An automatic multi-class coronary atherosclerosis plaque detection and classification framework.

Area of Science:

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

Background:

• Accurate detection and classification of atherosclerotic plaques are crucial for managing coronary artery disease.
• Existing methods often fail to differentiate between various plaque types, limiting early intervention strategies.

Purpose of the Study:

• To develop an automated framework for multi-class detection and classification of coronary atherosclerosis plaques.
• To improve upon existing methods by distinguishing between non-calcified, calcified, and mixed plaque types.

Main Methods:

• Utilized computed tomography angiography (CTA) to extract transverse cross-sections.
• Developed a novel random radius symmetry (RRS) feature vector for data augmentation.
• Implemented a multi-class coronary plaque classifier using RRS features and support vector machines (SVM).

Main Results:

• The RRS feature vector combined with SVM achieved high performance on the Rotterdam Coronary Datasets.
• Achieved an average precision of 92.6% ± 1.9% and an average recall of 94.3% ± 2.1% in classifying plaque types.
• Outperformed traditional intensity features and random forest classifiers in plaque classification tasks.

Conclusions:

• The proposed framework offers a robust computer-aided diagnostic tool for multi-class coronary plaque detection and classification.
• This advancement can significantly aid in the early diagnosis and intervention of coronary artery diseases.
• The RRS feature extraction method shows promise for enhancing medical image analysis in cardiology.