Towards component-based validation of GATE: aspects of the coincidence processor

Area of Science:

• Medical Imaging
• Computational Physics

Background:

• GATE is widely used for Monte Carlo simulations in emission tomography.
• Previous GATE validations focused on whole systems, potentially masking sub-system errors.

Purpose of the Study:

• To assess the accuracy of the GATE PET coincidence generation sub-system in isolation.
• To evaluate GATE's performance against independent generators and ground truth data.

Main Methods:

• Modeled a Siemens mCT scanner in GATE.
• Developed independent coincidence generators at Toshiba Medical Research Unit (TMRU) and UC Davis.
• Compared true, scattered, and random coincidences generated by GATE (using multiple and single window methods) with TMRU, UC Davis, and ground truth data.

Main Results:

• TMRU and UC Davis generators demonstrated strong agreement with ground truth.
• GATE achieved reasonable accuracy with the 'single window method' and without the 'delayed coincidence option' for randoms estimation.
• Other GATE parameter combinations led to substantial inaccuracies.

Conclusions:

• The GATE PET coincidence generator can be accurate when configured appropriately.
• Careful selection of coincidence detection methods and random coincidence estimation is crucial for reliable GATE simulations.
• Independent validation of sub-systems is essential for ensuring the overall accuracy of simulation software.