HKUMed AI tool uses single blood test to predict six cardiovascular diseases up to 15 years early

Researchers at the LKS Faculty of Medicine of the University of Hong Kong (HKUMed) have developed an artificial intelligence tool called CardiOmicScore that uses data from a single blood test to predict the risk of six major cardiovascular diseases up to 15 years before clinical onset, according to a university press release (med.hku.hk).

The system forecasts future risk for coronary artery disease, stroke, heart failure, atrial fibrillation, peripheral artery disease and venous thromboembolism. It was detailed in a peer-reviewed study published in Nature Communications (nature.com).

Cardiovascular diseases remain the leading global cause of death, responsible for approximately 19.8 million fatalities in 2022. Conventional risk assessment relies on factors such as age, blood pressure and smoking status, while polygenic risk scores capture inherited genetic predisposition. Both approaches have limitations: clinical indicators may miss early molecular changes, and genetic scores remain fixed from birth and do not reflect lifestyle or environmental influences, the researchers noted (eurekalert.org).

CardiOmicScore addresses this by applying deep learning to multiomics data—genomics, metabolomics and proteomics—drawn from UK Biobank participants. The model analysed 2,920 circulating proteins and 168 metabolites in blood samples. These molecules act as real-time indicators of immune activity, metabolism and vascular health, according to the HKUMed team (sciencedaily.com).

Professor Zhang Qingpeng, Associate Professor in the Department of Pharmacology and Pharmacy at HKUMed and co-lead of the work, stated: “Genes determine where we start—they define our baseline health risk. However, proteins and metabolites reflect our current physical health. 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” (med.hku.hk).

In testing, proteomics-based scores achieved C-index values of roughly 0.69–0.82 and metabolomics-based scores 0.64–0.74, compared with about 0.52–0.60 for polygenic risk scores. When combined with clinical data such as age and sex, the multiomics scores improved discrimination (ΔC-index approximately 0.005–0.102) across the six outcomes. Among higher-risk individuals the model flagged elevated risk up to 15 years before symptoms appeared (mobihealthnews.com; nature.com).

The first author is Luo Yan of the HKU Musketeers Foundation Institute of Data Science. The team indicated that future work could involve external validation, including in Asian cohorts, and that a reduced set of key biomarkers might eventually allow practical clinical use from a small blood sample. Professor Zhang added that the goal is to shift health management “from reactive treatment to proactive prediction and intervention” (scitechdaily.com).

The findings highlight complementary contributions of different omics layers and identify specific proteins (such as those linked to cardiac stress and inflammation) and metabolites associated with risk, offering potential avenues for biomarker development.

Why this is uncovered

The research was announced via HKUMed and EurekAlert in March 2026, published in Nature Communications earlier that year, and later summarised by ScienceDaily, SciTechDaily and specialist outlets such as MobiHealthNews. It has received little to no pickup from major international wire services or general mainstream news organisations.