Lifestyle
AI Breakthrough: Boosting Heart Disease Detection by 20% Over Traditional Tests
Cardiovascular disease remains a leading cause of mortality in the United States, with one person succumbing to the condition every 34 seconds, as reported by the Centers for Disease Control and Prevention. While lifestyle and dietary modifications play a crucial role in heart health, early detection and prevention are equally vital.
In a significant advancement, researchers have developed an innovative AI tool that enhances the detection of heart disease, building on recent successes in cancer diagnosis. This groundbreaking tool, known as EchoNext, is designed to work alongside traditional electrocardiograms (ECG) to uncover hidden heart conditions.
Pierre Elias, a study author and assistant professor at Columbia University, highlighted a gap in heart disease diagnostics: “We have colonoscopies, we have mammograms, but we have no equivalents for most forms of heart disease.”
EchoNext is engineered to analyze ECG data, pinpointing patients who may need further testing, such as an echocardiogram, to identify structural heart issues. These problems can involve abnormalities in the heart’s valves, walls, chambers, or muscles, and pose significant health risks if left undetected.
Recent research published in the journal Nature underscores EchoNext’s capability, revealing that it “accurately identified structural heart disease from ECG readings more often than cardiologists.”
The ECG is a staple test in cardiology, measuring electrical signals and heartbeats to detect irregularities, blockages, or previous heart damage. Despite its routine use, Elias noted a common misconception: “We were all taught in medical school that you can’t detect structural heart disease from an electrocardiogram.”
EchoNext challenges this notion by using AI to determine when a follow-up cardiac ultrasound is necessary. The model was trained using over 1.2 million ECG-echocardiogram records from 230,000 patients, demonstrating high accuracy in identifying various heart conditions.
In comparative studies, EchoNext outperformed cardiologists by 20 percent in detecting structural heart issues. In a trial of 3,200 ECGs, the tool achieved a 77 percent success rate, compared to a 64 percent accuracy rate by a team of 13 doctors.
“EchoNext basically uses the cheaper test to figure out who needs the more expensive ultrasound,” Elias explained. “It detects diseases cardiologists can’t from an ECG. We think that ECG plus AI has the potential to create an entirely new screening paradigm.”
In a recent screening of 85,000 ECG patients, EchoNext identified over 7,500 individuals as high-risk for undiagnosed structural heart disease. Of those flagged, 55 percent were recommended for an echocardiogram, leading to a diagnosis rate nearly double that of traditional methods.
Elias emphasized the potential impact of this technology: “Using our technology, we may be able to turn the estimated 400 million ECGs that will be performed worldwide this year into 400 million chances to screen for structural heart disease and potentially deliver life-saving treatment at the most opportune time.”
Currently, Elias and his team are conducting a clinical trial across eight emergency departments to further validate EchoNext’s efficacy and prepare for its broader implementation in hospitals nationwide.
Let us know what you think, please share your thoughts in the comments below.
