BOIN12-PFS: A Bayesian Optimal Interval Phase I/II Design Incorporating Progression-Free Survival Endpoint

Modern oncology drug development is increasingly focused on identifying the optimal biological dose (OBD) rather than just the maximum tolerated dose (MTD). Early-phase oncology trials have conventionally relied on short-term endpoints, such as overall response rate (ORR), to evaluate efficacy. However, there is a growing perception that incorporating longer-term endpoints, such as progression-free survival (PFS), into early-phase dose-finding trials can provide more informative guidance for dose optimization. This study proposes a new design, BOIN12-PFS, to identify the optimal dose based on PFS, short-term efficacy, and toxicity. As an extension of the BOIN12 design, BOIN12-PFS is expected to be comparatively straightforward to implement in clinical oncology dose-finding trials. Simulation studies demonstrate that the BOIN12-PFS design offers advantages over existing designs in terms of the probability of selecting the OBD, the average number of patients allocated to the OBD, reducing overdosing, and requiring a smaller total sample size across various realistic settings.