Diffusion tensor imaging (DTI) has demonstrated potential in detecting microstructural changes in pathological tissues in the central nervous system. In particular, there has been increased interest in DTI of the spinal cord, as visualization of spinal cord tracts can play a role in the diagnosis and prognosis of neurodegenerative disease or trauma. Development of DTI in the spinal cord has struggled, however, due to the need in acquiring high-resolution images in clinically feasible acquisition times. Additionally, it is desirable to acquire as many axial slices as possible along the length of the cord to assess level- and tract-specific changes . Multiband excitation allows simultaneous acquisition of slices without increasing the acquisition time while potentially maintaining high SNR². Here, we implement multiband excitation with DTI to increase the number of axial slices to enable characterization of the entire cervical spinal cord and brainstem in a clinically feasible scan time. To our knowledge, this work is the first to report the use of multiband excitation for DTI acquisition of axial slices in the cervical spinal cord at 3T.

Acquisition: A healthy volunteer (23 years old, female) participated in this study under a protocol approved by the institutional review board. Imaging was performed on a 3T whole body Philips scanner (Philips Achieva, Best, Netherlands). A two-channel body coil was used in multi-transmit mode for excitation and a 16-channel SENSE neurovascular coil was used for reception. The DTI sequences studied consisted of a single-shot EPI with a reduced field-of-view outer volume suppression technique³, sampling 15 non-collinear diffusion-weighted directions at an effective b-value of 750 s/mm². The protocol included a cardiac gated, spin-echo acquisition with the following relevant parameters: flip angle=90°, SENSE factor=1.5, resolution=1x1 mm², slice thickness=5 mm, TE/TR=55 ms/5 beats (~5000 ms), number of dynamics=2, slices=30 (brain stem-C6 coverage). In order to evaluate the efficiency of multiband excitation, two DTI acquisitions were performed: 1) with multiband (9 minutes) and 2) without multiband (18 minutes). For the multiband sequence, a multiband factor of 2 was applied to simultaneously excite two slices (FOV/2 apart). To mitigate fold-over artifacts, no multiband excitation was applied when acquiring the b=0 volume. Figure 1 shows the acquired FOV with both scans.

Processing: Each diffusion weighted acquisition was registered to a b=0 volume using the FLIRT package from FSL v5.0.2.1 (FMRIB, Oxford, UK). Diffusion tensor calculation was estimated with a nonlinear fit in Camino⁴. Mean DTI indices were calculated over the spinal cord for both acquisitions and compared to the acquisition without multiband.

Figure 2 shows the fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) maps for the pons through C4 of the cervical segment. Despite cutting the acquisition time in half, all maps provide high contrast between the spinal cord and surrounding cerebrospinal fluid (CSF); the FA and AD maps provide the ability to distinguish white matter from gray matter inside the spinal cord. The mean SNR value for white matter calculated over all the DTI indices for the C1-C4 segment was reduced by an average of 10% for multiband compared to the no multiband case (multiband: 6.98, no multiband: 7.73), which is less than the theoretical SNR penalty expected when reducing the scan time by 50%. Considering both white matter and gray matter (i.e. whole cord), the SNR is similar between multiband and no multiband (multiband: 6.60, no multiband: 6.14). Additionally, it should be noted that no fold-over artifacts are detected in any of the levels of all the maps. Table 1 lists the mean DTI values derived for the brainstem and C1-C4 segments. The multiband acquisition yields DTI measurements that are in good agreement to the no multiband reference and literature values⁵, with a mean difference of less than 3% for the cervical cord segment. This study demonstrates the potential and feasibility of multiband excitation with diffusion imaging in the brainstem and cervical spinal cord. With multiband excitation, we are able to cover multiple axial slices over the length of the cord and brainstem in half the scan time without sacrificing measurement certainty. The provided coverage with multiband excitation permits investigation of relationships between the midbrain, spinal cord and cerebellum, and implementation of more advanced, multi-shell diffusion protocols (NODDI, DBSI, q-space) in clinically acceptable scan times.