The Diagnosis of Cardiovascular Disease Using Simple Blood Biomarkers Through AI and Big Data

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Cardiovascular Medicine

Background:

• Cardiovascular disease (CVD) is a major global cause of death.
• Current diagnostic methods, like imaging, are costly and often overused in asymptomatic individuals.
• There is a need for cost-effective, non-imaging methods for CVD risk stratification.

Purpose of the Study:

• To develop an Artificial Intelligence (AI)-based solution for cardiovascular disease (CVD) risk stratification.
• To utilize routine blood biomarkers as a pre-imaging risk assessment tool.
• To identify cost-effective, non-imaging risk factors for CVD using UK Biobank data.

Main Methods:

• Anonymized data from over 500,000 UK Biobank (UKB) patients were analyzed.
• A dataset of 701 features including demographics, blood tests, and clinical assessments was curated.
• A hybrid XGBoost classifier with a scalable loss function was employed for risk stratification.

Main Results:

• The AI model achieved 0.83 accuracy, 0.82 sensitivity, and 0.84 specificity in diagnosing CVD comorbidities.
• Key predictive biomarkers identified were blood pressure, Body Mass Index (BMI), and age.
• The study demonstrated the feasibility of using non-imaging data for CVD risk assessment.

Conclusions:

• AI-powered analysis of routine blood biomarkers offers a cost-effective approach to CVD risk stratification.
• This method can reduce reliance on expensive and potentially overused imaging modalities.
• The findings highlight the potential of leveraging large datasets like UKB for advancing cardiovascular health.