Researchers at the University of Hong Kong developed CardiOmicScore, an artificial intelligence blood test that predicts cardiovascular disease up to 15 years before symptoms appear. The test analyzes thousands of proteins and metabolites in blood samples to assess risk for six major conditions: heart attack, stroke, heart failure, atrial fibrillation, peripheral artery disease, and aortic stenosis.

The approach differs fundamentally from traditional genetic risk scores. While DNA-based assessments remain constant throughout a person's life, CardiOmicScore captures dynamic biological changes reflecting current health status, lifestyle choices, and environmental factors. This flexibility allows the test to identify accumulating disease risk as people age and their circumstances shift.

The test's early detection window represents a major advantage in cardiovascular medicine. Current screening methods typically identify disease only after significant damage occurs. A 15-year advance warning gives patients and clinicians ample time to implement preventive interventions. Lifestyle modifications, medication adjustments, and closer monitoring could all begin years before critical events.

The Hong Kong team trained their AI model on proteomic and metabolomic data, leveraging machine learning to identify patterns humans cannot discern. These molecular signatures reflect underlying physiological changes associated with cardiovascular deterioration long before clinical diagnosis becomes possible.

Several limitations merit consideration. The research appears preliminary based on available information, raising questions about validation across diverse populations. Cardiovascular risk varies substantially by ancestry, geography, and healthcare access. Whether CardiOmicScore performs equally well in different demographic groups remains unclear. The test's clinical implementation also depends on cost, accessibility, and whether insurance covers screening for asymptomatic individuals.

The technology builds on growing recognition that multi-marker blood tests outperform single biomarker approaches. Previous research demonstrated that panels measuring dozens or hundreds of proteins provide better disease prediction than traditional markers like cholesterol levels alone.

If