Key facts
- A new AI tool can detect signs of heart failure and heart valve disease from ECGs in under two seconds.
- The AI analyzes millions of patient records to extract more information from ECGs than the human eye.
- Trials involving 67,000 patients showed the AI identified up to 81% of those with heart failure and 90% with heart valve disease.
- The technology could help fast-track patients for echocardiograms, potentially leading to earlier diagnosis and treatment.
- The AI tool is not a standalone diagnostic device but provides a strong indication of potential heart conditions.
Doctors have developed an artificial intelligence tool capable of detecting heart disease from routine electrocardiograms (ECGs) in less than two seconds. This groundbreaking technology, trained on millions of patient records, analyzes ECGs to extract subtle information that the human eye cannot typically perceive, identifying signs of heart failure and heart valve disease.
ECGs, which have been a standard medical tool for a century to record heart rate and rhythm, have historically been unable to detect heart disease, a diagnosis usually requiring an echocardiogram, a type of ultrasound scan that can involve long waiting times. The new AI tool offers the potential to significantly speed up this diagnostic process.
The development was presented at the European Society of Cardiology annual congress in Munich. Early diagnosis of heart failure and heart valve disease is crucial for timely intervention with lifesaving medicines. The AI's ability to analyze ECGs, one of the most common medical tests performed globally, makes it a potentially significant advancement.
In a trial involving 67,000 patients in the US, the AI tool demonstrated a high success rate, identifying up to 81% of individuals with heart failure and up to 90% with heart valve disease. Dr. Sonya Babu-Narayan of the British Heart Foundation highlighted the excitement around AI's potential to deliver rapid ECG readouts and identify high-risk patients for early intervention, emphasizing that while it won't detect every condition, it can help prioritize those most likely to have an abnormality.
While the AI cannot definitively diagnose or rule out these conditions on its own, it provides a strong indication, enabling patients identified as high-risk to be rapidly referred for echocardiograms. Professor Fu Siong Ng of Imperial College London noted that this prioritization could significantly reduce waiting times for scans, allowing for earlier treatment. He also suggested a broader application where the AI could be run on all hospital ECGs to opportunistically flag individuals at risk, even if their initial ECG was for unrelated reasons.
Dr. Ahmed El-Medany described the tool as a "superhuman AI" and indicated that future challenges include developing handheld AI-led ECG readers for healthcare professionals. The conference also heard about related AI advancements, such as facial video analysis for detecting undiagnosed high blood pressure and type 2 diabetes.