Symptomatic and Asymptomatic Carotid Plaques Classification using CT Images and Hybrid Deep Transfer Learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Cardiovascular Research

Background:

• Carotid plaques are a major cause of stroke.
• Accurate classification of plaque symptoms is crucial for risk stratification.
• Current methods for plaque analysis have limitations.

Purpose of the Study:

• To develop and validate a data-driven model for classifying symptomatic and asymptomatic carotid plaques.
• To leverage deep transfer learning for enhanced feature extraction from Computed Tomography (CT) images.
• To improve the accuracy of carotid plaque characterization for clinical decision-making.

Main Methods:

• A hybrid deep transfer learning framework combining Convolutional Neural Network (CNN) architectures was employed for feature extraction.
• Features capturing local/global textures and morphological characteristics of the carotid artery were extracted.
• Extracted features were utilized as inputs for various machine learning models and evaluated on real-world data.

Main Results:

• The proposed model demonstrated feasibility and effectiveness in classifying carotid plaques.
• The deep transfer learning approach successfully captured complex plaque characteristics.
• The evaluated machine learning models showed promising performance in plaque classification.

Conclusions:

• The developed model shows significant potential for improving clinical diagnosis of carotid artery disease.
• This AI-driven approach can enhance the assessment of stroke risk associated with carotid plaques.
• Further research can explore the integration of this model into routine clinical workflows.