VOLUMETRIC LOCALISATION OF DENSE BREAST TISSUE USING BREAST TOMOSYNTHESIS DATA

Area of Science:

• Medical Imaging
• Biophysics
• Computational Biology

Background:

• Accurate breast tissue composition analysis is crucial for mammography and cancer detection.
• Distinguishing between dense fibroglandular tissue and fatty adipose tissue remains a challenge in breast imaging.

Purpose of the Study:

• To develop and validate a method for localizing dense breast tissue using combined digital breast tomosynthesis (DBT) and projection imaging data.
• To assess the accuracy of the proposed method in determining breast tissue composition.

Main Methods:

• Utilized reconstructed digital breast tomosynthesis (DBT) volumes and density estimation from projection images.
• Employed software breast phantoms generated with fractal Perlin noise for verification.
• Used the PENELOPE Monte Carlo package for creating projection images and estimated dense tissue volume from central projection images.

Main Results:

• The developed method accurately determined the composition of 75±5% of voxels.
• Combined DBT and projection data effectively localized dense tissue within simulated breast phantoms.
• Density images guided the placement of dense voxels using DBT images as templates.

Conclusions:

• The combined approach of DBT and projection imaging offers a promising method for accurate breast tissue composition analysis.
• This technique has the potential to improve the interpretation of mammographic images and enhance early cancer detection.
• Further validation with clinical data is warranted to confirm the efficacy of this method in real-world scenarios.