Rare diseases affect more than 300 million people globally (that’s about 10% of the world’s total population) – yet the lack of reliable, high-quality data and research remains significant. At ISPOR 2025, experts gathered to examine how generative AI (GenAI) and large language models (LLMs) could help tackle the unique challenges of rare diseases during a panel titled Rare but Common: Generative AI’s Potential on Data, Evidence, and Insight Generation in Rare Diseases.
The conversation underscored the urgency of innovation in rare disease research – and the increasingly vital role GenAI plays in powering it.
Xiaoyan Wang, PhD, FAMIA, Chief Scientist and Senior Vice President, Life Science Solutions at IMO Health, joined panelists in Montreal, Canada, to explore how advanced AI is transforming rare disease research across three key areas:
- Data: Unlocking richer, more relevant rare disease data from electronic health records (EHRs) and publicly available sources through GenAI and LLMs
- Evidence: Creating rare disease-specific knowledge graphs to efficiently and accurately identify and synthesize real-world data (RWD) related to rare disease, including natural history and disease pathways
- Insights: Generating actionable intelligence to manage unmet needs and support earlier clinical trial recruitment
The conversation underscored the urgency of innovation in rare disease research – and the increasingly vital role GenAI plays in powering it.
IMO Health presents at ISPOR 2025
In addition to the panel, IMO Health’s AI Science team had the opportunity to present several recent AI-driven initiatives with applications across HEOR (health economics and outcomes research) and life sciences:
- May 14: AI-powered systematic literature reviews – dramatically reducing review time from months to days, with 0.993 accuracy
- May 14: Automated extraction of Kaplan-Meier survival curves – applying GenAI and computer vision
- May 15: AI-driven detection of CHAPLE disease – leveraging real-world clinical data to support earlier diagnosis
Closing the rare disease evidence gap
The panel also highlighted novel AI strategies reshaping what’s possible in HEOR:
- Real-world evidence generation using federated learning (a type of machine learning where the model is trained on data from multiple decentralized sources, such as hospitals and research centers)
- Digital twins (virtual representations of real patients) to mitigate recruitment challenges like high costs, patient risks, and lengthy timelines
- Public data and social listening to uncover more sources of evidence and data
For patients navigating rare diseases – where diagnosis timelines are long and clinical trials sparse – AI isn’t just a plus. It’s a necessity. The technology offers an unparalleled opportunity to overcome persistent hurdles in the rare disease space, generating data, evidence, and insights faster and smarter.