VALIDATION OF A SIMULATION PROCEDURE FOR GENERATING BREAST TOMOSYNTHESIS PROJECTION IMAGES

Area of Science:

• Medical Imaging
• Radiological Physics
• Computational Imaging

Background:

• Optimizing diagnostic performance in breast tomosynthesis (BT) is crucial.
• Accurate simulation of BT imaging is essential for parameter evaluation.
• Existing simulation methods require validation against real-world data.

Purpose of the Study:

• To validate a simulation procedure for generating BT projection images.
• To compare contrast and sharpness between simulated and real BT images.
• To assess the fidelity of a Monte Carlo-based simulation for BT imaging.

Main Methods:

• Developed a simulation procedure using the Penelope Monte Carlo code.
• Modeled the geometry of a Siemens BT system.
• Acquired real BT images using the Siemens system for comparison.
• Quantitatively compared contrast and sharpness metrics between simulated and real images.

Main Results:

• Demonstrated good agreement in sharpness between simulated and real reconstructed image planes.
• Observed higher contrast in simulated projection images compared to real ones.
• Identified discrepancies in contrast that warrant further investigation.

Conclusions:

• The developed simulation procedure can generate BT images.
• Further modifications are needed to improve the realism of image noise and contrast in simulations.
• The simulation shows potential for BT system evaluation but requires refinement for contrast accuracy.