A single blood test analyzed by artificial intelligence may soon warn people about serious heart and circulation problems more than a decade before symptoms ever appear. Researchers at the University of Hong Kong’s Li Ka Shing Faculty of Medicine (HKUMed) have developed CardiOmicScore, an AI-powered tool that turns a routine blood draw into a personalized forecast of future cardiovascular risk.
The system uses deep learning to sift through thousands of proteins and metabolites circulating in the blood, molecules the researchers describe as the body’s real time recorders of immune, metabolic, and vascular health. Trained on data from the UK Biobank, including 2,920 proteins and 168 metabolites, CardiOmicScore can estimate a person’s risk for six major cardiovascular diseases: coronary artery disease, stroke, heart failure, atrial fibrillation, peripheral artery disease, and venous thromboembolism.
What sets the tool apart from older genetic risk scores is that it captures the body’s present state rather than a fixed, inherited baseline. ‘Genes determine where we start, they define our baseline health risk. However, proteins and metabolites reflect our current physical health,’ said Professor Zhang Qingpeng, who led the research. ‘Our AI tool is designed to decode these complex molecular signals, enabling doctors and patients to identify risks much earlier, which can potentially change the trajectory of disease through timely lifestyle modifications and early prevention.’
Cardiovascular disease remains the world’s leading cause of death, linked to nearly 19.8 million deaths in 2022 alone. Standard checkups often rely on age, blood pressure, and smoking history, measures that can miss the subtle biological shifts that precede a diagnosis. By flagging elevated risk up to 15 years in advance, the researchers hope CardiOmicScore could give patients and doctors a much wider window to intervene through lifestyle changes or closer monitoring, shifting cardiac care from reactive treatment toward proactive prevention.
The findings, led by Professor Zhang and first author Luo Yan of the HKU Musketeers Foundation Institute of Data Science, were published in Nature Communications. The team says a single small blood sample could eventually be enough to generate a full cardiovascular risk profile, a shift they describe as moving precision medicine from a static, gene-centered model to one based on a person’s dynamic, real-time biology.
