Dyania Health today announced a collaboration with NVIDIA to put AI clinical reasoning to work inside health systems, on their own infrastructure, at the scale of their entire patient population. Dyania's Synapsis AI reads a patient's full medical record, across years of notes, operative reports, imaging reports and lab results, and answers clinical questions with evidence from the chart. NVIDIA accelerated computing makes it possible to do that across large patient populations behind the hospital firewall, without transferring patient records outside the health system. Together, the companies aim to give health systems the ability to learn from the full histories of the patients they have treated and turn that experience into earlier identification of care needs, more informed treatment decisions and better care for the next patient, while keeping those records within their own systems.
Turning clinical experience into better care
Health systems hold decades of clinical experience in their medical records, but much of it is difficult to analyze at scale. Much of the information that matters is in narrative text rather than structured fields, and interpreting it often requires manual review. Registries, outcomes studies and predictive models can be limited by how many charts a team can read by hand. Extracting and interpreting the clinical information needed to build them remains a major bottleneck.
Synapsis AI performs that reasoning. It is built to work the way a specialist works through a chart: longitudinally, across every document, resolving what a patient actually had, when, and with what evidence. Doing this for one patient is a reasoning problem. Doing it for an entire health system, continuously, inside the hospital's own environment, requires substantial computing capacity. Dyania brings the clinical reasoning engine and implementation experience; NVIDIA brings the accelerated computing platform and the software that lets it run securely and efficiently on premises.
Synapsis AI is built on the NVIDIA AI Enterprise software stack end to end. Dyania trains its proprietary clinical models with the NVIDIA NeMo framework for model training and reinforcement learning. Inside the health system, Synapsis AI runs on NVIDIA TensorRT and TensorRT-LLM on the hospital's own NVIDIA accelerated computing infrastructure, so inference stays inside the hospital's private network.
Under the collaboration, the companies will work to bring Synapsis AI on NVIDIA infrastructure to health systems in the U.S. and internationally, publish joint case studies, and integrate Synapsis AI with NVIDIA's healthcare software stack. Areas the companies are evaluating include NVIDIA Confidential Computing for training and inference, NVIDIA Dynamo for distributed inference serving, Multi-Instance GPU for partitioning GPUs across concurrent clinical workloads, and scaling to many NVIDIA Blackwell Ultra GPUs as deployments move from batch processing toward near-real-time clinical insight at the point of care.
"The opportunity for AI in healthcare is to give care teams more time and insight to focus on patients," said Chelsea Sumner, Healthcare AI Startups Ecosystem Lead at NVIDIA. "Dyania Health brings deep clinical expertise to that opportunity, using NVIDIA's AI platform to support research, help identify gaps in care and build a more complete understanding of patients' needs."
"This is a watershed moment for patient care. Every patient's story is spread across years of notes, reports and results, and for the first time we can reason across all of it at once, the way a specialist does, for every patient in a health system rather than one chart at a time," said Eirini Schlosser, founder and CEO of Dyania Health. "Synapsis AI reads the full record, draws the clinical conclusions and shows the evidence behind them, and NVIDIA gives us the computing power to do that inside the hospital, with every record staying where it belongs. What we have shown with Cleveland Clinic is that a health system can learn from the patients it has already treated to give the next patient the right care sooner."
Predicting how long replacement heart valves will last
The first work from the collaboration is a Cleveland Clinic study of 7,812 patients who received an aortic valve replacement between 2005 and 2021. The study addressed a central concern for patients: How long will the replacement valve last, and will another procedure be needed? To answer this, researchers reviewed up to 16 years of patient care, most of it in narrative notes and echocardiogram reports that can be challenging to capture in registries.
Running on NVIDIA accelerated computing inside Cleveland Clinic's environment, Synapsis AI identified the patient population and answered 88 clinical questions for each patient, covering the valve and its characteristics, co-existing conditions, laboratory trends, medications, follow-up ultrasound measurements and later clinical events. The study produced a detailed clinical dataset in far less time than a manual review would take. Additionally, a predictive machine-learning model was trained that is capable of predicting valve failure and reintervention risk timelines for every patient.
The model found that the type of valve implanted was the strongest determinant of how quickly it deteriorated, while age and other medical conditions shaped whether and when a second procedure was performed. The analysis also examined how follow-up information, such as heart ultrasounds following surgery, can substantially raise predictive accuracy around future interventions. The model estimates an individual patient's risk at three, five and 10 years and revises the outlook with each new ultrasound. The tool is used to help inform care but requires external validation from clinicians.
"Using technology to bring together and analyze information that can help us predict the duration of implanted heart valves is a promising step towards enhanced outcomes and more personalized care for patients," said Lars G. Svensson, MD, PhD, Chief, Heart, Vascular & Thoracic Institute at Cleveland Clinic.
The heart valve study illustrates the broader potential of the collaboration. The same technology could help health systems learn from their own experience to understand which patients would benefit from certain therapies, which treatments hold up over time, and why outcomes differ between patients who look alike on paper. Looking ahead, Synapsis AI agents will continuously train predictive models of this kind, so that recommendations for the next recommended action in a patient's care keeps improving in accuracy as precision medicine advances.
ACS Emergency General Surgery Verification associated with fewer deaths and shorter stays