A new artificial intelligence tool can identify warning signs of heart failure and heart valve disease from routine electrocardiograms in less than two seconds, giving doctors another way to decide which patients may need faster cardiac scans.
Researchers at Imperial College London developed the system with support from the British Heart Foundation. The AI analyses subtle electrical patterns within ECG recordings that doctors may not detect through routine visual assessment.
In testing involving about 67,000 patients in the United States, the system identified up to 81% of heart failure cases and 90% of patients with heart valve disease. Researchers presented the findings at the European Society of Cardiology congress in Munich.
A standard ECG records the electrical activity of the heart and helps doctors check heart rhythm, rate and other abnormalities. Researchers believe the same recordings contain additional information that can point to underlying disease before patients receive more detailed tests.
The Imperial team trained its AI models using more than 1.6 million ECGs from Brazil linked to patient medical records, along with several million recordings from the United States. The large dataset allowed the system to learn patterns linked to heart conditions that normally require further examination.
Dr. Ahmed El-Medany, who led the Imperial analysis, described the technology as ‘superhuman’ because it can detect patterns within ECG traces that clinicians cannot reliably recognise by eye.
Researchers expect the tool to work as a screening system rather than a final diagnostic test. Patients flagged as higher risk could receive an echocardiogram sooner, while doctors would continue using scans and clinical examinations to confirm any diagnosis.
Professor Fu Siong Ng of Imperial College London said some patients currently wait several months for an echocardiogram after referral. Using AI to identify patients with stronger warning signs could help clinicians decide who needs earlier testing.
Dr Sonya Babu-Narayan, clinical director at the British Heart Foundation, said the technology could support faster detection and treatment but cautioned that it ‘will never pick up all those with heart disease.’
The team is also studying whether AI could analyse ECGs performed for unrelated medical reasons and identify patients who may have undiagnosed heart failure or valve disease. This use could let routine ECG tests serve as an additional screening method without requiring a separate initial procedure.
Researchers are now working on clinical validation and possible handheld ECG readers for healthcare professionals. Imperial's research group has also created Cardiovolt.ai to support the technology's development.
Further testing across different patient groups and healthcare systems will be required before the AI can become part of routine clinical care.
Also Read: Apple Watch Series 12, Ultra 4 Could Get New Health Tracking Features