Population Pharmacokinetics Analysis of Regorafenib and Its Active Metabolites From the Phase III Correct Study of Metastatic Colorectal Cancer

Abstract

Background: Regorafenib (REG) is an oral multikinase inhibitor that targets several kinases involved in tumor angiogenesis, oncogenesis, and tumor microenvironment signaling. The aim of this study was to develop a population pharmacokinetics (PK) model for REG and its two pharmacologically active metabolites M-2 and M-5, and to use this model for evaluation of covariate effects in the phase III CORRECT study. In the CORRECT study, REG demonstrated significant improvement in overall survival and progression-free survival vs placebo in patients with treatment-refractory metastatic colorectal cancer. Methods: A population PK model for REG, M-2, and M-5 was developed with NONMEM, based on PK data from the densely sampled, open-label, phase I dose-escalation study that investigated the safety and PK ofREG in cancer patients (study 11650). Subsequently, this model was applied to the sparsely sampled CORRECT study, by systematically estimating one or more parameters, while fixing the remaining parameters, until the best model was obtained. A model-based covariate analysis was performed to explore the impact ofpatient demographics (sex, weight, body mass index, height, age, ethnic group) and baseline parameters (kidney and liver function tests, protein levels, hematocrit/hemoglobin levels) on drug exposure in the CORRECT study. PK data from 67 densely sampled patients (study 11650) and from 381 sparsely sampled patients (CORRECT study), with concentration values for REG, M-2, and M-5, were used. Results: PK ofREG, M-2, and M-5 were described by a 2-compartmental model that provided an adequate fit of the plasma concentration–time profiles from both studies. Derived PK parameter estimates were used to calculate the individual exposure (Cavmd) ofREG, M-2, and M-5 in the CORRECT study. Overall, a moderate to high variability in the REG PK was observed in the CORRECT study, with a coefficient of variance of clearance of44% and even higher variability in exposure to M-2 and M-5. Covariate analysis of the CORRECT study showed that higher bilirubin levels at baseline were associated with higher individual exposure ofREG and M-2 (Cavmd –14 to +18% when comparing the 5th and 95th percentiles of the observed bilirubin distribution to the median baseline bilirubin levels; the overall variability was –48.7 to +79.9%). Higher body weight was associated with lower exposure ofM-2 and M-5, and exposure ofM-5 was higher in women (Cavmd 77%) than in men (the overall variability ofM-2 and M-5 were –69.8 to +149% and –89.2 to +348%, respectively). The magnitude of observed covariate effects was small compared with the overall high variability in exposure. The other tested covariates did not have an impact. Conclusion: The combined population PK model adequately described the PK profiles ofREG and its two pharmacologically active metabolites M-2 and M-5, in the phase I 11650 study and in the phase III CORRECT study. The effects of the covariates tested in population PK analysis of the CORRECT data were not considered clinically relevant in light of the overall high variability in exposure