An assessment of intra-scanner reproducibility of myelin-sensitive MRI measures
AUTHORS: Rahmanzadeh R, Weigel M, Lu PJ, Melie-Garcia L, Nguyen TD, La Rosa F, Wang Y, Bach Cuadra M, Radue EW, Gaetano L, Magon S, Kappos L, Kuhle J, Granziera C
ECTRIMS 2021 - Multiple Sclerosis Journal, 27(2_SUPPL): 430-431, October 2021
Background: Myelin-sensitive quantitative MRI techniques(qMRIs) provide more specific measures of demyelination and remyelination than conventional MRI. However, there are currently few studies assessing their relative reproducibility in clinical scanners.
Objectives: To assess and to compare the test–retest reproducibility of myelin-sensitive qMRIs acquired within and across scanning sessions in a single 3T MR system.
Methods: In a 3T Prisma MR system, ten healthy subjects HS) underwent (i) myelin water imaging(MWI) using Fast Acquisition with Spiral Trajectory and Adiabatic T2prep,(ii) magnetization transfer saturation(MTsat) imaging, (iii) quantitative T1 mapping(qT1) using MP2RAGE. Two scans were conducted 1h apart without repositioning(S1 and S2) and a third scan(S3) was performed 1 week later(± 3 days).
In five regions of interest (ROIs: genu and splenium of the corpus callosum(CC), putamen, head of caudate and thalamus), we calculated the intra-class correlation coefficient(ICC) for each qMRI using a two-way mixed model with consistency of agreement within and across sessions (S1-S2, S2-S3).
White matter (WM) voxel-wise comparisons of qMRI measures for S1 vs S2 and S2 vs S3 were performed using Threshold-Free Cluster Enhancement(TFCE) clustering (P<0.05). The same analysis for cortex using a vertex-wise approach was performed using FreeSurfer and compared using a General Linear model(GLM) (P<0.05).
Results: Three HS were excluded from further analyses due to motion artefacts. In the remaining 7 HS: qT1 ICC was the highest, followed by MWF and MTsat(qT1- S1-S2: 0.95, S2-S3: 0.91; MWF- S1-S2: 0.92, S2-S3: 0.87; MTsat- S1-S2: 0.85, S2-S3: 0.82). The best reproducible measure was qT1 in the CC genu(S1-S2: 0.93, S2-S3: 0.95) and the one with lowest ICC was MTsat in the caudate head(S1-S2: 0.83, S2-S3: 0.74).
For all qMRI measures, WM voxel-wise analysis exhibited no clusters of significant differences for S1 vs S2 and S2 vs S3. Cortical surface analysis showed very small areas of significant differences in S1 vs S2 and in S2 vs S3, which were slightly more pronounced in qT1, followed by MTsat and MWF.
Conclusion: qT1 from MP2RAGE, MWF and MTsat have high within scanner reproducibility, both with and without repositioning. qT1 appeared to be the most reproducible measure in ROI analyses, whereas MWF was the most reproducible in cortical surface analyses. Also, all qMRI measures showed high intra and inter-session reproducibility in voxel-wise WM comparisons.