Researchers at the University of Hong Kong have developed an AI-powered blood test, called CardiOmicScore, that can flag the risk of six major cardiovascular diseases up to 15 years before symptoms appear. The tool analyzes 2,920 circulating proteins and 168 metabolites from a single blood sample, using deep learning to combine genomic, proteomic, and metabolomic data drawn from UK Biobank records.
The six conditions covered are coronary artery disease, stroke, heart failure, atrial fibrillation, peripheral artery disease, and venous thromboembolism — together responsible for roughly 19.8 million deaths worldwide each year. Unlike fixed genetic risk scores, CardiOmicScore captures a person's current biological state, so it can be updated as health, lifestyle, and environment change over time.
"Genes determine where we start, but proteins and metabolites reflect our current physical health," said lead researcher Professor Zhang Qingpeng, whose team published the findings in Nature Communications. The researchers say the approach substantially outperforms conventional polygenic risk scores and could shift cardiovascular care from reactive treatment toward proactive prediction and early intervention.