AI Demonstrates High Accuracy in Identifying Rheumatic Heart Disease from Echocardiograms
January 29, 2024A study led by Children’s National Hospital in Washington reveals that artificial intelligence (AI) can diagnose early-stage rheumatic heart disease with accuracy comparable to expert cardiologists based on the analysis of echocardiograms.
Rheumatic heart disease claims the lives of nearly 300,000 individuals annually, primarily in low- and middle-income countries. It primarily affects children or young people and results from heart damage following rheumatic fever—an immune response to group A streptococci infections.
Early detection of rheumatic heart disease is crucial, as it can be effectively treated with antibiotics. Access to cardiac specialists can be challenging in many countries where the disease persists. Therefore, Kelsey Brown, a pediatric cardiology fellow at Children’s National Hospital, and colleagues explored the effectiveness of convolutional neural networks in diagnosing early rheumatic heart disease compared to expert evaluations.
The study, published in the Journal of the American Heart Association, involved testing the AI’s efficacy using 511 echocardiogram images of children, with 282 having rheumatic heart disease and 229 without. The AI was trained to identify nearly 40 challenging features in echocardiograms associated with early-stage rheumatic heart disease, focusing on signs of mitral valve regurgitation where the heart valve fails to close properly.
The AI demonstrated a capability to detect mitral regurgitation in approximately 90% of the children, aligning with the accuracy of expert cardiologists. Kelsey Brown, MD, the first and co-lead author, emphasized the potential of the technology to extend the reach of cardiologists globally. The AI system could quickly screen children for signs of rheumatic heart disease, leading to early intervention and preventive antibiotics.
In the coming months, Craig Sable, co-lead author and interim division chief of Cardiology at Children’s National Hospital, along with collaborators, will initiate a project to deploy the AI technology in clinics in Uganda. The aim is to detect and treat early-stage rheumatic heart disease cases before severe symptoms manifest.
Craig Sable emphasized the potential impact of the AI technology in preventing rheumatic heart disease by identifying affected individuals in the early stages. Early detection allows for timely intervention, involving the administration of affordable monthly penicillin, which can significantly reduce the risk of severe consequences. Sable expressed optimism that, once the technology is developed and deployed on a broader scale, it holds great promise to deliver highly accurate care in economically disadvantaged countries and contribute to the global efforts to eradicate rheumatic heart disease.
