Model-informed dose selection in early oncology drug development

Abstract

Drug developers in oncology indication has been struggling with a low probability of success for decades. In my thesis, we studied how to improve the clinical proof-of-concept success rate with quantitative dosing rationale. We reviewed the classical dose selection strategies and the emerging transition of mathematical model-informed optimization strategies in anti-cancer drug development. Then we presented two case studies: first-in-human dose selection of avelumab, a monoclonal antibody, and the recommended phase II dose determination of tepotinib, a small-molecule kinase inhibitor. It showed how quantitative tools informed the dose selection with precision, and streamlined the early drug development path by avoiding unnecessary dose steps. The accumulating clinical data also proved, in both cases, successful dose strategies. By integrating preclinical in-vitro/in-vivo and clinical data across multiple development stages, the quantitative frameworks interpret the accumulating knowledge with great synergy, showing a high potential of model-informed optimization strategy to improve efficiency in drug development.