Researchers have developed an artificial intelligence tool that analyzes a routine electrocardiogram in less than two seconds to detect hidden heart failure and heart valve disease, according to findings presented at the European Society of Cardiology annual congress in Munich. The technology extracts subtle electrical patterns from standard tracings that the human eye cannot typically see, potentially offering a rapid way to prioritize patients for further diagnostic testing.
How the AI Tool Analyzes Routine ECGs
The traditional electrocardiogram (ECG) has recorded the heart’s electrical activity, rate, and rhythm for a century, but it cannot independently detect structural heart disease, which usually requires an echocardiogram ultrasound scan. To overcome this limitation, researchers trained an AI system using more than 1.6 million ECGs from Brazil and several million recordings from the United States, linking electrical patterns with subsequent patient diagnoses. According to a US study involving approximately 67,000 patients, the AI tool successfully identified up to 81 percent of individuals with heart failure and up to 90 percent of those with heart valve disease. Dr. Ahmed El-Medany, a clinical research fellow at Imperial College London who led the analysis, described the system as superhuman due to its ability to detect patterns that clinicians cannot consistently identify from an ECG alone.
Prioritizing Patients and Reducing Wait Times
Cardiologists emphasize that the technology serves as a screening and prioritization tool rather than a standalone diagnostic device. Because patients often wait several months for a heart ultrasound scan after referral, Professor Fu Siong Ng of Imperial College London noted that the tool could help hospitals identify high-risk individuals faster and more urgently. Dr. Sonya Babu-Narayan, clinical director of the British Heart Foundation, which funded the trial, stated that while the AI ECG will not detect every heart condition, it provides a solution to fast-track patients most likely to have an abnormality. Beyond targeted referrals, researchers suggest the system could run opportunistically on all hospital ECGs to flag un-suspected cases of heart failure and valve disease.
Following the presentation in Munich, the development team is exploring how to integrate the technology into everyday medical equipment. According to researchers, future applications could include handheld AI-led ECG readers for healthcare professionals. Imperial College London has also established a spinout company named Cardiovolt.ai to support the transition from research systems to clinical applications, though further testing remains necessary before wider deployment in routine healthcare settings.