Classification of breast computed tomography data.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Breast tissue composition is crucial for risk assessment, disease detection, and monitoring changes.
• Accurate differentiation of breast tissues (skin, fat, glandular) is essential for quantitative analysis.

Purpose of the Study:

• To develop and validate an algorithm for classifying breast tissue into skin, fat, and glandular components using dedicated breast CT data.
• To enable a more quantitative analysis of breast composition and glandular tissue patterns.

Main Methods:

• A dedicated breast CT scanner was used to acquire data from 55 volunteers and patients.
• Breast CT voxel data underwent median filtering and histogram analysis with a two-compartment Gaussian fit.
• Region-growing algorithms classified tissues based on histogram values and architectural features.

Main Results:

• The algorithm demonstrated excellent agreement (97.7%) with radiologist segmentation.
• Analysis revealed glandular tissue fraction rarely exceeded 50%, even in dense breasts.
• Most individuals exhibit a 70% fat-30% glandular composition, rather than 50%-50%.

Conclusions:

• The developed algorithm provides accurate breast tissue classification from CT data.
• This method enhances quantitative breast composition analysis and understanding of tissue distribution.
• Findings suggest a common fat-dominant composition in most individuals, impacting risk assessment.