Coronary Plaque Characterization From Optical Coherence Tomography Imaging With a Two-Pathway Cascade Convolutional...

Area of Science:

• Cardiovascular Imaging
• Artificial Intelligence in Medicine
• Biomedical Engineering

Background:

• Coronary atherosclerotic plaque stability is determined by its morphology and tissue composition.
• Intravascular optical coherence tomography (OCT) enables plaque assessment but requires expert interpretation of large datasets.
• Accurate plaque characterization is crucial for predicting cardiovascular events.

Purpose of the Study:

• To develop an automated convolutional neural network (CNN) method for coronary plaque component characterization.
• To utilize OCT imaging for extracting tissue features (fibrous, lipid, calcification) from plaques.
• To enhance diagnostic efficiency and accuracy in coronary plaque analysis.

Main Methods:

• Development of a novel CNN architecture named TwopathCNN.
• Implementation of TwopathCNN in a cascaded structure for improved performance.
• Validation using in vivo OCT imaging data for plaque component classification.

Main Results:

• The proposed TwopathCNN method achieved an average F1-score of 0.86 and accuracy of 0.88.
• The TwopathCNN architecture and cascaded structure demonstrated significant performance improvements (p < 0.05) over conventional methods.
• The CNN approach showed higher efficiency compared to traditional machine learning techniques.

Conclusions:

• The developed CNN method effectively and robustly characterizes coronary plaque composition from OCT images.
• The TwopathCNN architecture in a cascaded setup offers superior performance for plaque characterization.
• This automated approach shows promise as an efficient diagnostic tool for coronary plaque detection.