2882
Diffusion Weighted MRI in Cervical Spine with Metal Hardware at 3T1Medical College of Wisconsin, Milwaukee, WI, United States
Synopsis
Surgical repair of the cervical spinal cord to correct instability induced through trauma or degenerative disease often precludes follow-up MRI due to severe artifacts caused by metal stabilization hardware. Postoperative imaging is essential to monitor the hardware positioning, disease progression and new complications that may occur after surgery. In this study, we investigate the imaging capabilities of the recently developed multi-spectral diffusion weighted PROPELLER technique within the spinal cord immediately adjacent to metallic instrumentation. In addition, we assess the quantitative stability of this approach relative to other conventional methods in cohort of normal controls.
Introduction
Spinal fusion hardware is used in surgical procedures to treat several degenerative cervical spinal cord diseases like cervical spondylosis, fractures and bone abnormalities [1]. Quantitative metrics provided by Diffusion Tensor Imaging (DTI), such as apparent diffusion coefficient (ADC) and fractional anisotropy (FA) can help detect the severity of the spinal cord injury [2]. DTI are traditionally acquired using full field of view (FOV) single shot echo planar imaging (EPI) sequences or reduced field of view (rFOV) EPI sequences [3]. Both conventional approaches lack the ability to image near metal implants, due to the severe susceptibility artifacts that result in such images. Recently, a multi-spectral diffusion imaging scheme using the non-Cartesian periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) technique was introduced to acquire diffusion images near metal implants in musculoskeletal applications [4]. In this work, we demonstrate the feasibility and applicability of this approach to image the spinal cord in the immediate vicinity of fusion hardware.Methods
The MSI-PROPELLER technique was compared to the traditional EPI and rFOV EPI (FOCUS-EPI sequence from GE Healthcare, Milwaukee) techniques in normal controls without any implants and with FOCUS-EPI in instrumented myelopathy patients. Data from a cohort of 5 normal controls without any implants or history of any spinal injuries and 5 patients with post-surgical follow up of Cervical Spondylotic Myelopathy (CSM) were collected. 4 of the 5 patients received spinal fusion with bilateral screws across 4 or 5 posterior segments while 1 patient received fusion with anterior screws in 2 segments. 2 datasets with b=350 s/mm2 and b=600 s/mm2 was acquired for each of the three techniques in case of normal controls. The imaging parameters were: field of view=12cm, matrix size = 64x64, slice thickness=3 mm, diffusion directions=3, TR=4s for EPI and FOCUS-EPI and a TR=2s for MSI-PROPELLER and, in-plane resolution: 1.875mm. 3 spectral bins were acquired for the MSI-PROPELLER technique. For instrumented myelopathy patients, one T2 weighted (b=0) and 3 diffusion-weighted (b=350s/mm2) images in x, y and z directions were acquired. The matrix size for FOCUS-EPI was 84x84 resulting in an in-plane voxel resolution of 1.43mm. Apparent diffusion coefficient (ADC) maps were calculated using a mono-exponential diffusion model.Results
Fig 1 shows the DWI acquisition for a patient with bilateral screws in the 4 posterior segments. The FOCUS-EPI technique suffers from in-plane as well as through plane distortion and artifacts in the spinal cord (region in white box) due to its proximity to screws as seen in Fig 1d. The combination of off-resonant and on-resonant multi-spectral data in MSI-PROPELLER (Fig 1e) yields artifact free image of the spinal cord. The distribution of mean ADC values in the spinal cord for the normal control cases obtained using EPI, FOCUS-EPI and MSI-PROPELLER for both b values and instrumented cases is shown Fig 2. The mean ADC obtained from b=350s/mm2 dataset is higher than the one obtained from b=600s/mm2 for all the methods. The mean ADC values are very close to each other for b=600s/mm2 which demonstrates the accuracy of the technique. Fig 3 shows the ADC maps for the instrumented cases. There is greater variability in the ADC values in the instrumented cases which could be an indication of the severity of the cord injury. For example, the case with the highest ADC value appears to have cord compression on the Axial T1. The FOCUS-EPI method is unusable due to severe artifacts whereas MSI-PROPELLER yields artifact free images of the cord. The case with very high ADC (Fig 3b) could be indicative of edema whereas the one with low ADC (Fig 3d) could indicate restriction.Discussion
Although the MSI PROPELLER technique improved visualization of the spinal cord compared to DWI-EPI, there are some limitations. At low b values, other gradients like the slice select gradient might also affect the signal weighting which could result in the elevated ADC in the MSI-PROPELLER technique. However, this can be accounted for in b value calculations in future. The accuracy of the ADC estimation and its correlation with underlying pathology and neurological status needs to be further validated on a larger cohort of patients with spinal instrumentation.Conclusion
In this work, we have demonstrated the feasibility of MSI-PROPELLER technique to acquire DWI in the cervical spinal cord near fixation hardware. The comparison of MSI-PROPELLER with EPI and FOCUS-EPI in normal controls validated the applicability of MSI-PROPELLER technique in imaging cervical spinal cord.Acknowledgements
Research reported in this publication was supported by Daniel M Soref Charitable Trust and NIH R21EB023415-01A1. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.References
[1] Omeis, I., DeMattia, J. A., Hillard, V. H., Murali, R., & Das, K. (2004). History of instrumentation for stabilization of the subaxial cervical spine. Neurosurgical focus, 16(1), 1-6.
[2] Mamata, H., Jolesz, F. A., & Maier, S. E. (2005). Apparent diffusion coefficient and fractional anisotropy in spinal cord: age and cervical spondylosis–related changes. Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine, 22(1), 38-43.
[3] Zaharchuk, G., Saritas, E. U., Andre, J. B., Chin, C. T., Rosenberg, J., Brosnan, T. J., ... & Fischbein, N. J. (2011). Reduced field-of-view diffusion imaging of the human spinal cord: comparison with conventional single-shot echo-planar imaging. American Journal of Neuroradiology.
[4] Koch, K. M., Bhave, S., Gaddipati, A., Hargreaves, B. A., Gui, D., Peters, R., ... & Kaushik, S. S. (2018). Multispectral diffusionâweighted imaging near metal implants. Magnetic resonance in medicine, 79(2), 987-993.
Figures
