Population pharmacokinetic and pharmacodynamic modeling and simulation of the investigational anticancer agent indisulam

Abstract

Indisulam is an investigational anticancer agent that is currently being evaluated in phase II clinical studies. The aim of this thesis was to develop a mechanism-based pharmacokinetic and pharmacodynamic model for indisulam-induced myelosuppression and to apply this model as a tool for treatment optimization. This may assist further clinical development of this drug. Indisulam has a highly non-linear concentration-time profile, indicating complex underlying pharmacokinetic processes. The profound non-linearity and the wide variability between patients contributed to poorly predictable drug exposure after treatment with indisulam. A semi-physiological population pharmacokinetic model of indisulam was developed to gain further knowledge and understanding of the disposition of indisulam. Neutropenia and thrombocytopenia were identified as the dose limiting toxicities of indisulam. A subset of patients developed grade 4 neutropenia, the most severe grade according to the Common Toxicity Criteria (CTC) and were at risk of developing serious and rapidly progressing infections. A semi-physiological population pharmacokinetic and pharmacodynamic model was used to describe the time profile of neutropenia and thrombocytopenia after administration of indisulam. Treatment with indisulam was complicated, mainly because the severity of indisulam-induced haematological toxicity was poorly predictable. An extensive covariate analysis was performed to identify pharmacokinetic and pharmacodynamic determinants of indisulam. Low body surface area, Japanese race, variant CYP2C genotype, low baseline neutrophil and thrombocyte counts and female sex were identified as clinically relevant risk factors. An algorithm for dose individualization was developed, which may assist in safe dosing of indisulam. Indisulam is currently evaluated in combination regimens. Indisulam in combination with capecitabine was well tolerated in the first treatment cycle, but severe myelotoxicity was observed after the second treatment cycle. A time-dependent drug-drug interaction was identified by population PK-PD modeling. The combination of 550 mg/m2 indisulam and 1250 mg/m2 capecitabine BID was predicted to be safe and feasible for future studies. The combination of indisulam and carboplatin was not tolerable in a 3-weekly regimen, because a delay of re-treatment was frequently required to allow recovery from myelosuppression from previous cycles. A PK-PD model was developed and used to evaluate 3-weekly and 4-weekly administration regimens of various doses of the combination indisulam-carboplatin. This study supported the selection of 500 mg/m2 indisulam in combination with 6 mg.min/ml carboplatin in a 4-weekly regimen as the recommended dose for future studies. A safe dose of indisulam in various administration schedules was previously defined in four parallel phase I dose escalation studies. We proposed a new clinical trial design to increase the efficiency of a phase I program. The method proved to be equally safe as the conventional design, the number of patients treated at a dose level below the recommended dose was reduced and the recommended dose for further evaluation could be precisely determined. This demonstrates that population PK-PD modeling and simulation may support optimization of clinical trial designs. Population PK-PD modeling and simulation has contributed to a better understanding of the clinical pharmacology of indisulam. Dose individualization may improve treatment with indisulam. Further clinical development may be supported by optimization of combination chemotherapy.