Optimizing timelines of phase I dose finding studies through cohort size simulations

6568 Background: A challenge of phase I studies is to select a dose for phase II in a timely manner, while ensuring patient's safety. There is also a desire to treat more patients at potentially efficacious doses. There are many obstacles to optimal study delivery: low starting dose; recruitment suspension while a cohort is in observation; need to replace non-evaluable (NE) patients; need to expand an initially small cohort to clarify safety signals. We have conducted simulation exercises quantifying: time for dose decision based on chance of a dose-limiting toxicity (DLT) and a patient being NE; the expected number of DLTs at a dose level. Methods: Consider a dose finding study design based on DLTs. Initially, 1 to 6 patients are enrolled at a dose level and observed for a pre-specified period of time. Depending on the number of DLTs, the next step is: (1 DLT) expand to 6 patients; (0 or 2 or more DLTs) explore another dose level or stop. In this presentation, we set observation period = 28 days; time to...