Health
Researchers develop AI tool to identify heart disease signs from routine ECGs
Presented at the European Society of Cardiology congress, the model analyzes electrocardiograms in seconds to help prioritize patients for ultrasound scans.
The short version
- Researchers presented an AI model trained on millions of electrocardiograms (ECGs) that flags signs of heart failure and valve disease in under two seconds.
- In a US trial of 67,000 patients, the tool detected up to 81% of heart failure cases and up to 90% of heart valve disease cases.
- The tool cannot provide a definitive diagnosis on its own, but researchers aim to use it to prioritize high-risk patients for echocardiograms and screen routine hospital ECGs.
- The next steps for development include designing handheld AI-assisted ECG readers for medical professionals.
Key facts
- Researchers presented an AI tool at the European Society of Cardiology annual congress in Munich capable of reading ECG results in under two seconds to detect potential heart failure and valve disease.[The Guardian]
- A trial of 67,000 US patients showed the model identified up to 81% of individuals with heart failure and up to 90% of those with heart valve disease.[The Guardian]
- The development and trial were funded by the British Heart Foundation.[The Guardian]
- Researchers at the University of Tokyo and Institute of Science Tokyo separately presented research demonstrating that AI analysis of five-second facial videos could detect high blood pressure and type 2 diabetes.[The Guardian]
What remains uncertain
- The AI tool cannot independently diagnose heart conditions and requires confirmation via standard ultrasound scans (echocardiograms).[The Guardian]
- It remains to be seen how effectively handheld AI-led ECG readers can be designed and integrated into real-world healthcare settings.[The Guardian]
Sources
- ‘Superhuman’ AI tool spots heart disease in less than 2 secondsThe Guardian - Science