Turn on cookies to sign in
Signing in saves your progress to your Curie account. We can only do that with cookies on — turn them on to continue.
Dyania Health and Nvidia are teaming up to run clinical reasoning AI directly inside hospital networks, keeping sensitive patient data behind the firewall. Their Synapsis AI platform reads full medical records — notes, labs, imaging — and answers complex clinical questions with evidence-backed answers. A real-world test at Cleveland Clinic analyzed nearly 8,000 heart valve patients across 16 years, building a predictive model for valve deterioration and repeat procedures.
Dyania Health and Nvidia are joining forces to deploy Synapsis AI — a clinical reasoning platform — directly inside hospital IT systems, so patient data never has to leave the building. Unlike general-purpose AI tools, Synapsis reads a patient's full longitudinal record (clinical notes, imaging reports, labs, operative summaries) and answers specific clinical questions while citing supporting evidence from the chart. The system runs on Nvidia's accelerated computing infrastructure on-premises, avoiding cloud transfers and reducing cybersecurity and privacy risks.
The partnership's real-world proof of concept comes from Cleveland Clinic, where Synapsis AI analyzed up to 16 years of records from nearly 8,000 aortic valve replacement patients — a task that would have been "humanly impossible" through manual chart review. The AI answered 88 clinical questions per patient and helped build a predictive model estimating valve deterioration and re-intervention risk at 3, 5, and 10 years post-surgery.
By the Numbers:
Why it matters: Health systems are sitting on mountains of unstructured clinical data they can't easily use. This partnership offers a privacy-preserving, scalable way to turn that data into actionable insights — from personalized care decisions to clinical trial recruitment — without outsourcing sensitive records to external AI services.
