Boston Children's Hospital's Manton Center for Orphan Disease Research used OpenAI's o3 Deep Research model to reanalyze unsolved pediatric rare disease cases, surfacing new diagnostic leads for expert review. The target problem is concrete: roughly 50% of rare disease patients never receive a diagnosis, even after modern genetic testing.
The method matters as much as the outcome. Researchers are not using AI to replace clinicians but to comb through existing case data and generate hypotheses that specialists then evaluate. This distinction, AI as a research accelerator rather than a decision-maker, is the core of what makes this approach defensible and scalable in a clinical setting.
The full conversation reveals how the workflow was structured, what kinds of leads the model actually generated, and where the process broke down. If you work in genomics, rare disease research, or clinical AI deployment, the specifics of how o3 was integrated into an expert review pipeline are worth your time.
[WATCH ON YOUTUBE →]