Understanding and Streamlining Dose Finding: From Dose Simulation to Dose Estimation.

Abstract

Understanding drug dosing to fulfill safety and efficacy requirements in a patient population is an essential part of dose finding in clinical practice and drug development. The majority of current dose finding methods are simulation-based, which can be time consuming and resource intensive. Model-based simulations also do not guarantee that the dose, that will optimally fulfill the safety and efficacy endpoints, will be found. In this work, an advanced dose estimation approach based on the OptiDose concept is utilized to understand therapeutic doses in a patient population and to streamline dose finding. To demonstrate the potential of this concept, two illustrative case studies are presented, each representing specific challenges for dose finding, including different safety and efficacy requirements, complex dose-response relationships, and changes of dynamics in special populations. For both applications, the distributions of therapeutic doses in the population were estimated, such that therapeutic population doses can be determined for any proportion of patients to fulfill the safety and efficacy endpoints. In addition, the therapeutic population dose with which 50% and 95% of the population will fulfill the safety and efficacy requirements was estimated. Dose estimation for both drug applications was implemented in Monolix. The presented OptiDose approach has the potential to identify the "optimal" dose for any pharmacometric and clinical pharmacology scenario, allowing to guide clinical practice and facilitate dose selection in drug development, particularly in special populations such as pediatric or cancer patients. As such, we suggest moving from current "dose simulation" approaches toward a more efficient "dose estimation" paradigm.