Reduction of collimator correction artefacts with bayesian reconstruction in spect

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Single photon emission computed tomography (SPECT) imaging quality is often limited by poor resolution.
• Collimator correction improves SPECT resolution but can introduce ringing artefacts, degrading image quality.

Purpose of the Study:

• To investigate the efficacy of Bayesian reconstruction methods in reducing collimator correction artefacts in SPECT.
• To compare the performance of different Bayesian methods against standard reconstruction.

Main Methods:

• Applied and tested three Bayesian reconstruction methods: smoothing prior, median root prior, and anatomical prior.
• Evaluated physical and visual performance using phantom and patient studies.

Main Results:

• All three Bayesian methods successfully reduced collimator correction artefacts.
• Smoothing and median root priors resulted in slightly lower contrast compared to standard reconstruction.
• The anatomical prior method yielded images with good resolution and contrast.

Conclusions:

• Bayesian reconstruction, particularly the anatomical prior method, offers a promising approach to enhance SPECT image quality by mitigating artefacts while preserving resolution and contrast.