PS25: Designing Biomarker-Adaptive Trials: Phase II Innovation to Phase III Confirmation

Biomarker-adaptive clinical trial designs are reshaping how evidence is generated for targeted and precision therapies, particularly in oncology and other therapeutic areas characterized by substantial biological heterogeneity. By explicitly integrating biomarker information into the design, conduct, and analysis of clinical trials, these approaches aim to improve development efficiency, increase the probability of technical and regulatory success, and better align clinical programs with mechanisms of action and patient benefit. At the same time, biomarker-adaptive designs introduce important statistical and operational challenges. These include control of multiplicity across biomarker-defined hypotheses, robustness to biomarker misclassification and uncertain prevalence, interpretability of adaptive decisions, and the need for regulatory-acceptable decision and analysis frameworks that support labeling claims. This parallel session will highlight recent statistical advances of biomarker-adaptive designs across the clinical development continuum. Topics will include rigorous Phase II signal-finding designs, confirmatory Phase III strategies that allow dual development paths for all-comer and biomarker-defined populations, and a regulatory perspective on evidentiary standards, multiplicity, and decision-making frameworks. Collectively, the session will emphasize how principled statistical methodology enables impactful, transdisciplinary decision-making in modern precision medicine programs. The session is structured to follow the typical development pathway: beginning with Phase II optimization of biomarker-guided designs, progressing to confirmatory Phase III trials with adaptive population-level decision rules, and concluding with a regulatory perspective informed by recent submissions and interactions. Impact and Intended Audience This session directly supports the RISW 2026 theme of "Impactful Transdisciplinary Decision-making in the Evolving Digital Era." It brings together statistical methodology and regulatory science to illustrate how biomarker-adaptive designs can be responsibly implemented to guide high-stakes development decisions. The intended audience includes clinical statisticians and regulatory scientists working in oncology and other precision-medicine areas, as well as academic researchers interested in adaptive design methodology. Attendees will gain insight into state-of-the-art Phase II and Phase III biomarker-guided designs with rigorous multiplicity control, and direct regulatory perspectives on evidentiary standards and decision-making.