Digital Twins for Accelerating Clinical Trial Innovation in Alcohol-Associated Hepatitis

About This Grant

PROJECT SUMMARY Alcohol-associated hepatitis (AH) is a severe form of liver disease and a leading cause of liver-related mortality in the U.S., yet the therapeutic landscape remains critically underserved. Clinical trials for new AH therapies are profoundly impaired by low patient enrollment and poor retention, creating a major bottleneck that has stalled therapeutic advancement for decades. To address these challenges, this proposal leverages an innovative approach using state-of-the-art artificial intelligence (AI) to create "digital twins" – virtual patient representations that can simulate disease trajectories and treatment response. Our central hypothesis is that digital twins, generated via advanced diffusion models, can effectively emulate randomized clinical trial results and simulate control group responses, enabling in-silico trial designs that reduce enrollment requirements while maintaining statistical rigor and privacy. Our strong preliminary data support this hypothesis, demonstrating the successful generation of high-fidelity synthetic hepatology datasets that preserve critical clinical relationships, survival outcomes, and patient privacy. Furthermore, we have successfully executed a full in-silico trial using digital twins, providing a robust proof-of-concept for our proposed methods. Building on this foundation, this project will pursue two specific aims: • Aim 1: Develop digital twins from a comprehensive, multi-institutional AH dataset and perform a target trial emulation of the landmark STOPAH study to validate the framework's ability to reproduce the findings of a large-scale randomized controlled trial. • Aim 2: Validate the digital twins for accurately simulating control group responses in both a completed (retrospective) and an ongoing (prospective) AH clinical trial conducted by the NIH-funded AlcHepNet consortium. This research is innovative as it represents the first systematic application of digital twins to overcome long- standing clinical trial barriers in AH. Successful completion will provide a validated platform to accelerate therapeutic development, reduce trial costs, and inform regulatory science for AI-driven clinical research. The institutional environment at Mayo Clinic provides unparalleled mentorship, access to extensive clinical datasets, and computational resources. This K08 mentored career development award is essential for my transition into an independent physician-scientist pioneering AI-driven precision medicine to improve outcomes for patients with alcohol-associated liver disease.