Background/Objectives: Spinal diffusion tensor imaging (sDTI) remains a challenging method for the selective evaluation of key anatomical structures, like pyramidal tracts (PTs) and dorsal columns (DCs), and for reliably quantifying diffusion metrics such as fractional anisotropy (FA), radial diffusivity (RD), mean diffusivity (MD), and axial diffusivity (AD). This prospective, single-center study aimed to assess the reproducibility, robustness, and reliability of an optimized axial sDTI protocol, specifically intended for long fiber tracts. Methods: We developed an optimized Stejskal–Tanner sequence for high-resolution, axial sDTI of the cervical spinal cord at 3.0 T. Using advanced standardized evaluation and post-processing methods, we estimated DTI values for PTs, DCs, and AHs at the level of the second cervical vertebra. Reliability was evaluated through repeated measurements in 16 healthy volunteers and by comparing results from two 3.0 T scanners (Magnetom Skyra and Magnetom Prisma, Siemens Healthineers, Erlangen, Germany). Reproducibility was assessed using paired t-tests, intraclass correlation coefficients (ICCs), Bland–Altman analysis, and coefficients of variation (CVs). Results: The optimized sDTI protocol demonstrated high consistency for FA between test–retest sessions and across scanners. For the Skyra, the DC region showed the highest reliability (average ICC = 0.858) followed by the PT region (average ICC = 0.789). On the Prisma, the PT region reached an average ICC of 0.854, with the DC region at 0.758. Pooled inter-scanner data indicated good-to-excellent agreement, particularly in the PT region (average ICC = 0.860). FA CVs remained low (<10%) across all regions and scanners. RD showed good-to-excellent ICC values for PTs and DCs (average ICC for Skyra 0.642 and 0.769 and 0.926 and 0.830 for Prisma, respectively) but showed a higher CV between 14.6 and 19.4% for these two scanners. Conclusions: Improved sDTI offers highly reproducible FA measurements for all metrics with scanner independence, supporting its potential as a robust tool for detecting and monitoring spinal cord pathologies.
Ruff et al. (Sat,) studied this question.