PS32: Innovative Designs in Oncology: Adaptive Enrichment, AI/ML-Driven Borrowing, and Regulatory Perspectives

Innovative trial designs are reshaping oncology drug development by enabling more efficient studies, optimizing patient selection, and accelerating timelines while preserving scientific rigor. This session begins with a Bayesian AI/ML framework for creating "digital controls" that enable robust, privacy-preserving borrowing of external data to strengthen randomized controlled trials. Building on this, we explore an adaptive seamless Phase 2/3 subgroup enrichment design that leverages surrogate endpoints for expedited decision-making in heterogeneous patient populations. The discussion then turns to a regulatory perspective, highlighting considerations and challenges in evaluating and implementing such innovative designs in oncology. The session concludes with an integrated discussion synthesizing methodological innovation, applied strategies, and regulatory insights, offering attendees a multi-stakeholder view of how these approaches can support precision medicine and modernize oncology clinical trials.