Measurement of T2* in the human spinal cord at 3T

Abstract

Purpose: To measure the transverse relaxation time T₂* in healthy human cervical spinal cord gray matter (GM) and white matter (WM) at 3T.

Methods: Thirty healthy volunteers were recruited. Axial images were acquired using an averaged multi-echo gradient-echo (mFFE) T₂*-weighted sequence with 5 echoes. We used the signal equation for an mFFE sequence with constant dephasing gradients after each echo to jointly estimate the spin density and T₂* for each voxel.

Results: No global difference in T₂* was observed between all GM (41.3 ± 5.6 ms) and all WM (39.8 ± 5.4 ms). No significant differences were observed between left (43.2 ± 6.8 ms) and right (43.4 ± 5.5 ms) ventral GM, left (38.3 ± 6.1 ms) and right (38.6 ± 6.5 ms) dorsal GM, and left (39.4 ± 5.8 ms) and right (40.3 ± 5.8 ms) lateral WM. However, significant regional differences were observed between ventral (43.4 ± 5.7 ms) and dorsal (38.4 ± 6.0 ms) GM (p < 0.05), as well as between ventral (42.9 ± 6.5 ms) and dorsal (37.9 ± 6.2 ms) WM (p < 0.05). In analyses across slices, inferior T₂* was longer than superior T₂* in GM (44.7 ms vs. 40.1 ms; p < 0.01) and in WM (41.8 ms vs. 35.9 ms; p < 0.01).

Conclusions: Significant differences in T₂* are observed between ventral and dorsal GM, ventral and dorsal WM, and superior and inferior GM and WM. There is no evidence for bilateral asymmetry in T₂* in the healthy cord. These values of T₂* in the spinal cord are notably lower than most reported values of T₂* in the cortex.

Keywords: 3 Tesla; T2* mapping; healthy controls; multi-echo gradient-echo imaging; relaxometry; spinal cord.

FIGURE 1

Imaging and data analysis from one subject. (A) Mid-sagittal slice showing the axial placement of 12 5-mm slices. (B) Five gradient echoes from a mid-C3 slice with TEs = 7.2, 16.0, 24.9, 33.7, and 42.5 ms, respectively. (C) Combined mFFE image of the same slice, showing excellent contrast between gray matter, white matter, and cerebrospinal fluid. (D) Spatial map of the estimated spin density, S₀, in arbitrary units of intensity. In this subject, mean S₀ in gray matter across slices is 18.3% higher than in white matter (13815±961 vs. 11677±879). (E) Spatial map of the estimated transverse relaxation time, T₂*, in ms. High values of S₀ and T₂* around the edge of the cord are due to partial volume effects with cerebrospinal fluid.

FIGURE 2

Outlines of spinal cord gray and white matter across vertebral levels in one subject, and the corresponding eroded gray and white matter ROIs (in red and cyan, respectively) used in the T₂* calculations. Gray matter masks are eroded to minimize partial volume effects with adjacent white matter, and white matter ROIs are similarly constructed to minimize partial volume effects with both gray matter and cerebrospinal fluid.