Michigan-Developed AI Model Helps Diagnose Undetected Heart Disease From Simple EKG
michiganmedicine.org
|
Contributed by: Kate Gamble
Summary
Researchers at the University of Michigan have developed an AI model capable of diagnosing coronary microvascular dysfunction (CMVD) using a 10-second electrocardiogram (EKG) strip, a significant advancement over traditional imaging methods. CMVD is commonly underdiagnosed and can lead to serious cardiovascular issues, but the new model can accurately predict myocardial flow reserve, the gold standard for diagnosing this condition. The innovative use of self-supervised learning to analyze over 800,000 EKG waveforms enhances the model's accuracy and accessibility, offering healthcare professionals a cost-effective and efficient tool for early detection of CMVD. This development underscores the potential for AI to transform cardiac diagnostics, particularly for conditions that are difficult to detect with conventional techniques.
