AI Tool Detects Hidden Heart Disease in Two Seconds

Researchers at Imperial College London developed an AI tool capable of analyzing electrocardiograms in under two seconds to identify signs of heart failure and valve disease invisible to human clinicians.

In a U.S. study of 67,000 patients, the system successfully flagged up to 81% of heart failure cases and 90% of heart valve disease cases using models trained on over 1.6 million ECGs.

The technology, now being developed for clinical use by spinout company Cardiovolt.ai, is intended as a triage tool to prioritize high-risk patients for follow-up echocardiograms rather than providing a definitive diagnosis.