Coronary plaque phenotype associated with positive remodeling

Area of Science:

• Cardiovascular Imaging
• Artificial Intelligence in Medicine
• Atherosclerosis Research

Background:

• Positive remodeling is a key vascular adaptation in atherosclerosis.
• Coronary computed tomography angiography (CTA) can detect positive remodeling.

Purpose of the Study:

• To identify plaque features associated with positive remodeling using machine learning.
• To assess the prognostic value of positive remodeling in cardiovascular events.

Main Methods:

• Analysis of 426 patients with coronary CTA and optical coherence tomography (OCT).
• Utilized four machine learning models (GBM, RF, DL, SVM) to analyze 15 plaque features.
• Variable importance ranking identified key features linked to positive remodeling.

Main Results:

• Lipid index and maximal calcification arc were consistently important features across ML models.
• These features correlated with positive remodeling, with greater influence at lower ranges.
• Increased positive remodeling plaques were linked to higher LDL cholesterol and a 2.10-fold increased risk of cardiovascular events over 5 years.

Conclusions:

• Significant lipid accumulation and minimal calcification are crucial for positive remodeling.
• The extent of positive remodeling predicts future cardiovascular events.