Researchers at Boston Children’s Hospital and Harvard utilized OpenAI’s o3 Deep Research model to identify 18 new diagnoses among children with previously unsolved rare genetic diseases, according to a study published in NEJM AI. The research involved reanalyzing 376 de-identified pediatric cases that had undergone genetic testing and expert review, ultimately yielding diagnoses across multiple conditions including neurodevelopmental disorders, rare neuromuscular diseases, sudden unexpected death in pediatrics, and early-onset psychosis.
The o3 Deep Research model synthesizes genetic data, clinical presentations, and medical literature to aid clinicians and researchers rather than serve as a direct consumer diagnostic tool. John Brownstein, Chief Innovation Officer at Boston Children’s Hospital, noted the challenge lies in cognitive limits rather than effort. “We combine genetic information, phenotypic information, literature search, and the reasoning of AI to deliver diagnoses to families that were once left without any answers,” Brownstein said.
The NEJM AI publication represents a significant advancement in a broader initiative at Boston Children’s Hospital, which announced in May that its AI efforts had already resulted in over 40 rare disease diagnoses previously deemed unsolvable. The collaboration with OpenAI commenced in early 2025 and received $50 million in funding to support these initiatives.
Boston Children’s Hospital has integrated AI across its operations, with over one-third of employees using AI tools daily. The hospital estimates it has saved approximately 60,000 hours due to AI-enabled workflows, translating to more than $7 million in labor savings. Rare diseases affect an estimated 300 million people globally, and the challenge of achieving timely diagnoses persists, often resulting in families enduring long “diagnostic odysseys” before receiving answers.
