Artificial Intelligence Has Learned to Predict Sudden Death## Contribution

Breakthrough in cardiology: AI from Johns Hopkins University predicts sudden cardiac death with almost 90% accuracy

A revolutionary development by scientists from Johns Hopkins University could save thousands of lives. The artificial intelligence they created is capable of predicting the risk of sudden cardiac death with 89% accuracy in patients with hypertrophic cardiomyopathy - a hereditary heart disease that becomes a merciless killer of young people and athletes. The research results are published in the prestigious journal Nature Cardiovascular Research (NCR).

Modern clinical methods identify patients at risk of fatal arrhythmias in only half of cases. The innovative MAARS model takes diagnostics to a new level, applying deep learning to analyze cardiac MRIs and medical indicators. The system is able to detect scars in heart tissue that are invisible to conventional methods, which significantly increase the probability of a fatal outcome.

"Currently, many patients die because they weren't protected in time, while others wear unnecessary defibrillators all their lives that don't benefit them," emphasized project leader Natalia Trayanova.

The uniqueness of the development lies not only in the accurate determination of risk level but also in the ability to explain the reasons for its predictions, which allows cardiologists to develop personalized treatment plans. In the future, scientists plan to adapt the technology for diagnosing other cardiovascular diseases.