Health professionals without prior ultrasound training successfully performed quality heart scans in mobile units traveling to rural U.S. communities, according to preliminary research presented at the American Heart Association's Scientific Sessions 2026. The findings suggest a promising strategy to extend echocardiography—ultrasound imaging of the heart—to underserved rural populations and enhance their participation in heart disease research.
Echocardiography provides live, moving images of the heart's structure and function. As the American Heart Association notes, it is commonly ordered for conditions such as heart murmurs, damaged valves, unexplained chest pain, congenital heart defects, or a history of heart attack or stroke. However, delivering this service in rural areas has been challenging due to a shortage of expert sonographers and cardiovascular imaging specialists. "The conventional model depends on an expert sonographer working in a hospital setting or imaging center, which can be limited in rural areas," said study author Sanjana Ravi, M.D., an internal medicine and pediatrics resident at Duke University School of Medicine.
The study, part of the larger RURAL (Risk Underlying Rural Areas Longitudinal) study, involved 4,597 volunteers aged 25-65. Fourteen non-sonographers—nurses, nursing assistants, and CT technologists—completed a two-hour training on AI-guided ultrasound. They then performed scans in a mobile health unit that visited 10 rural counties across Alabama, Kentucky, Louisiana, and Mississippi between November 2021 and June 2026. The AI software provided real-time guidance, and images were reviewed remotely by experts at the Duke Clinical Research Institute Imaging Core Laboratory.
Results showed that 98% of the scans included all six standard echocardiography views required by the study protocol, and more than 99% were rated adequate for measurements by core lab experts. Diagnostic-quality images for individual views ranged from 77% to 94%. Scans were completed within the 15-minute target window 81% of the time, with a median time of 9.5 minutes. Critical findings were identified in 133 scans, most commonly an ejection fraction below 50%, indicating reduced heart pumping function. Of these, 99% were returned to participants within three days.
"We were surprised at how quickly and well the nurses and CT technologists who had never done a heart ultrasound were able to adapt to this technology and produce quality images for review by experts," Ravi said. Lee H. Schwamm, M.D., FAHA, a member of the American Heart Association's volunteer committee on AI and Technology Innovation, noted that "1 in 5 Americans lives in a rural area, which limits their access to critical diagnostic testing and specialty care. In this hybrid model, community-based non-experts were able to create a human-AI partnership with the acquisition software to enable the capture of echocardiogram images that could be read by off-site experts." Schwamm, who was not involved in the study, added that if confirmed in further studies, this approach could serve as a model for other specialties and diagnostic modalities. He also pointed to the Association's New guidance offered for responsible AI use in health care as relevant context for such innovations.
The study has limitations, including no direct comparison between sonographer-acquired and AI-guided non-sonographer images, use of only one software brand, and involvement of a single core laboratory. Results may not be generalizable to non-rural populations or settings outside core-lab-supported research. The findings are considered preliminary until published in a peer-reviewed journal. The research will be presented at Scientific Sessions 2026, and additional details are available in the abstract poster MDP2290.
