A new study published in The Lancet Digital Health found that an AI model can predict patients' risk of developing or worsening disease and their risk of early death using electrocardiograms.
Here are seven things to know about the study:
1. Researchers at Imperial College London and Imperial College Healthcare NHS Trust developed the AI model to help physicians identify disease earlier and prioritize urgent cases for treatment.
2. The research team used millions of ECGs, previously taken as part of routine care from international sources, to train the AI model. It accurately predicted which patients experienced new disease, whose conditions worsened and who subsequently died.
3. The model was trained to read and identify patterns within the electrical signals traveling between different chambers of the heart and can understand these patterns with more "complexity" and "subtlety" than a cardiologist.
4. Known as AI-ECG risk estimation, the model correctly predicted the risk of death in the 10 years following the ECG in 78% of cases. In the remaining cases, outcomes could have been influenced by unknowable factors, such as subsequent treatment or an unpredicted cause of death.
5. The system can also predict future health issues, such as heart rhythm problems, heart attacks, and heart failure, as well as the risk of death from non-heart-related causes.
6. "Our analysis shows that the AI can tell us a lot about not only the heart but also what is going on elsewhere in the body and may be able to detect accelerated aging," said Arunashis Sau, MD, a cardiologist and academic clinical lecturer at ICL's National Heart and Lung Institute, who led the research.
7. Imaging and genetic information were also analyzed in the study, helping researchers confirm that the AI predictions were linked to real biological factors in the heart's structure and function.