Dyania Health's AI Reads Decades of Patient Records; Cleveland Clinic Study Covers 7,800 Valve Cases

NVIDIA Blog · rss · 2026-11-13

An NVIDIA blog post profiles Dyania Health, a New Jersey clinical AI company whose Synapsis AI is a clinical reasoning engine that works across every free-text note, operative report, imaging and pathology read, and discharge summary in a patient's record, linking findings years apart and outputting source-traceable clinical signals for care teams.

Data stays on-prem: Synapsis AI runs on NVIDIA-accelerated computing inside the health system's own firewall, so records are processed onsite and never leave the hospital. Models are trained with NVIDIA NeMo and optimized with TensorRT / TensorRT-LLM for the hospital's own GPU servers.
Assistive by design: clinicians still review the evidence and make every decision.
Real-world use: Cleveland Clinic used it in a study of more than 7,800 aortic valve replacement patients from 2005–2021, reasoning across each record (including yearly post-surgery echocardiograms) to build predictive models of valve deterioration and reintervention risk at 3, 5 and 10 years — chart review that would have taken clinicians years.
Broader applications: trial eligibility screening, closing cardiology and oncology care gaps, and populating clinical registries.

Original post →

Dyania Health's AI Reads Decades of Patient Records; Cleveland Clinic Study Covers 7,800 Valve Cases

More from Apps