← Back to Home

Bringing Echocardiography to Rural Communities: Study of 4,597 People Shows Feasibility of AI Guidance Paired with Remote Experts

Can healthcare staff without ultrasound experience obtain usable images at mobile examination sites? A rural U.S. study offers preliminary findings while highlighting the continued importance of expert quality control and follow-up care.

By SURL BioNews

For rural residents, barriers to heart examinations sometimes include the distance to an imaging center and a shortage of local specialists. A U.S. study involving 4,597 adults attempted to bring examinations into communities: healthcare staff without ultrasound experience acquired cardiac images with real-time artificial intelligence (AI) guidance, then sent them to experts elsewhere for review. The results suggest that this division of responsibilities could expand cardiac imaging services in rural research settings.

The analysis comes from the RURAL cohort study. The study’s official website lists Sanjana Ravi of Duke University’s imaging core team as the first author of the related abstract, which focuses on AI-guided echocardiography and core laboratory quality assurance in mobile rural settings. The American Heart Association released the findings on November 2, and the study is scheduled for presentation at its 2026 Scientific Sessions on November 6–9; it has not yet undergone journal peer review as a full paper.

Echocardiography uses sound waves to show the heart’s structure and movement. In this study, AI helped staff on site acquire images, while experts remained responsible for interpretation and quality control. Fourteen nurses, nursing assistants, and computed tomography technologists received two hours of training, with three requiring additional training after review of their initial images. Senior sonographers and physicians at the Duke Clinical Research Institute’s core laboratory reviewed the acquired images.

In the study, 98% of examinations obtained all six views required by the protocol, and experts rated more than 99% as adequate for measurements. However, being suitable for measurement and achieving diagnostic quality in every view are different thresholds: the proportion meeting diagnostic quality standards ranged from 77% to 94% across views. These findings therefore support the feasibility of the overall workflow for obtaining images needed for research; they do not yet justify claiming that its diagnostic performance is equivalent to that of professional sonographers.

Efficiency and reporting results also tested the workflow. The median image acquisition time was 9.5 minutes, and 81% were completed within the allotted 15 minutes. The study also identified significant abnormalities requiring prompt reporting in 133 examinations, with 99% of those results communicated to participants within three days. The same American Heart Association report published by Medical Xpress also lists these figures, but it is a republication from the same source, rather than independent validation by another study.

Several key questions remain. The team did not directly compare the images with those acquired by professional sonographers, nor did it systematically assess differences when different operators scanned the same person. The study used only one software brand and one core laboratory, so performance with other equipment or in settings without equivalent expert support still requires separate validation. In particular, these participants were volunteers undergoing research examinations, and the findings cannot yet be assumed to extend to routine clinical services.

The study provides a more concrete picture of one possible approach to rural imaging services: trained staff on site, AI assistance with image acquisition, and remote experts maintaining quality and reporting abnormalities. However, the study did not assess whether delivering results to residents led to medical visits, treatment, or improvements in health. Turning access to examinations into benefits from care will require tracking that subsequent course.

References

  1. American Heart Association
  2. The RURAL Cohort Study
  3. Medical Xpress