Atherosclerosis plaque tissue classification using self-attention-based conditional variational auto-encoder...

Area of Science:

• Cardiovascular Imaging
• Artificial Intelligence in Medicine
• Medical Image Analysis

Background:

• Coronary artery disease (CAD) involves atherosclerosis, characterized by plaque buildup in arterial walls.
• Optical Coherence Tomography (OCT) is a high-resolution imaging technique used by cardiologists to visualize intracoronary plaque.
• Manual interpretation of OCT images is time-consuming, subjective, and prone to inter-observer variability, hindering widespread adoption and increasing diagnostic errors.

Purpose of the Study:

• To develop an automated method for accurate classification of atherosclerosis plaque tissues from OCT images.
• To enhance the diagnostic utility and reduce errors associated with OCT-based plaque analysis.
• To introduce a novel deep learning framework for improved cardiovascular disease diagnosis.

Main Methods:

• A Self-Attention-Based Conditional Variational Auto-Encoder Generative Adversarial Network (APC-OCTPI-SACVAGAN) was developed for plaque tissue classification.
• The APC-OCTPI-SACVAGAN model was trained and executed in MATLAB.
• The method classifies OCT images into five categories: Fibro calcific plaque, Fibro atheroma, Thrombus, Fibrous plaque, and Micro-vessel.

Main Results:

• The APC-OCTPI-SACVAGAN method demonstrated superior performance compared to existing techniques.
• Achieved higher accuracy rates of 16.19%, 17.93%, 19.81%, and 1.57%.
• Showcased improved Area Under the Curve (AUC) values (16.92%, 11.54%, 5.29%, and 1.946%) and significantly reduced computational time (28.06%–39.185% lower).

Conclusions:

• Automated classification of atherosclerosis plaque using APC-OCTPI-SACVAGAN offers a significant advancement in OCT image analysis.
• This AI-driven approach has the potential to increase the clinical adoption of OCT and decrease diagnostic errors.
• The proposed method provides a robust and efficient solution for precise plaque characterization in cardiovascular imaging.