Validation of a neutropenia PK/PD model built from intravenous vinflunine and its application to design phase I trials with oral vinflunine
Laurent Nguyen, Aurelie Petain and Christian Puozzo
Pharmacokinetic department of Oncology Division, Institut de Recherche Pierre Fabre, Castres, France
Objectives: Vinflunine (VFL) is a fluorinated microtubule inhibitor and its dose limiting toxicity is neutropenia [1]. A semi-physiological population PK/PD model of chemotherapy-induced myelosuppression had been previously developed from early phase I data with intravenous (IV) VFL [2]. The first objective of the study was to qualify this PK/PD model based on larger datasets from VFL IV studies with and without PK information. The second objective was to use this model for designing phase I studies with oral VFL and exploring new dosing schedules.
Methods: The model was developed from single agent studies (n=210 patients and 423 administrations) with different schedules (D1q1W, D1q3W, D1/D8q3W). A linear four-compartment model was used to fit the PK data. The time course of neutrophils was fitted using the established PD model [2] consisting in 1 proliferative, 3 transit and 1 circulating compartments plus a feed-back loop. Data were analysed with the NONMEM program (version VI.0) using the FOCE interaction method.
An internal validation was performed using VPC for both PK and PD data. An external validation was performed on clinical data not including PK: frequencies of toxicity grades were compared between simulations and observed IV data.
Simulations of fractioned oral doses were performed by mimicking oral PK profiles (information obtained from a pilot bioavailability study) and by comparing haematological toxicity among different schedules with the same dose intensity.
Results: Both the VFL concentrations and the neutrophil data were well described by the PK/PD model. VFL PK showed moderate inter-patient (CV ~ 25%) and little inter-occasion (CV < 10%) variabilities whereas PD variabilities were higher (CV > 20%). External validation demonstrated that mild or severe grades of neutropenia were well predicted following IV administration of VFL.
Concerning oral VFL simulations, lower doses and more continuous administrations (shorter wash out interval) were associated with better tolerability (e.g.: D1-5, D8-12 q3W vs D1-D5q3W). Constant exposure (daily regimen) showed also good tolerability.
Conclusions: The validation confirmed the consistency of the model with IV VFL and enabled to implement and use new clinical designs for oral VFL. The current preliminary results in phase I trials are in line with the simulations.
References:
[1] Bennouna J et al. Clin Cancer Res. 14 (6):1625-32, 2008
[2] Friberg LE et al. J Clin Oncol 20:4713-4721, 2002
