Unifying the odyssey: artificial intelligence for rare disease diagnosis and therapy.

Purpose: To summarize current challenges in rare disease (RD) diagnosis and therapy, highlight recent advances in artificial intelligence (AI) for RDs, and propose a model future state for RD patient care. Methods: Multidisciplinary expert-led narrative review summarizing modern practical challenges and rate-limiting steps in RD patient care, citing key clinical and research considerations with respect to regulatory and economic constraints. Results: Over 10,000 known RDs collectively affect 1 in 10 Americans, a total of over 30 million people. Annually, RDs account for over $1 trillion of annual US healthcare expenditures. Despite advances in genomic medicine, it takes 5-8 years on average to obtain an accurate diagnosis, and less than 5% of RDs currently have FDA-approved therapies. In this article, we review the history of RD diagnosis and current healthcare gaps underlying the major failures in patient care. Next, we will highlight emerging advances in genomic medicine and AI that are rapidly changing the RD landscape. Finally, we propose a target future state that integrates agentic AI for diagnosis and therapy with human-in-the-loop feedback. Conclusions: The rare disease diagnostic and therapeutic odyssey represents healthcare's most persistent failure mode. Ongoing challenges for clinical implementation involve biological modeling, manufacturing bottlenecks, and clinical trial design. We propose strategies for artificial intelligence to restructure the traditional sequence of diagnosis-then-therapy into a proactive orchestrated system delivering personalized cures at scale.