Researchers from Yale University Develop AI Model to Detect Heart Disease
Yale researchers have developed a new artificial intelligence model that diagnoses the potentially fatal ATTR-CM heart condition using electrocardiogram images instead of raw signals.
Researchers at Yale University have announced the development of a new artificial intelligence-based model to detect ATTR-CM, a fatal heart condition caused by protein accumulation that manifests with symptoms such as shortness of breath and fatigue.
Heart Disease and Diagnostic Challenges
Shortness of breath, fatigue, and chest tightness can result from excessive exercise or be signs of more serious conditions such as heart disease.
Transthyretin cardiac amyloidosis, or ATTR-CM for short, is a condition that occurs due to the accumulation of misfolded proteins in the heart and can be fatal if not diagnosed in time.
New Artificial Intelligence-Powered Approach
Published in the Journal of the American Medical Association, the study included Yale researchers Philip Croon, Bruno Batinica, and Lovedeep Dhingra.
The researchers developed an artificial intelligence model that directly analyzes images of electrocardiogram (ECG) tracings instead of raw numerical data or signals.
Working Principle and Testing Process of the Model
This newly developed artificial intelligence model undergoes its learning process by examining thousands of images containing patterns associated with the disease.
Tested across eight different cohorts in the United States and Europe, the model can check the quality of the graph and detect abnormalities.
Future Goals and Approval Processes
While Yale professor James Duncan stated that the results obtained are promising, he pointed out that certain hurdles remain for real-world applications.
The research team is currently working to secure FDA approval and is collaborating with health systems to enable the model to be used on clinicians' smartphones.