Comparison of several MRI markers revealed that a simple, generalizable midbrain linear measure and midbrain area accurately support routine differential diagnosis between progressive supranuclear palsy (PSP) and non-PSP parkinsonisms.

Background/Objectives: This study evaluated the reproducibility of brain morphometric measurements obtained with different head coils and acceleration factors on a PET/MR scanner, with the aim of supporting reliability in longitudinal neuroimaging studies following major scanner hardware upgrades and introduction of higher-channel-density coils for improved image quality and shorter acquisition times. Methods: Fifteen healthy subjects underwent MPRAGE imaging on a 3T PET/MR scanner using a 16 channel head/neck coil with acceleration factor 2 and a 32 channel PET-transparent head coil with acceleration factors 2 and 4. Cortical thickness and subcortical volumes were measured with FreeSurfer. We performed correlation analyses to test the association between the values of cortical thickness and subcortical volumes derived from MPRAGE images acquired with either the 16- or the 32 channel coil. Test–retest variability and intraclass correlation coefficient were calculated to assess the reproducibility of the measurements for each brain region generated by each sequence and coil. Results: Cortical thickness and subcortical volume measurements from images acquired with different coils and acceleration factors showed high correlation (R = 0.96–1.00, p < 0.001) between protocols, with median test–retest variability below 4% for cortical thickness and a structure-dependent pattern for subcortical volumes, with higher variability in smaller structures such as the accumbens, reaching approximately 7.9% in cross-coil comparisons. Conclusions: Overall, brain morphometry measurements were reproducible across acquisition conditions. Coils with a higher number of channels were associated with improved signal-to-noise ratio, shorter acquisition times, and maintained quantitative consistency.