Optimized Clinical Feature Analysis for Improved Cardiovascular Disease Risk Screening.

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Cardiovascular disease (CVD) poses a significant global health burden.
• Accurate and efficient CVD risk prediction is crucial for timely intervention.
• Current risk assessment tools can be time-consuming, requiring extensive clinical data input.

Purpose of the Study:

• To develop a clinical decision support tool for accurate CVD risk prediction.
• To minimize the number of clinical features required for risk assessment.
• To enhance clinician efficiency and patient risk evaluation.

Main Methods:

• A robust feature selection approach was employed to identify key CVD risk factors.
• A machine learning model was developed using an optimized set of five critical features.
• Explainable artificial intelligence (XAI) techniques were utilized to derive patient-specific insights.

Main Results:

• The developed model achieved state-of-the-art performance with an AUROC of 91.30%.
• High sensitivity (89.01%) and specificity (85.39%) were recorded, indicating robust predictive power.
• The five selected features demonstrated consistency across various predictive models.

Conclusions:

• The study presents an effective approach for CVD risk prediction with minimal feature requirements.
• The tool facilitates personalized patient interventions by highlighting specific high-risk factors.
• This approach supports shared decision-making between clinicians and patients through data-driven insights.