Power-law Beta Calculations for Realistic Mammography Images from Egs_cbct Simulation.

Area of Science:

• Medical Imaging
• Computational Physics
• Biophysics

Background:

• Mammography is crucial for early breast cancer detection.
• Glandular tissue in mammograms creates anatomic noise, potentially obscuring lesions.
• Previous research characterized this noise using a power-law form with β-exponent values between 1.5 and 4.0.

Purpose of the Study:

• To validate the egs_cbct simulation code's ability to replicate mammography images.
• To quantitatively assess simulated mammography images by extracting β-values and comparing them to clinical data.

Main Methods:

• Breast phantoms were modeled using in silico imaging software for the Virtual Clinical Trial for Regulatory Evaluation project.
• The egs_cbct code simulated 400 mammography images of these phantoms.
• Anatomical noise power spectrum analysis was performed on 200 regions of interest per image.

Main Results:

• The mean β-value for simulated images was 3.42 (range: 2.36–4.41), consistent with clinical data.
• The simulated American College of Radiology (ACR) phantom showed comparable image quality to a real mammography unit.
• The extracted β-values from simulated images align with previously reported values from authentic mammograms.

Conclusions:

• The egs_cbct software successfully simulates mammography images, as evidenced by the β-value analysis.
• This validates the use of egs_cbct for generating realistic mammograms in virtual trials.
• The findings support the use of egs_cbct in regulatory evaluation for mammography technologies.