Integrated quantitative pharmacology for treatment optimization in oncology

This thesis describes the development and application of quantitative pharmacological models in oncology for treatment optimization and for the design and analysis of clinical trials with respect to pharmacokinetics, toxicity, efficacy and cost-effectiveness. A recurring theme throughout this thesis is the quantitative integration of available knowledge during the development of a model. The utility of such integrative approaches relates to the fact that usually some relevant pre-existing knowledge exists, for instance related to physiology, pharmacokinetics or pharmacodynamics. Across different chapters in this thesis we have studied how such knowledge can be integrated in models to increase their predictive value or clinical utility. A second motivation for integrative approaches is the multi-factorial character of treatment of individual patients, or the design of clinical trials. The first part of this thesis demonstrated the development of model-based approaches for the design and analysis of studies in special patient populations including children, pregnant patients and patients with renal impairment. A clinical trial simulation was conducted for a drug-drug interaction study in pediatric oncology patients. Subsequently we describe the development of a model quantifying changes in pharmacokinetics for four frequently used anticancer agents in pregnant cancer patients. Finally we showed how integration of established physiological changes during pregnancy can be used to support the prediction and analysis of pharmacokinetic studies in pregnant patients. In the next part we developed quantitative dynamical models for toxicities. A semi-physiological model for eribulin-induced neutropenia was developed and applied to derive improved dosing strategies. A second toxicity model described trastuzumab-induced cardiotoxicity, which was then used to improve cardiac monitoring protocols. In the last part of this thesis we first describe the development of a disease-progression model for castration-resistant prostate cancer in patients treated with eribulin, and the subsequent integration of models for toxicity, efficacy and cost-effectiveness that can be potentially used to support decision-making during early drug development. The last chapters describe the development of a generic simulation framework for cost-effectiveness models and an evaluation of the impact of structural uncertainty in such models