Assessment of cervical spondylotic myelopathy using diffusion tensor magnetic resonance imaging parameter at 3.0 tesla.

STUDY
DESIGN: Cross-sectional study.
OBJECTIVE: To assess spinal cord condition in patients with cervical spondylosis (CS), using diffusion tensor imaging parameter. SUMMARY OF BACKGROUND DATA: Although myelopathy is a common symptom after CS, clinically objective assessment for determination of surgical intervention is not straightforward.
METHODS: Twenty-six patients with CS and 30 normal control subjects were enrolled. Diffusion tensor imaging was obtained using a single-shot fast spin-echo-based sequence at 3.0 T. Mean diffusivity (MD) and fractional anisotropy (FA) were measured in the axial plane at 6 spinal levels. To evaluate MD and FA in patients with CS considering the normal variation at each spinal level and between spinal levels, MD and FA at the most compressed spinal level were transformed to normalized values with a z score. Presence of myelopathy was predicted with the MD and FA z scores. Diagnostic validity of MD and FA was compared with receiver operating characteristic analysis. More effective parameter and the optimal cutoff value for prediction were determined.
RESULTS: In normal subjects, MD and FA were significantly different between spinal levels. In patients with myelopathy, an MD increase or an FA decrease was demonstrated in most cases. Although both an MD increase and an FA decrease had diagnostic validity for myelopathy, receiver operating characteristic analysis demonstrated a higher sensitivity and specificity for prediction of an MD increase than an FA decrease (areas under the curve for MD and FA were 0.903 and 0.760, respectively). An MD z score of 1.40 was considered to be the best diagnostic cutoff value with 100% sensitivity and 75% specificity.
CONCLUSION: Myelopathy can be predicted with high accuracy with diffusion tensor imaging parameter, with the MD z score at the most compressed spinal level. LEVEL OF EVIDENCE: 3.