Quantitative analysis of breast parenchymal patterns using 3D fibroglandular tissues segmented based on MRI

Area of Science:

• Radiology and Medical Imaging
• Oncology
• Biomedical Engineering

Background:

• Mammographic density and breast parenchymal patterns are linked to breast cancer risk.
• Quantitative analysis of parenchymal patterns has been limited due to a lack of reliable parameters.
• Three-dimensional breast MRI offers a solution to tissue overlapping issues in analysis.

Purpose of the Study:

• To analyze fibroglandular tissue morphology using 3D breast MRI.
• To identify quantitative parameters for differentiating breast parenchymal patterns.
• To explore the potential of these parameters in breast cancer risk assessment.

Main Methods:

• Analyzed four morphological parameters (circularity, convexity, irregularity, compactness) in 230 patients using 3D breast MRI.
• Classified patients into Intermingled (Type I) and Central (Type C) breast patterns.
• Used t-tests, histograms, and ROC analysis to differentiate between patterns based on parameters.

Main Results:

• All four morphological parameters significantly differentiated between Type I and Type C breast patterns.
• Compactness achieved the highest Area Under the Curve (AUC) of 0.84 in ROC analysis.
• Combining all four parameters increased the AUC to 0.94, indicating strong differentiation.

Conclusions:

• Morphological parameters from 3D MRI can quantitatively distinguish between intermingled and central breast tissue patterns.
• These quantitative parameters may serve as valuable tools for characterizing breast parenchymal patterns.
• This approach could facilitate further research into the relationship between parenchymal patterns and breast cancer risk.