AI can help detect rheumatic heart disease

Credit: Unsplash+.

A groundbreaking study led by researchers at Children’s National Hospital has demonstrated the potential of artificial intelligence (AI) to detect rheumatic heart disease (RHD) with the same accuracy as a cardiologist.

Detailed in the latest edition of the Journal of the American Heart Association, this research highlights how advanced deep learning technology can be applied to address a significant health inequity.

The newly developed AI system, a collaboration involving Ph.D. student Metin Yarici, combines novel ultrasound probes with portable electronic devices.

These devices are equipped with algorithms capable of diagnosing RHD through echocardiograms. This innovation could empower healthcare workers, even those without specialized medical degrees, to detect RHD in endemic regions effectively.

RHD, often caused by repeated Strep A bacterial infections, can lead to severe heart damage if not treated early.

While almost eradicated in high-income countries, RHD remains a significant concern in low- and middle-income nations, affecting 40 million people and causing nearly 400,000 deaths annually. Early detection is key, as the condition is treatable with penicillin.

Kelsey Brown, M.D., a cardiology fellow at Children’s National and co-lead author of the study, emphasizes the practicality and life-saving potential of the single ear-ECG. This new method enables quick screening for signs of RHD, leading to timely specialized care and treatment.

The challenge in diagnosing RHD lies in the necessity of a highly trained cardiologist to interpret echocardiograms. In many impoverished countries, access to such specialized care is limited, often leading to undetected conditions and severe complications.

The AI algorithm developed at Children’s National addresses this gap by identifying mitral regurgitation, a key indicator of RHD, in up to 90% of children with the condition.

In March, a pilot program led by Craig Sable, M.D., interim division chief of Cardiology at Children’s National, will be implemented in Uganda.

This program incorporates AI into the echo screening process for children, using a handheld ultrasound probe, a tablet, and a laptop equipped with the new algorithm.

The research team at Children’s National, including AI experts Dr. Pooneh Roshanitabrizi and Marius George Linguraru, employed machine learning and deep learning techniques to develop this novel algorithm.

It is trained to interpret ultrasound images of the heart to detect RHD, analyzing features that are not visible to the human eye.

For instance, while current guidelines for diagnosing RHD use weight categories as a surrogate for heart size, the AI algorithm can adjust for the heart’s size as a continuously variable factor.

This precision is expected to significantly enhance the capabilities of healthcare workers in diagnosing RHD more quickly and accurately.

The development of this AI system faced challenges, such as teaching the AI to handle the clinical differences in ultrasound images and the complexities of evaluating color Doppler echocardiograms.

However, the team has successfully taught the machine to learn and interpret this data, potentially better than the human eye and brain.

This AI-driven approach promises to improve the global distribution of medical expertise, particularly in low- and middle-income countries with fewer cardiologists.

It signifies a major step towards achieving equity in medicine, allowing for widespread, accurate, and early detection of RHD, and ultimately saving countless lives.

If you care about heart disease, please read studies that herbal supplements could harm your heart rhythm, and how eating eggs can help reduce heart disease risk.

For more information about heart health, please see recent studies that apple juice could benefit your heart health, and results showing yogurt may help lower the death risks in heart disease.

The research findings can be found in Journal of the American Heart Association.

Copyright © 2024 Knowridge Science Report. All rights reserved.