Count rate performance of brain-dedicated PET scanners: a Monte Carlo simulation study

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Biomedical Engineering

Background:

• Renewed interest in brain-dedicated Positron Emission Tomography (PET) systems, especially for combined PET/Magnetic Resonance Imaging (MRI).
• Design of a brain-dedicated PET insert for the Siemens Magnetom 7T ultra-high field MR scanner.

Purpose of the Study:

• Investigate the count rate performance of various detector configurations for a brain-dedicated PET scanner.
• Evaluate the impact of detector geometry, deadtime, coincidence timing, and discriminator settings on performance.
• Analyze the influence of activity outside the field-of-view (FOV) on scanner performance.

Main Methods:

• Monte Carlo simulations were employed to model PET scanner detectors.
• Simulated detectors used lutetium oxyorthosilicate (LSO) scintillator with varying areas (0.04 to 101.37 cm²), thicknesses (10-20 mm), and crystal pitch (~2 mm).
• Evaluated scanner singles rate, sensitivity, and noise equivalent count rate (NECR) as a function of activity.

Main Results:

• Scanners with few crystal elements exhibited reduced sensitivity due to inter-detector scattering.
• Count rate performance of scanners with large detectors was primarily limited by detector deadtime.
• A model predicting scanner count rate performance based on detector characteristics was developed.

Conclusions:

• Detector design critically influences the count rate performance of brain-dedicated PET scanners.
• Trade-offs exist between detector size, sensitivity, and deadtime limitations.
• The developed model aids in optimizing PET insert design for combined PET/MR systems.