A single blood sample, analyzed by an artificial intelligence model, can now forecast the risk of six major cardiovascular diseases up to 15 years before clinical onset — potentially transforming how we prevent the world's leading cause of death.
Most heart attacks and strokes happen without warning. The underlying disease — atherosclerosis, the gradual hardening of arteries — can develop silently for decades before producing its first symptom. A new AI-powered blood test developed at the University of Hong Kong (HKUMed) aims to change that by reading the molecular signals of future disease long before any organ is damaged.
Called CardiOmicScore, the system analyzes a single blood sample using AI to integrate thousands of molecular signals — proteins, metabolites, genetic markers, and other biological data. Unlike traditional risk scores that ask about age, cholesterol, and smoking history, this approach captures the body's real-time biological state. The model can predict coronary artery disease, stroke, heart failure, atrial fibrillation, and two other major cardiovascular conditions with a lead time of up to 15 years.
Cardiovascular disease remains the leading cause of death globally, claiming an estimated 18 million lives each year. Most of those deaths are from conditions that develop over decades but are only detected at the moment of crisis. A predictive test that works at scale — requires only a routine blood draw and costs no more than existing lab panels — could shift cardiology from a reactive discipline to a preventive one. The HKUMed team is now working on clinical validation in larger and more diverse populations, with the goal of making the test available in routine health screenings.
The ability to see a heart attack coming 15 years in advance would be one of the most powerful preventive tools in modern medicine. CardiOmicScore may be the first step toward making that a reality.