Intrascanner Reproducibility of an SPM-based Head MR-based Attenuation Correction Method

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiology

Background:

• Current MR-based attenuation correction (AC) methods show comparable accuracy to CT-based AC.
• Robustness, specifically intrascanner reproducibility, is a critical metric for clinical applicability but often unquantified.
• The SPM-based pseudo-CT approach is a promising MR-based AC method for brain PET/MRI.

Purpose of the Study:

• To evaluate the intrascanner reproducibility (repeatability) of the SPM-based pseudo-CT approach for brain MR-based attenuation correction.
• To quantify the robustness of this MR-AC method using established statistical metrics.

Main Methods:

• Retrospective analysis of 22 subjects with 3 serial 18F-FDG PET/MRI visits within 2 months.
• Generation of pseudo-CT mu-maps from MR images for each visit.
• Reconstruction of PET data using mu-maps from all three visits.
• Assessment of repeatability using relative changes (RC), Intraclass Correlation Coefficient (ICC), Reproducibility Coefficient (RDC95%), and Bland-Altman Limits of Agreement (LoA).

Main Results:

• Absolute relative changes in reconstructed PET images were within ~2%.
• Brain cortex and cerebellum showed the largest variability at ~3%.
• Differences in reconstructed PET images across visits were not statistically significant (p=0.90).

Conclusions:

• This study provides the first assessment of the repeatability of the SPM-based pseudo-CT approach for brain MR-AC.
• The method demonstrates excellent repeatability and robustness, alongside previously shown quality and ease of implementation.
• The SPM-based pseudo-CT approach is well-suited for routine brain PET/MRI research and clinical studies.