Revealing the nature of cardiovascular disease using DERGA, a novel data ensemble refinement greedy algorithm

Area of Science:

• Cardiology
• Artificial Intelligence
• Biomedical Data Science

Background:

• Cardiovascular disease (CVD) poses a significant global health challenge.
• Accurate prediction of CVD risk is essential for early intervention and management.
• Identifying key predictive parameters is crucial for developing effective diagnostic tools.

Purpose of the Study:

• To determine the most crucial parameters associated with cardiovascular disease (CVD).
• To employ a novel data ensemble refinement procedure to uncover the optimal pattern of these parameters for high prediction accuracy.
• To enhance early diagnosis and optimize resource utilization in CVD risk assessment.

Main Methods:

• Analysis of data from 369 patients (281 with CVD or at risk, 88 healthy).
• Utilized an artificial intelligence-based algorithm, including DERGA and Extra Trees, for pattern identification.
• Evaluated the predictive power of various parameters, including Platelet-derived Microvesicles (PMV), hypertension, age, smoking, dyslipidemia, and Body Mass Index (BMI).

Main Results:

• A six-parameter combination, including PMV, hypertension, age, smoking, dyslipidemia, and BMI, effectively discerns CVD likelihood.
• The highest prediction accuracy achieved was 98.64%.
• Platelet-derived Microvesicles (PMV) alone yielded 91.32% accuracy, while a ten-parameter model achieved 97.83% accuracy.

Conclusions:

• The DERGA algorithm effectively identifies critical parameters for CVD risk assessment.
• This approach accelerates CVD risk assessment, enabling earlier diagnosis and reducing the need for extensive lab tests.
• The findings enhance understanding of CVD susceptibility and optimize resource allocation in healthcare.