Aspects of three dimensional reconstruction for a multi ring positron tomograph

Area of Science:

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

Background:

• Inter-plane septa in multi-ring PET scanners reduce random and scattered coincidences.
• Septa typically limit coincidence detection to adjacent detector rings, excluding lines of response between non-adjacent rings.

Purpose of the Study:

• To evaluate the impact of removing septa on image quality and signal-to-noise ratio (SNR) in a clinical PET scanner.
• To assess the feasibility of 3D reconstruction with septa removed for enhanced data acquisition.

Main Methods:

• Utilized an 8-ring, 15-slice ECAT 931/08-12 scanner with a configuration option to remove septa.
• Acquired full 64 sinograms with septa removed, using a rotating pin source for detector normalization and transmission data.
• Reconstructed phantom studies using a fully three-dimensional (3D) algorithm and compared results with septa-in data.

Main Results:

• Signal-to-noise ratio (SNR) improved by a factor of 2.8 in the center of a uniform 20 cm diameter cylinder.
• SNR decreased in regions distant from the center axially due to increased scatter and random coincidences.
• An overall improvement in SNR was observed within 6 cm of the tomograph's 10 cm axial length.

Conclusions:

• Removing septa in PET imaging can significantly enhance SNR in central regions, particularly when using 3D reconstruction.
• The benefits of septa removal are spatially dependent, with potential trade-offs in axial regions due to increased scatter and randoms.
• Optimized 3D reconstruction strategies are crucial for maximizing the advantages of septa-less PET acquisition.