research
A Blood Signature Forecasts Six Heart Diseases
CardiOmicScore combines thousands of proteins and metabolites to identify elevated cardiovascular risk years ahead.
Summary
CardiOmicScore combines thousands of proteins and metabolites to identify elevated cardiovascular risk years ahead.
A University of Hong Kong team trained a deep-learning model on UK Biobank measurements spanning 2,920 proteins and 168 metabolites. The resulting CardiOmicScore predicted risk across six cardiovascular diseases and, for people it classified as elevated risk, could signal danger as far as 15 years ahead. In the reported analysis it outperformed a polygenic risk score. The score remains a research model derived from one large cohort, not an approved clinical blood test or a substitute for established assessment.
Why it matters
CardiOmicScore combines thousands of proteins and metabolites to identify elevated cardiovascular risk years ahead.
Limits and context
- The score remains a research model derived from one large cohort, not an approved clinical blood test or a substitute for established assessment.
Key claims
CardiOmicScore combines thousands of proteins and metabolites to identify elevated cardiovascular risk years ahead.
Qualification: The score remains a research model derived from one large cohort, not an approved clinical blood test or a substitute for established assessment.
Evidence: source-2026-07-20-008
Sources
- University of Hong Kong via ScienceDaily: CardiOmicScoreUniversity of Hong Kong via ScienceDaily · secondary reporting
Corrections
No corrections have been recorded for this story.