Improved cardiovascular risk prediction using targeted plasma proteomics in primary prevention.

Area of Science:

• Cardiovascular disease research
• Proteomics and biomarker discovery
• Personalized medicine

Background:

• Accurate identification of individuals at high cardiovascular (CV) risk is crucial for personalized medicine.
• Traditional risk factors and single biomarkers have limitations in predicting CV events effectively.
• Novel proteomic technologies offer potential for improved CV risk prediction.

Purpose of the Study:

• To compare the predictive performance of a protein-based risk model against a traditional clinical risk model for CV events.
• To validate the findings in independent prospective cohorts for clinical implementation.

Main Methods:

• Utilized proximity extension assay to measure 368 proteins in 822 individuals from the EPIC-Norfolk cohort and 702 from the PLIC cohort.
• Developed prediction models using tree-based ensemble and boosting methods: protein-based, clinical risk-based, and combined.
• Evaluated model performance using area under the curve (AUC) for myocardial infarction and 3-year event prediction.

Main Results:

• A 50-protein panel significantly outperformed the clinical risk model in predicting myocardial infarction (AUC 0.754 vs. 0.730) in the EPIC-Norfolk cohort.
• The protein model showed superior prediction for 3-year CV events (AUC 0.803 vs. 0.732) compared to the clinical model.
• The superior predictive value of the protein panel was confirmed in the PLIC validation cohort (AUC 0.705 vs. 0.609).

Conclusions:

• A proteome-based model demonstrates superior performance in predicting CV events in primary prevention compared to models based on clinical risk factors.
• The findings support the potential clinical utility of proteomic profiling for enhancing cardiovascular risk assessment.
• Further validation in large prospective cohorts is recommended for widespread clinical implementation in CV prevention.