The effect of regularization in motion compensated PET image reconstruction: a realistic numerical 4D simulation...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Respiratory motion significantly impacts PET image quality and diagnostic accuracy.
• Reconstruct-transform-average (RTA) and motion-compensated image reconstruction (MCIR) are common PET motion correction methods.
• Theoretical advantages of MCIR over RTA exist, but unregularized comparisons show mixed results regarding noise and accuracy.

Purpose of the Study:

• To compare the performance of regularized RTA and MCIR for respiratory motion correction in PET imaging.
• To evaluate the impact of regularization on image quality, bias, and root mean square error.
• To assess the effectiveness of these methods across various noise levels, lesion sizes, and iterations.

Main Methods:

• A realistic numerical 4D simulation study was conducted.
• The median-root-prior incorporated in the ordered subsets maximum a posteriori one-step-late algorithm was used for regularization.
• Performance was evaluated based on bias, root mean square error, contrast-to-noise ratio (CNR), and standard deviation.

Main Results:

• Regularized MCIR demonstrated reduced bias and root mean square error in lesion reconstruction compared to regularized RTA.
• Both methods achieved similar contrast-to-noise ratio and standard deviation with appropriate regularization.
• These findings were consistent across different noise levels, lesion sizes, and iteration counts.

Conclusions:

• Motion-compensated image reconstruction (MCIR) with appropriate regularization parameters offers superior lesion accuracy in PET imaging.
• Regularized reconstruct-transform-average (RTA) is a viable alternative for motion compensation, effectively reducing bias and mean square error.
• The choice between regularized MCIR and RTA may depend on specific clinical and technical requirements.