Researchers at Imperial College London have created an AI tool capable of interpreting an electrocardiogram (ECG) in under two seconds, detecting signs of heart failure and valve disease that cardiologists often miss. This innovation aims to accelerate diagnosis and prioritize patients for urgent follow-up testing.
- AI detects heart failure with up to 81% accuracy from ECGs
- Identifies valve disease cases with up to 90% accuracy
- Developed at Imperial College London, now commercializing through Cardiovolt.ai
What happened
Imperial College London researchers have developed an artificial intelligence system trained to interpret ECG results in under two seconds. The AI can flag signs of heart failure and valve disease that are not visibly discernible to cardiologists using standard ECG analysis. This breakthrough was shared at the European Society of Cardiology congress in Munich.
The training involved a large dataset of 1.6 million ECGs from Brazil linked to patient outcomes, complemented by over 67,000 US patient records. The AI demonstrated the ability to identify 81% of heart failure cases and 90% of valve disease cases from ECGs—tests initially not designed to diagnose these conditions.
Why it matters
Heart failure and valve disease require timely diagnosis and treatment to prevent serious health consequences. However, current diagnostic pathways rely heavily on echocardiograms, which require specialized equipment and trained staff, causing delays of several months for many patients.
By analyzing routinely collected ECGs, which are cheap and quick to perform, the AI provides a scalable method for fast-tracking patients likely to have structural heart problems. This triage tool could reduce diagnostic backlogs, allowing clinicians to prioritize echocardiography appointments and initiate treatment earlier.
What to watch next
The AI technology is being commercialized via the spinout company Cardiovolt.ai, aiming to integrate the software into handheld ECG devices. This would expand its use beyond hospitals into primary care and community settings, potentially identifying patients who have not yet been referred for heart evaluations.