activity for HDM201 is independent of treatment schedule. Preclinical and clinical model-derived TSC were comparable (48 ng/mL vs. 74 ng/mL) and demonstrating TSC as a translatable metric for antitumor activity in clinic. Schedule independency was further substantiated from modeling of clinical serum growth differentiation factor-15 (GDF-15) as a downstream marker of p53 pathway activation. Equivalent cumulative induction of GDF-15 was achieved across schedules when normalized to an equivalent total dose. These findings allow for evaluation of optimal dosing schedules by maximizing the total dose per treatment cycle while mitigating safety risk with periods of drug holiday. This approach helped guide a phase I dose escalation study in the selection of an optimal dose and schedule for HDM201.