An anthropomorphic breast model for breast imaging simulation and optimization

Area of Science:

• Medical Imaging Physics
• Computational Biology
• Radiological Sciences

Background:

• Computer simulations for x-ray breast imaging require realistic phantoms to model patient anatomical variability.
• Existing phantoms may not adequately capture the diverse anatomical features of different patients.

Purpose of the Study:

• To develop a three-dimensional (3D) anthropomorphic breast phantom with realistic and randomizable anatomical features.
• To create a versatile platform for optimizing x-ray-based breast imaging systems through computer simulation.

Main Methods:

• A voxelized breast model was created with skin, fat, glandular tissue, ductal trees, masses, and microcalcifications.
• Randomized breast morphology and compression models were incorporated.
• Monte Carlo (MC) simulation code was adapted for x-ray imaging systems.
• Simulated projections were reconstructed for mammography, tomosynthesis, and computed tomography (CT), with dose maps generated.

Main Results:

• Simulated mammograms closely matched real mammograms in appearance and features.
• The phantom demonstrated suitable properties for anatomical backgrounds in noise-power spectra.
• Reconstructed tomosynthesis and CT images, along with dose maps, provided data for optimization studies.
• Dosimetry results revealed differences in dose distribution across modalities and compression levels.

Conclusions:

• The developed anthropomorphic breast phantom and MC simulation platform provide a realistic model for breast imaging systems.
• This integrated framework is expected to be a powerful tool for optimizing volumetric breast imaging systems.