Repeatability of Total Sodium Concentration In Brain White Matter Measured with 3T MRI Using a 3D Spiral K-Space Trajectory
Abstract
Purpose
To assess the repeatability of brain white matter sodium concentration (TSC) measurements using 3T sodium MRI with the Fermat Looped Orthogonally Encoded Trajectories (FLORET) 3D spiral.
Methods
Sodium MRI requires ultra-short echo times (TE) due to rapid T₂ relaxation. A FLORET 3D spiral acquisition was used to achieve ultra-short TE with high signal-to-noise (SNR) efficiency. Data acquisition: Fifteen healthy participants were scanned twice (mean interval: 7 days) on a 3T MRI system. T1 weighted (T1w) anatomical proton images and sodium FLORET images (TE=0.193ms, repetition time=130ms, isotropic resolution=4.75mm3, 12 signal acquisitions, and one noise-only acquisition) were acquired. Four calibration phantoms containing 30, 60, 90, and 120mM of sodium were included in the field of view. Data analysis: Sodium images were linearly registered with each other and averaged. Proton T1w images were then non-linearly registered to the averaged sodium images. TSC maps were created using linear fit of phantom signal intensities. White matter masks were segmented from proton T1w were mapped to TSC maps. Statistical analysis: Coefficient of variation (CoV) and Bland–Altman analyses assessed repeatability of white matter TSC.
Results
Data from one participant was excluded due to poor image registration. FLORET sodium images demonstrated a mean white matter SNR of 3.2±0.5. The overall average white matter TSC was 30.0±4.0mM (scan 1: 29.6±4.3mM, scan 2: 30.5±3.8mM). The average CoV was 7.5%, and Bland-Altman analysis revealed that TSC differences between scans were symmetrically distributed about zero.
Conclusion
This study was the first to evaluate FLORET-based TSC repeatability, demonstrating an average CoV of 7.5%. Measured TSC values (30.0±4.0mM) were consistent with a prior meta-analysis that obtained a pooled mean of 37.3mM using a variety of acquisition methods. Reliable sodium MRI techniques are needed to detect altered TSC as a biomarker of neuronal viability in diseases such as multiple sclerosis, cancer, and stroke.