Simulation study of a high-performance brain PET system with dodecahedral geometry.

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Biomedical Engineering

Background:

• Spherical PET systems offer optimal sensitivity but are impractical for brain imaging.
• Dedicated brain PET scanners are crucial for neurological disease diagnosis.
• Existing systems face limitations in sensitivity, resolution, or manufacturability.

Purpose of the Study:

• To design and simulate a novel dodecahedral PET scanner as a practical alternative to spherical systems.
• To evaluate the imaging performance, including sensitivity and spatial resolution, of the proposed dodecahedral scanner.
• To compare the dodecahedral PET system with other dedicated brain PET systems.

Main Methods:

• Monte Carlo simulations were used to generate data for the dodecahedral PET system (11 regular pentagon detectors).
• Image reconstruction was performed using a fully three-dimensional maximum-likelihood expectation maximization (3D-MLEM) algorithm.
• Performance was assessed using metrics like sensitivity, spatial resolution, image contrast, and noise levels, including simulations with the Hoffman phantom.

Main Results:

• The dodecahedral PET system demonstrated high sensitivity across the entire field of view.
• Achieved a spatial resolution of 1.98 mm with a depth-of-interaction resolution of approximately 6.67 mm.
• Showed improved image contrast and reduced noise compared to other dedicated brain PET systems, indicating clinical potential.

Conclusions:

• The proposed dodecahedral PET scanner is a promising technology for high-sensitivity and high-resolution brain imaging.
• This design offers a feasible manufacturing approach with potential for widespread clinical application.
• The system's performance suggests it could enable lower injected radiation doses in Positron Emission Tomography scans.