#14. Global Trial & Error: Can AI Help? (05/30/2025)

In this episode, we delve deep into the recent FDA Oncologic Drugs Advisory Committee (ODAC) decision against expanding glofitamab (Columvi) for relapsed/refractory diffuse large B-cell lymphoma (DLBCL), despite overall positive Phase III STARGLO trial results, highlighting critical challenges in global clinical trial design. Stark regional efficacy disparities, particularly between Asian and non-Asian cohorts, underscored the limitations of current methodologies in addressing geographic heterogeneity. This report delves into the potential scientific underpinnings of these disparities, including molecular variations in DLBCL subtypes (e.g., ABC vs. GCB prevalence), pharmacokinetic factors influenced by patient characteristics like BMI, and inconsistencies in chemotherapy backbone administration across trial sites. It further explores the urgent need for advanced computational approaches to overcome these challenges. Emerging artificial intelligence (AI) technologies offer transformative solutions, such as federated learning for enhanced diverse patient recruitment, AI-generated synthetic control arms for regional validation, multi-omic integration for predictive biomarker discovery, AI-driven adaptive trial designs, blockchain for data integrity, and virtual patient simulations. The report emphasizes that integrating these AI-driven tools is crucial for developing therapies with demonstrated efficacy across diverse populations, aligning with regulatory expectations for robust, generalizable evidence in the era of precision oncology. Produced by Dr. Jake Chen.