How AI Is Helping Care Teams Get Patients the Right Care, Sooner

Health systems and startups are deploying AI across pediatric research, radiology, smart hospitals and home care to trim routine work from clinical schedules. NVIDIA says clinicians remain the decision-makers, reviewing AI outputs while its accelerated computing underpins the applications. One concrete case: Dyania Health, a clinical AI company based in New Jersey, built Synapsis AI to tackle one of healthcare's oldest bottlenecks — decades of patient history trapped in unstructured text.

Hospital records accumulate as physicians' notes, operative reports, imaging and pathology reads and discharge summaries, forcing clinicians to read through years of charts to reconstruct a single patient's story. Synapsis AI works as a clinical reasoning engine across every note, report and lab result, linking findings that may be separated by years to answer clinical questions the way a specialist reviewing the chart would. The output is a set of clinical signals, each traceable to its source in the record, indicating which patients need attention and why. Dyania positions it as assistive: clinicians review the evidence and make every decision.

Deployment runs on NVIDIA-accelerated computing inside a health system's own firewall, processing the entire record set onsite so patient data never leaves the hospital. Dyania builds and trains its clinical models with NVIDIA NeMo open libraries, then optimizes them with NVIDIA TensorRT and TensorRT-LLM to run faster and more efficiently on the hospital's own GPU servers. The stated benefit is that chart review comes off clinical teams' desks, freeing time for patient encounters — while outputs still require clinician confirmation before any action.

Cleveland Clinic put the platform to work in a study of more than 7,800 aortic valve replacement patients treated between 2005 and 2021. Synapsis AI reasoned across each patient's record, including every echocardiogram added in the years after surgery, to build a predictive model of valve deterioration and reintervention risk at the three-, five- and ten-year marks following the procedure. NVIDIA says the platform completed chart review across the full cohort in a fraction of the time such work would take clinicians by hand. Dr. Lars Svensson, chief of the Heart, Vascular & Thoracic Institute at Cleveland Clinic, called bringing together and analyzing information to predict the duration of implanted heart valves a promising step toward enhanced outcomes and more personalized care.

The open questions center on generalization: whether the retrospective modeling at Cleveland Clinic holds up prospectively across other procedures and institutions, and how care teams validate traceable-but-voluminous clinical signals in daily workflow. Also unresolved in this account is how such onsite GPU deployments scale cost across smaller health systems without in-house infrastructure, and where clinician review time actually lands when chart review volume grows alongside AI-generated signals.