‘Superhuman’ AI tool spots heart disease in less than 2 seconds

Doctors have unveiled an artificial‑intelligence system that can detect heart failure and heart‑valve disease from a standard electrocardiogram in under two seconds, a speed that far exceeds human interpretation. Trained on millions of routine ECG recordings, the tool extracts subtle patterns invisible to the naked eye and flags patients who are likely to have these conditions. In a U.S. trial of 67,000 individuals, the algorithm identified up to 81 % of heart‑failure cases and up to 90 % of valve‑disease cases, offering a rapid “read‑out” that could direct high‑risk patients toward definitive echocardiography much sooner than current practice allows.

The breakthrough was presented at the European Society of Cardiology’s annual congress in Munich, highlighting its potential to transform early diagnosis for two of the most common forms of heart disease. Traditional ECGs, while ubiquitous—about one billion are performed worldwide each year—cannot diagnose structural heart problems, which normally require an echocardiogram that patients often wait months to receive. By using the AI model to prioritize those most likely to have heart failure or valve disease, clinicians can fast‑track echocardiograms, reducing diagnostic delays that currently jeopardize timely treatment. Researchers emphasized that the system is not a definitive diagnostic tool but a powerful screening aid that can highlight high‑risk individuals for urgent follow‑up.

Experts foresee broader applications beyond the initial target conditions. Imperial College’s Prof Fu Siong Ng suggested the AI could be run on every ECG performed in a hospital, opportunistically flagging unsuspected cases of heart failure or valve disease, thereby catching disease earlier in patients who presented for unrelated reasons. Dr Sonya Babu‑Narayan of the British Heart Foundation called the technology “exciting,” noting its potential to save lives by enabling earlier intervention. Future work aims to embed the algorithm into handheld ECG devices for point‑of‑care use, while the Munich conference also showcased related AI advances, such as five‑second facial‑video analysis from Japanese researchers that can detect undiagnosed hypertension and type‑2 diabetes, underscoring a growing wave of AI‑driven diagnostic tools poised to reshape preventive medicine.

⚡ Effects Interpreter