Machine Learning Algorithm to Predict Obstructive Coronary Artery Disease: Insights from the CorLipid Trial.

Area of Science:

• Cardiovascular Medicine
• Biomarkers
• Machine Learning

Background:

• Predicting coronary artery disease (CAD) severity remains a clinical challenge.
• Accurate risk assessment is crucial for timely intervention and patient management.

Purpose of the Study:

• To develop and validate a machine learning (ML) predictive algorithm for determining obstructive CAD severity.
• To integrate clinical characteristics with novel metabolic biomarkers for enhanced risk prediction.

Main Methods:

• Developed an XGBoost ML algorithm using 73 biochemical, metabolic, anthropometric, and medical history variables.
• Analyzed serum levels of ceramides, acyl-carnitines, fatty acids, galectin-3, adiponectin, and APOB/APOA1 ratio.
• Validated the algorithm on 958 patients from the CorLipid trial, categorizing them by obstructive (SYNTAX score > 0) and non-obstructive CAD (SYNTAX score = 0).

Main Results:

• The ML model achieved an Area Under the Curve (AUC) of 0.725 (95% CI: 0.691−0.759).
• The algorithm successfully identified patients with obstructive CAD.
• Incorporation of metabolic features improved predictive capabilities.

Conclusions:

• A machine learning model combining clinical and metabolic features can effectively estimate the pre-test likelihood of obstructive CAD.
• This approach offers a promising tool for non-invasive CAD risk stratification.
• Further validation in diverse populations is warranted.