Innovative AI Tool Enhances Heart Disease Detection Using Standard ECG Data
A cutting-edge AI tool analyzes routine ECG data to detect hidden structural heart diseases, promising earlier diagnosis and treatment opportunities. Developed by Columbia University researchers, EchoNext offers a cost-effective, non-invasive method to identify patients needing further cardiac evaluation, potentially saving countless lives.
Recent advancements in artificial intelligence (AI) have led to the development of a novel, cost-effective screening tool capable of identifying hidden structural heart diseases from routine electrocardiogram (ECG) data. Traditionally, ECGs are employed to detect abnormal heart rhythms, blockages, and previous heart attacks, but they have limitations in diagnosing structural issues such as valve disease and cardiomyopathy. To bridge this gap, researchers at Columbia University Irving Medical Center and NewYork-Presbyterian created EchoNext, an AI-powered system that analyzes standard ECG readings to determine which patients should proceed with an echocardiogram—a more detailed, ultrasound-based heart imaging test.
The AI model was trained on over 1.2 million paired ECG and echocardiogram data from 230,000 patients, enabling it to spot signs of structural heart problems more accurately than many cardiologists, even with AI assistance. In a validation study across four hospital systems, including several NewYork-Presbyterian campuses, EchoNext demonstrated high accuracy in detecting conditions like heart failure, valve disease, pulmonary hypertension, and significant heart thickening. When compared directly with 13 cardiologists analyzing 3,200 ECGs, EchoNext correctly identified 77% of structural heart diseases, outperforming human experts with a 64% accuracy rate.
In real-world applications, the AI system screened nearly 85,000 patients who had not previously undergone echocardiography. It flagged over 7,500 individuals as high-risk, and during follow-up, about 55% of these patients subsequently received their first echocardiogram, leading to a diagnosis of structural heart disease in nearly three-quarters of them. This approach has the potential to significantly increase early detection, potentially saving many lives by enabling timely treatment. The team has also shared anonymized data to support further research and is conducting clinical trials to implement EchoNext in emergency settings. Overall, this innovation exemplifies how AI can transform cardiac screening, making it more accessible and effective worldwide.
Stay Updated with Mia's Feed
Get the latest health & wellness insights delivered straight to your inbox.
Related Articles
Enhanced Prediction of Chronic Pain Using Machine Learning with Biological and Psychosocial Data
Cutting-edge machine learning research reveals that combining biological markers with psychosocial factors enhances the prediction and understanding of chronic pain conditions, supporting a holistic biopsychosocial approach.
Four States Urge FDA to Relax Restrictions on Mifepristone, the Abortion Pill
Four states have petitioned the FDA to remove restrictions on mifepristone, citing its proven safety for medication abortions and seeking to improve access for patients nationwide.
Brain Power and Cognitive Skills in Hockey: Position-Specific Brain Processing Revealed by Study
Recent research reveals that hockey players' cognitive processing speeds vary by position, and targeted brain training can enhance their performance and safety on the ice. Using portable neurotechnology, athletes can now measure and improve neural responses for better reaction times and decision-making.
Shorter Telomeres and the Risk of Age-Related Brain Diseases
Shorter telomeres are linked to a higher risk of age-related brain diseases, but healthy lifestyle choices may mitigate this risk. New research suggests that behavioral changes can help protect cognitive health as we age.