PHOENIX — Mayo Clinic researchers have developed and externally validated an artificial intelligence (AI) model that can identify a potentially significant heart obstruction from routine ultrasound videos without relying on specialized Doppler imaging . The technology could help clinicians flag patients with hypertrophic cardiomyopathy (HCM) who may need additional testing, particularly in settings where specialized echocardiography expertise is limited. Study findings are published in Circulation: Cardiovascular Imaging.
HCM is a genetic condition that causes the heart muscle to become abnormally thick. About two-thirds of these patients develop left ventricular outflow tract (LVOT) obstruction , which restricts blood leaving the heart, causing symptoms such as chest pain and shortness of breath with exertion or when lying flat. Knowing which patients develop LVOT obstruction is important because it influences treatment decisions and long-term management for patients with HCM.
"Measuring LVOT obstruction typically requires Doppler echocardiography, which depends on precise ultrasound-beam alignment and operator expertise," says Imon Banerjee, Ph.D. , an AI researcher at Mayo Clinic in Phoenix and senior author of the study. "We wanted to determine whether AI could recognize subtle patterns that are imperceptible to the human eye in routinely acquired B-mode ultrasound videos and identify patients with LVOT obstruction earlier, enabling timely confirmatory Doppler evaluation and referral when appropriate."
The study included 1,833 patients in the Mayo Clinic cohort. The model was tested in 275 patients and externally validated in 46 patients from a hospital in South Korea. The AI model used only resting, non-Doppler ultrasound videos to predict whether a patient had a potentially significant obstruction to blood leaving the heart. Researchers found that combining information from three standard ultrasound views improved the model's ability to distinguish patients with elevated LVOT gradients. The model also helped identify obstruction that may only appear when the heart is under stress.
The model maintained strong performance in the South Korean group despite substantial differences between that population and the patients used to develop the model, supporting further study of the technology across different patient populations and clinical settings.
In a subset of cases, the AI model identified obstruction more accurately than two expert echocardiographers who reviewed the same non-Doppler images. The findings highlight how difficult it can be to recognize LVOT obstruction from routine two-dimensional images without Doppler measurements.
"This technology is intended to complement, not replace, Doppler echocardiography," Dr. Banerjee says. "By enabling earlier identification of patients with potential LVOT obstruction, it could escalate timely detection and prompt confirmatory Doppler measurements, stress testing, or referral to an HCM specialty center. It also could support evaluation using portable ultrasound or in settings where comprehensive Doppler assessment may not be readily available, helping expand access to earlier screening and risk assessment."
Dr. Banerjee notes that the next steps include additional prospective validation across broader clinical settings, ultrasound platforms and patient populations.
The list of authors and disclosures may be found in the article, Beyond Doppler: Scalable AI Detection of LVOT Obstruction in HCM . This study received no external funding.
