A Review on Automatic Mammographic Density and Parenchymal Segmentation

Area of Science:

• Biomedical Imaging
• Radiology
• Computer-Aided Diagnosis

Background:

• Breast cancer is a leading cancer in women, with unknown causes.
• Early detection and risk assessment are crucial for management.
• Current risk models lack comprehensive integration of image-based factors like mammographic density.

Purpose of the Study:

• To review automatic mammographic tissue segmentation methodologies.
• To assess the evidence for risk assessment and density classification using segmentation.
• To analyze engineering progress and clinical impact of these methods.

Main Methods:

• Comprehensive literature review of automatic mammographic tissue segmentation techniques.
• Analysis of studies on risk assessment and density classification using segmentation.
• Evaluation of clinical translation and research gaps.

Main Results:

• Automatic segmentation methods have advanced significantly over two decades.
• Evidence supports the use of segmentation for mammographic risk assessment and density classification.
• Gaps exist in integrating image-based factors into non-image-based risk prediction models.

Conclusions:

• Automatic mammographic tissue segmentation is a progressing field with clinical potential.
• Further research is needed to fully incorporate image-based risk factors into comprehensive breast cancer risk models.
• Engineering advances are key to improving mammographic risk stratification and prevention strategies.