Med AI

The University of Virginia School of Medicine has been advancing artificial intelligence (AI) in healthcare since the early 2000s. From developing early warning systems that detect sepsis in premature infants to designing personalized approaches for managing complex diseases, our researchers are creating tools that measurably improve outcomes across the full continuum of care. Many of these UVA innovations are now in active use in hospitals throughout the United States and in 27 countries worldwide.

What sets UVA apart is our unwavering focus on clinical implementation. We do more than build sophisticated algorithms—we embed them into real-world hospital workflows, partner closely with frontline clinicians, and rigorously evaluate their impact on patient safety, quality, and outcomes.

CAMA (Center for Advanced Medical Analytics)

CAMA is a multidisciplinary UVA group using big data and AI to improve clinical outcomes and solve health system challenges. Some of their notable projects include:

HeRO (Heart Rate Observation System): Developed at UVA, this machine learning system predicts sepsis in preterm infants using heart rate variability. FDA-cleared in 2002, validated in multi-center clinical trials, and deployed in the U.S. and 27 other countries. HeRO paved the way for UVA Health to become a pioneer in predictive analytics.
CoMET (Continuous Monitoring of Event Trajectories) was developed at UVA by Dr. Randall Moorman, CoMET predicts patient deterioration (e.g., sepsis, respiratory failure) up to 12 hours in advance using real-time physiological data. The software was implemented on various floors at UVA Health, including the Medical Intensive Care Unit (MICU), Critical Care Unit (CCU), and others, starting in 2020.

Read about CAMA’s clinical AI applications

Cardiology and Endocrinology

CARNA (Characterizing Advanced Heart Failure Risk and Hemodynamic Phenotypes) is a free, publicly available AI risk assessment tool that helps predict heart failure risk for individual patients. It was developed by School of Medicine physicians Drs. Sula Mazimba and Kenneth Bilchick in collaboration with computer scientists Josephine Lamp and Yuxin Wu. The researchers have made their new tool available online for free at https://github.com/jozieLamp/CARNA.
Researchers at the UVA Center for Diabetes Technology, led by Dr. Boris Kovatchev, have developed an AI-powered artificial pancreas that automatically regulates insulin delivery for Type I diabetes without requiring patient input on meal timing or size. The system is currently being tested in a broader cohort across the United States, with plans for a larger study to further evaluate its clinical impact.
AI for Sudden Cardiac Death Risk: UVA researchers, including Kenneth Bilchick, are applying deep learning to cardiac MRI (CMR) images and ECGs to identify patients at high risk of sudden cardiac death, including those with mild or no known heart dysfunction, with the goal of improving early detection and prevention where current approaches may fall short.
Critical Care & Respiratory Monitoring

Respiratory Monitoring Projects (ARC study): Led by Shrirang Gadrey, MBBS, MPH, UVA researchers developed ARK (Analysis of Respiratory Kinematics), a wearable sensor system that quantifies labored breathing with high fidelity. Supported by a $4M NIH grant, the project aims to improve early detection of respiratory deterioration in adults and premature neonates, enabling timely intervention and advancing predictive analytics in critical care.

Emergency Medicine

George Glass, MD is conducting research on wearable AI devices designed to enhance patient monitoring in the emergency department. One example is a caregiver-operated finger probe that continuously tracks blood pressure and oxygen levels, helping clinicians detect and predict patient decompensation. These wearable technologies also have the potential to strengthen and streamline ED triage.

Neurology

Explainable AI (XAI): UVA researchers at School of Data Science are developing models for neurodevelopmental disorders (autism, ADHD, OCD). XAI helps doctors with explanations for an AI’s diagnoses to build trust and make informed decisions.
Targeting STING (Stimulator of Interferon Genes) to Combat Neurodegeneration: Led by John Lukens at UVA’s Harrison Family Translational Research Center, researchers are investigating how the immune molecule STING contributes to Alzheimer’s and other neurodegenerative diseases. Blocking STING in lab mice reduced plaque buildup, neuronal damage, and memory loss, highlighting a novel pathway for potential disease-modifying treatments.

Oncology

Brain Imaging AI: UVA researchers, including Dr. Bijoy Kundu and the UVA Cancer Center and Department of Biomedical Engineering team, are using AI-enhanced MRI and PET imaging to improve brain cancer care, particularly for glioblastoma. The project leverages Siemens’ advanced imaging technologies (e.g., Deep Resolve Swift Brain protocol) and AI algorithms to distinguish tumor progression from treatment-related changes with high accuracy.
Research by J. Kim Penberthy, David Penberthy, and Jennifer Bires is exploring how AI-powered tools, including virtual counselors, chatbots, telepsychiatry platforms, and wearable devices, can help detect mental health challenges in breast cancer patients, enabling earlier intervention and more scalable, personalized support.

Orthopedics

Springbok Analytics, developed by UVA professors Silvia Blemker, Craig Meyer, and Joe Hart, transforms MRI scans into color-coded 3D renderings of muscle volume to evaluate injury, recovery, and performance. Originating from UVA research, the technology has grown into a commercial company that leverages AI and machine learning to accelerate analysis, guide rehabilitation strategies, and support applications in sports medicine, clinical care, and neuromuscular disorders.

Artificial Intelligence at the School of Medicine