Accurate modeling of a DOI capable small animal PET scanner using GATE

Area of Science:

• Medical Physics
• Nuclear Medicine
• Computational Imaging

Background:

• Pre-clinical Positron Emission Tomography (PET) scanners require accurate simulation models for research and development.
• The Sedecal Argus scanner utilizes dual-layer detectors (LYSO and GSO) for Depth of Interaction (DOI) compensation.

Purpose of the Study:

• To develop and validate a Monte Carlo (MC) model of the Sedecal Argus pre-clinical PET scanner using GATE.
• To assess the model's accuracy in predicting scanner performance metrics and image quality.

Main Methods:

• Developed an MC model in GATE, incorporating detector geometry, pulse readout, and coincidence event selection.
• Created separate code for emulating digitized data processing, including time windows and data flow.
• Validated the model against experimental measurements using various phantoms, including the NEMA NU-4 phantom.

Main Results:

• The MC model demonstrated good agreement with experimental data for spatial resolution, sensitivity, and scatter fraction (within 7%).
• Count rate curves and NEMA phantom image analyses (noise, contrast, recovery coefficient) showed agreement within 9%.
• The model accurately reproduced the energy spectrum of coincidence events, including contributions from LYSO crystals.

Conclusions:

• The developed MC model, built on GATE and extended with custom code, provides an accurate simulation of the Sedecal Argus PET scanner.
• The validated simulation setup is a valuable tool for diverse research applications in pre-clinical PET imaging.
• This work confirms the feasibility of MC modeling for complex PET systems with high fidelity.