Dyania Health and Nvidia Deploy On-Premises AI for Cleveland Clinic Heart Records

Dyania Health and Nvidia have deployed an automated clinical data extraction system across more than 7,800 patient charts at Cleveland Clinic, running entirely on the hospital's internal servers, as reported by The Next Web.

The deployment addresses a persistent bottleneck in clinical informatics: most longitudinal patient data remains locked within unstructured clinical notes and scan narratives rather than structured database tables. By running Dyania's Synapsis AI platform directly on on-premises Nvidia hardware inside the hospital's network perimeter, the facility processed medical histories spanning up to 16 years without transferring confidential patient information outside its firewalls.

For the study, the system answered 88 specific clinical questions for each patient who underwent an aortic valve replacement between 2005 and 2021. The software automatically cross-references each answer back to the underlying medical chart, enabling clinicians to inspect the source documentation. Researchers used the resulting structured dataset to train a predictive model evaluating the risk of replacement valve failure at three-, five-, and 10-year intervals following surgery.

Nvidia claimed that an equivalent manual review across that volume of unstructured records would have taken clinicians years to complete. Lars Svensson, head of Cleveland Clinic's heart institute, described the initiative as "a promising step towards enhanced outcomes and more personalized care for patients."

Despite the operational speedup, the study established that the tool can assemble a research dataset and predict failure risk, not that it directly improves patient health outcomes. Dyania noted that the predictive model still requires independent validation across outside healthcare institutions before clinical adoption. Cleveland Clinic confirmed that human physicians continue to evaluate the AI-generated evidence and make all treatment decisions.

The clinical validation process will also face commercial scrutiny: Cleveland Clinic disclosed that it is an investor in Dyania Health, having taken an equity stake when first using the platform for clinical trial recruitment. While major tech companies and venture-backed startups are racing to deploy AI agents across hospital networks, health systems face persistent skepticism from clinicians demanding prospective proof of diagnostic utility and data governance before giving algorithms clinical authority.