University of Oxford AI predicts heart attacks up to 10 years in advance

Researchers at the University of Oxford have developed an artificial intelligence system capable of predicting heart attacks up to ten years in advance. The study, led by Professor Haralambos Antoniades, focuses on analyzing CT scans of patients presenting with chest pain.

By identifying subtle patterns in fatty tissue and minimal arterial narrowings that human experts might miss, the AI could potentially reduce heart attacks by over 20 percent and strokes or cardiac deaths by 8 percent. In a clinical trial involving 744 patients, cardiologists adjusted treatment plans for up to 45 percent of cases based on the AI-generated risk assessments.

This technology addresses a critical gap in current diagnostics, where many patients are sent home as healthy despite having high long-term risks. The implementation of such AI tools is expected to be cost-effective for healthcare systems while significantly improving patient outcomes.