A Bayesian population PK–PD model for ispinesib/docetaxel combination-induced myelosuppression

Abstract

Ispinesib, a kinesin spindle protein inhibitor, blocks assembly of a functional mitotic spindle, leading to G2/M arrest. Docetaxel promotes tubulin assembly into microtubules while inhibiting microtubule de-polymerization leading to mitotic arrest. Prolonged (> or =5 days) Gr 4 neutropenia and/or febrile neutropenia were the observed dose-limiting toxicities with both agents. Both agents are substrates and inhibitors of CYP3A4; thus, the potential for a drug-drug interaction exists. The goal was to fit a Bayesian population PK/PD model to characterize the relationship between the ispinesib/docetaxel combination and absolute neutrophil counts (ANC). Escalating doses of docetaxel (60-75 mg/m(2)) were administered over 1 h followed by a 1-h infusion of escalating doses of ispinesib (8-12 mg/m(2)) on a 21-day schedule. At least 3 pts were treated at each dose level. Limited PK samples were obtained. ANC were measured weekly on days 1, 8, 15, and 22. More ANC samples were taken from some subjects. The PK properties of ispinesib and docetaxel, and the relationship of PK with ANC were investigated using nonlinear mixed-effects models and Bayesian methods. With a limited dataset, informative prior distributions for the model parameters were needed. These prior distributions were formed using information from a previous study for ispinesib, and from the literature for docetaxel. Twenty-four pts were treated in this study. The PK of ispinesib and docetaxel were well characterized by a two-compartment model and a three-compartment model, respectively. There is no obvious PK interaction between ispinesib and docetaxel. The model for ANC consisted of a proliferating compartment, three transit compartments that represented maturation, and a compartment of circulating blood cells. This ANC model has been used previously for ispinesib given as monotherapy, and for other chemotherapeutic drugs in the literature. Using Bayesian methods, the model was successfully fit for the PK of both compounds and the PD simultaneously. The PK/PD model developed for ispinesib/docetaxel, may be used to examine different schedules, doses, and infusion times of both agents. Bayesian methods allow for the use of prior information available for the model parameters.