A depth-of-interaction rebinning method based on both geometric and activity weights

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Positron Emission Tomography (PET)

Background:

• Depth-of-interaction (DOI) measurement is crucial for PET scanner image quality.
• Current DOI rebinning methods map events based on line of response (LOR) correlation.
• There is a need to improve DOI rebinning by integrating object-specific information.

Purpose of the Study:

• To develop a novel DOI rebinning method for PET scanners.
• To enhance image quality by incorporating prior radioactivity distribution.
• To improve the efficiency of PET image reconstruction.

Main Methods:

• A DOI rebinning method utilizing both geometric and activity weights was proposed.
• Geometric weights were calculated based on LOR intersection areas.
• Activity weights were derived from prior reconstructed images to reflect radioactivity distribution.

Main Results:

• The proposed method demonstrated superior performance in Monte Carlo simulations across multiple phantoms.
• Image quality metrics including recovery coefficient (RC), contrast recovery coefficient (CRC), SSIM, and PSNR were significantly improved.
• The method achieved a maximum 32% improvement in RC and enhanced CRC, SSIM, and PSNR compared to existing methods.
• Incorporating prior activity distribution also accelerated image reconstruction speed.

Conclusions:

• A novel DOI rebinning method was developed, combining LOR correlation with prior activity distribution.
• The method achieved optimal image quality and reconstruction speed in simulations.
• Further evaluation on actual PET equipment is required to validate the findings.