This AI-powered stethoscope can detect pulmonary hypertension

A new artificial intelligence (AI) algorithm integrated into a digital stethoscope by Eko Health shows promise in detecting pulmonary hypertension (PH), according to a study published in the Journal of the American Heart Association (JAHA).

The study, led by researchers from Brown University alongside the team at Eko Health, evaluated the AI’s ability to detect elevated pulmonary systolic pressures (PASP), a key indicator of PH. The model was trained on approximately 6,000 heart sound recordings paired with echocardiographic pressure estimates.

The study highlighted the model’s capabilities, using phonocardiograms (PCGs) recorded with a standard handheld stethoscope to screen for PASP. The algorithm performed well, achieving an average area under the receiver operating characteristic (AUROC) curve of 0.79, with a sensitivity of 71% and a specificity of 73%.

“This innovative approach demonstrates how combining digital stethoscopes with advanced AI can lead to a low-cost, non-invasive, point-of-care screening tool for the early detection of pulmonary hypertension,” Gaurav Choudhary, MD, lead principal investigator of the study, professor and director of Cardiovascular Research at Brown University, said in an organizational release. “Our findings represent a significant advancement in clinical practice that can ultimately enhance patient care.”

PH, a condition characterized by elevated blood pressure in the arteries of the lungs, is often underdiagnosed due to the reliance on echocardiography, which is not always accessible in primary care settings.

Left untreated, PH can lead to heart failure, early disability or early mortality. Eko Health’s AI-enhanced stethoscope offers primary care physicians a noninvasive, rapid screening tool to aid in earlier detection of PH.

The study also highlighted the algorithm’s ability to pinpoint key acoustic markers associated with PH — including an altered duration between the two components of the second heart sound and a tricuspid valve regurgitation murmur. By mapping these abnormalities, the AI tool provides physicians with a transparent and easily interpretable assessment.

According to the news release from Eko Health, across the world, as many as 10% of people aged 65 and older are impacted by PH, with millions more affected under the age of 65. Eko Health’s AI approach to PH diagnosis has the potential to transform early detection, diagnosis and management of PH, leading to better patient care in primary care practices.