From a Routine ECG, AI Spots Hidden Signs of Heart Disease in Under Two Seconds

Researchers at Imperial College London have developed an artificial intelligence tool capable of analysing a routine electrocardiogram, or ECG, in under two seconds and detecting hidden signs associated with heart failure and heart valve disease.

The findings were presented at ESC Congress 2026 in Munich. Researchers say the system can extract information from an ordinary ECG that may not be visible to clinicians reviewing the tracing alone.

ECGs are widely used to record the heart's electrical activity and identify rhythm problems, but conditions such as heart failure and valve disease generally require additional investigation, including an echocardiogram. The new AI system is intended to help identify which patients should be prioritised for those more detailed tests.

In testing involving about 67,000 patients in the United States, the system identified up to 81% of heart failure cases and up to 90% of cases of heart valve disease, according to results reported by the research team.

The technology uses AI models trained on millions of ECG recordings linked to patient health data, allowing the system to recognise subtle patterns in electrical signals associated with diseases that may not be apparent during conventional ECG interpretation.

The two-second analysis does not provide a definitive diagnosis. Instead, the technology is designed as a screening and triage tool to flag patients who may require an echocardiogram or further clinical assessment.

A prospective study is now evaluating the technology in routine NHS care. The registered study plans to recruit 590 patients undergoing standard echocardiography and compare the AI assessment with their clinical findings.

Researchers hope the approach could eventually help identify heart disease earlier and allow health systems to prioritise patients who need specialist testing, using a test that is already routinely performed in hospitals and clinics.