Intramedullary high signal intensity (HSI) in T2-weighted (T2W) images has been widely used in the spinal cord evaluation in cervical spondylotic myelopathy (CSM) patients. However, it has discrepancies between the actual clinical status and the imaging findings, which limits its diagnosis accuracy and predictive capacity [1-4]. Diffusion tensor imaging (DTI) can reveal functional and pathologic information of the spinal cord through quantitative measurement of FA and ADC. Nonetheless, the conventional single-shot echo planar imaging (SS-EPI) DTI has low resolution and distortion. This study aims to propose an accurate quantitative evaluation method to help diagnosis of CSM patients, and to investigate the DTI metrics change after pressure release by operation. A navigated multi-shot EPI (MS-EPI) DTI sequence was used for the data acquisition and a recently proposed k-space phase variation correction method, SYMPHONY [5], was used for the image reconstruction. In addition, the correlation between the calculated DTI metrics and physical examination findings as well as preoperative and postoperative clinical scores was studied to determine whether DTI metrics are accurate predictors to CSM patients.

Neurologic Assessment All CSM patients were treated by anterior cervical discectomy and fusion. Modified Japanese Orthopedic Association (mJOA) scores were used for neurologic assessment of the severity of myelopathy, while neck disability index (NDI) and visual analogue scale (VAS) scores were used to evaluate the neck activities and the severity of their neck pain, respectively. Recovery after surgery was calculated using a recovery ratio: (postoperative mJOA score – preoperative mJOA score) / (17 – preoperative mJOA score)*100%. Greater than 50% of recovery ratio was considered indicative of a good outcome, whereas less than 50% was indicative of a fair outcome.

Data Acquisition All scans were performed on a Philips 3.0T scanner using a sixteen-channel SENSE neurovascular coil. Data were acquired from seven patients (5 male, 2 female) with CSM. The study was approved by the IRB of Beijing Chao-Yang Hospital. Before the surgery, sagittal T2W-TSE, axial multi-echo T2W-FFE, SS-EPI DTI and 2D-navigated interleaved EPI were scanned on each patient. The detailed scan parameters were listed in Table 1. Three months after the surgery, all patients were scanned again using the same imaging protocol.

Data Processing The Multi-shot DTI data were reconstructed using SYMPHONY. Afterwards, FA maps and eigenvalue maps (λ1, λ2, λ3) were calculated using DtiStudio [6]. Based on the eigenvalue maps, RD ((λ2+λ3)/2) and AD (λ1) were calculated. The sagittal T2W images were used to identify intramedullary HSI positions, and then ROIs were drawn at the same level of the HSI positions in b0 images, with the axial T2W images served as references to distinguish the boundary of CSF and the spinal cord. DTI metrics including FA, AD and RD were extracted from these ROIs. A paired, double-tailed Student T test was used to assess the statistical significance of these indices.

As shown in Fig. 1, the high resolution MS-EPI DTI in the sagittal view has less distortion induced by the metal implantation after the surgery compared with the SS-EPI DTI (e), and also shows FA maps with more homogenous colors (f). Fig. 2 shows the FA change between the preoperative and postoperative images for the same patient shown in Fig. 1. The left panel (a1-3, b1-3) shows the zoomed images from Fig. 1 (yellow box). The SS-EPI FA maps show strong artifacts, especially after surgery with metal implanted, and the FA decrease should not be correct as the patient did have neurologic improvement. In contrast, in MS-EPI DTI, the FA increased to normal values [7] in the lesion region after surgery and were the same with those in normal regions, with suppressed artifacts. Table 2 shows conventional MRI findings and neurologic assessments of all patients. Six of seven patients got a good outcome after the surgery therapy with the increased mean mJOA and the decreased VAS and NDI, while the intramedullary HSI did not change much. Meanwhile, these six patients showed increased FA and decreased AD and RD, which correlated with the neurologic assessment well. The paired, double-tailed Student T test showed significant difference between preoperative and follow-up assessment in Table 3. For normal regions, there are no significant change before and after surgery.Multi-shot DTI can provide images with higher resolution, less distortion, and more accurate DTI metrics, so it can perform better than SS-EPI DTI in CSM diagnosis and recovery monitoring. The DTI metrics form high resolution DTI can reflect the pathologic conditions of spinal cord quantitatively, and can potentially evaluate the function of spinal cords more accurately than HSI in T2W images.