Meta to fund radiology AI research at Ohio State
Teams of experts from Ohio State are partnering with Meta to advance research into the use of artificial intelligence in radiology.
Three researchers will be using models from Meta Superintelligence Labs to explore new approaches to chest X-ray reporting, early heart failure risk prediction and Alzheimer’s disease diagnosis. They represent three of five teams receiving funding and access to Meta’s Muse AI models, which will support projects aimed at improving clinical decision-making and making better use of information already available in patients’ medical records.
One project, led by Dufan Wu, PhD, associate professor of radiology at Ohio State’s College of Medicine, will investigate whether AI can draft chest X-ray reports using imaging and electronic health record data. The system will incorporate clinical context, including the reason for the examination, patient history and prior imaging to more closely mimic how radiologists interpret studies.
For that same project, Wu and Peter Kvam, PhD, an associate professor in Ohio State’s Department of Psychology, also will look at how radiologists interact with AI-generated reports in a simulated clinical setting. Researchers will examine when clinicians reject correct AI suggestions or accept incorrect information, with the goal of identifying ways to improve the safe and effective use of AI in radiology.
A second project will evaluate the potential of opportunistic heart failure risk assessments using routine CT scans. Experts will use the tools to help identify patients at risk of developing heart failure with preserved ejection fraction, which can be difficult to recognize early due to its overlapping symptoms.
Yuan Xue, PhD, Yuchi Han, MD, and Xia Ning, PhD, will lead that project. The team will combine CT-derived measures of body fat distribution, liver fat, muscle composition and vascular calcification with medical records and electrocardiogram data when available. The model will be designed to generate risk estimates and summaries of the evidence supporting its predictions, even with missing patient information.
The third project will focus on Alzheimer’s disease and related dementias. Ning, Xue and Douglas Scharre, MD, are developing a clinician-guided, multi-agent AI system to analyze the trajectory of patients’ cognitive histories, evaluate clinical evidence, identify diagnostic uncertainties and recommend next steps in the diagnostic process. The system will use Meta’s open-weight Muse models as a locally deployable foundation and draw on curated records from Ohio State’s Memory Disorders Clinic, which handles more than 5,600 patient visits annually. Clinicians will remain central to decision-making.
These projects will be funded by a grant from Meta worth approximately $2 million.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.