PS38: Integrating Biomarker Discovery and Dose-optimization to Drive Drug Development Success

Clinical trials have a complex and lengthy development timeline. In the advent of targeted therapies and the FDA Project Optimus guideline, there has been an emphasis on the appropriate dosage for the "targeted patients". This had led to the pharmaceutical industry having to simultaneously optimize the biomarker and dose decision in earlier phases of drug development. Often, limited data is used to inform a decision from early phase trials to be taken forward to late phase trials; therefore, there has been a shift to draw from multiple data sources. Such direction has motivated drug developers to be innovative in statistical and design solutions while taking advantage of available data. However, it is important to be diligent by striking a balance between utilizing historical data while being true to our data, and ensure the external data resemble the indication and mechanism of action of interest for the current study. Industry strategies are evolving to address early decision-making for advancing to Phase 3 development by making the most of Phase 1/2 (single-arm treatment trials) data, with a focus on the biomarker and optimal dose. It is essential to promote efficient designs that can make better informed decisions going from early to late phase trials. Effective identification of the patient population that will benefit most from a treatment is essential. This area has grown rapidly with multiple modalities, advanced computational resources, and explosion of data. Because Phase 1/2 trials often have limited sample sizes and single-arm designs, adaptive designs and Bayesian approaches are warranted. Approaches to account for patient heterogeneity and subgroup analysis are needed to understand the biomarker and sources of noise to further facilitate the patient enrichment and dosing decisions. These novel approaches will improve design efficiency to ensure the drug will be targeted for the right patient population while maintaining safety. This session holds broad interest to both pharmaceutical practitioners and researchers in the biomarker and dose-optimization domains. This session will focus on various topics relevant for early phase decisions to inform late phase trials including adaptive designs, accounting for heterogeneity, optimal decision criteria, and subgroup approaches to identify the optimal dose and enriched population. This breadth of topics will lead to a fruitful session covering what statistical approaches and designs to consider for patient selection and dose selection. We will have 4 speakers on the following topics: Speaker 1: Dose optimization design accounting for unknown patient heterogeneity and biomarkers in cancer clinical trials Speaker 2: Bayesian Adaptive Design for Continuous Biomarker Cutoff Determination with Cross-Cohort Borrowing Speaker 3 A Bayesian adaptive enrichment design with historical control for phase II clinical trials Speaker 4 will discuss the regulatory perspectives on the statistical and design approaches to optimize dose and biomarkers simultaneously. Impact This session aligns with the RISW 2026 theme, "Impactful Transdisciplinary Decision-Making in the Evolving Digital Era," by showing how statistics, regulatory science, and clinical development intersect to leverage data for more informed choices that enhance future trial design.