Rare Hope for Even Rarer Diseases: Innovation, Evidence, and Regulatory Perspectives
Moderator
Xiaoyan Wang, PhD, Tulane University, New Orleans, LA, United States
Speakers
Chunhua Weng, Columbia University, New York, NY, United States; Qi Liu, PhD, MStat, FCP, US Food and Drug Administration, Silver Spring, MD, United States; Sanket Shah, PhD, MD, Jazz Pharma, Gaithersburg, MD, United States
Purpose Recent advances in natural language processing (NLP) and large language models (LLMs) offer promising pathways to accelerate research, enhance phenotyping, and strengthen evidence packages for regulatory and HTA submissions. Yet the rapid adoption of AI technologies raises critical methodological, operational, and regulatory questions for rare disease drug development. This workshop aims to (1) clarify how emerging AI capabilities can support small-population evidence generation; (2) showcase real-world applications across academia, industry, and regulatory science; and (3) identify transparency, quality, and compliance considerations essential for integrating AI-generated evidence into regulatory and HTA decision-making. Description Rare diseases face persistent evidence gaps due to small patient populations, heterogeneous clinical presentations, and fragmented data sources. Advances in NLP and generative/agentic AI offer new opportunities to uncover real-world insights, automate evidence synthesis, and support more agile development pathways. The session will begin with context on the current AI landscape and the evolving methodological needs specific to small-population research. Dr. Wang (Tulane/NouStarX) will present her research on how NLP and LLMs support underserved and small population research, highlighting how large-scale clinical text analysis and federated learning can strengthen cohort identification and study readiness. Dr. Weng (Columbia) will present the use of fine-tuned LLMs for early diagnosis of rare diseases and genetic testing recommendations using RAG models and human phenotype ontology. Dr. Shah (Jazz Pharmaceuticals) will provide an industry perspective, including opportunities and constraints when integrating AI enabled workflows into value and market access strategies. Finally, Dr. Liu (FDA) will share a regulatory viewpoint on expectations for transparency, robustness, and methodological rigor when AI contributes to evidence packages in rare disease programs. The session concludes with a Q&A and digital polling designed to encourage audience interaction.
Topic
Clinical Outcomes, Health Policy & Regulatory, Methodological & Statistical Research