MRI of the thoracic spinal cord in multiple sclerosis at 7T
Jennifer Lefeuvre1,2, Qi Duan1, Jacco A de Zwart1, Peter van Gelderen1, St├ęphane Lehericy2, Steven Jacobson1, Daniel S Reich1, and Govind Nair1

1NINDS, NIH, Bethesda, MD, United States, 2INSERM U1127/CNRS UMR7225, CENIR, Brain and Spine Institute, Paris, France


MRI of the thoracic spinal cord (t-spine) is challenging especially due to motion, flow, and susceptibility artifacts. Improving the image quality and resolution in t-spine has the potential to improve visualization of anatomy, as well as lesions in neurodegenerative diseases such as multiple sclerosis (MS). Towards this end, we have developed high-resolution t-spine imaging techniques at 7T with a custom built RF transmit-receive coil and navigator based frequency correction.


Imaging the t-spine at high field in neurodegenerative disease


MRI of the thoracic spinal cord (t-spine) is challenging not only because of the small cross-sectional area compared to its length, but also due to motion, flow, and susceptibility artifacts from surrounding tissue and cerebrospinal fluid (CSF). While patients with a variety of central nervous system diseases, such as multiple sclerosis (MS) and HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP), show clinical symptoms originating in the t-spine, routine imaging studies generally underestimate the extent of corresponding damage, particularly when it takes the form of focal lesions. This may be a factor in the well described clinico-radiological paradox, wherein clinical disability cannot be completely explained by the imaging study1. Improving the image quality and resolution in t-spine has the potential to improve lesion visualization in these diseases. We have previously reported the design of a spine array coil with an electric dipole transmitter and 8 receive channels2. Here we report its application in imaging persons with MS.


T-spine was imaged in one healthy volunteer (25 y.o. M) and two persons clinically diagnosed with relapsing-remitting MS (30 y.o., M and 39 y.o., F) on 3T (Siemens or Philips) and 7T MRI systems (Siemens) for comparison. At 3T, t-spine was imaged using the spine array coil (product coil) with optimized protocols3 such as T1-MPRAGE in the sagittal plane (Siemens: 3D TFL, TR/TE/TI/FA: 3000/6.8/900 ms/9°, 1 mm isotropic) and T2*-GRE (Siemens: TR/TE/FA = 775/11 ms/20°, in-plane resolution 0.58 mm or Philips: 2D-GRE, TR/TE/FA = 700-804/3.2-6.4 ms/28°, in-plane resolution 0.3 mm) in the axial plane. At 7T, t-spine was imaged using a custom-built spine array coil with T1-MPRAGE (3D TFL, TR/TE/TI/FA = 2300/1.8/950 ms/7°, 1 mm isotropic) in the sagittal plane and both multislice T2*-GRE (TR/TE/FA = 33/9.3ms/35°, in-plane resolution of 0.31 mm, with and without navigator correction) and 3D-T2*-radial-GRE (Siemens WIP 528L, TR/TE/FA = 20/9.2ms/10-13°, in-plane resolution 0.27mm) in the axial plane.

Results and Discussion

The custom-built coil was found to have a suitable excitation and receive profile to obtain good signal from the t-spine without exceeding SAR limits in both subjects. Figure 1 A and B show comparison of sagittal T1-MPRAGE images obtained from the 3T and 7T scanners, respectively, in a MS patient. While MS lesions are clearly identified at both field strengths (red arrows), they appear more conspicuous at 7T, with at least one additional lesion only visible at the higher field strength. Anatomical structures, including the gray and white matter, were more clearly identified at 7T (Fig 1D) than at 3T (Fig 1C) in the axial T2* images from the healthy volunteer. Similarly, the MS lesion was better delineated, and its location more clearly identified, on the axial T2* images from 7T (Fig 1F) compared to 3T (fig 1E).


No acknowledgement found.


1. Barkhof, Curr Opin Neurol. 2002 Jun;15(3):239-45.

2. Duan et. al. Magn Reson Med. 2015 Oct;74(4):1189-97

3. Nair et. al. Am J Neuroradiol. 2013 Nov-Dec; 34(11): 2215–22.


Comparison of t-spine images acquired on 3T and 7T from 39 y.o. F (A-B) and 30 y.o. M (E-F) with MS, and 25 y.o. M HV (C-D). Lesions and anatomy are more conspicuous on the 7T Siemens (B, D, F), than from 3T Siemens (A) or 3T Philips (C, E).

Figure 2: Frequency correction during reconstruction, either with a navigator (B) or using a radial scan (C) improved the quality of the axial scan (A), allowing better lesion visualization or grey-white matter differentiation.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)