Comparison between quantitative measurements of diffusion tensor imaging and T2 signal intensity in a large series of cervical spondylotic myelopathy patients for assessment of disease severity and prognostication of recovery.

Abstract

Cervical spondylotic myelopathy (CSM) is a common cause of spinal cord dysfunction. Recently, it has been shown that diffusion tensor imaging (DTI) may be a better biomarker than T2-weighted signal intensity (T2SI) on MRI for CSM. However, there is very little literature on a comparison between the quantitative measurements of DTI and T2SI in the CSM patient population to determine disease severity and recovery. A prospective analysis of 46 patients with both preoperative DTI and T2-weighted MRI was undertaken. Normalized T2SI (NT2SI), regardless of the presence or absence of T2SI at the level of maximum compression (LMC), was determined by calculating the T2SI at the LMC/T2SI at the level of the foramen magnum. Regression analysis was performed to determine the relationship of fractional anisotropy (FA), a quantitative measure derived from DTI, and NT2SI individually as well their combination with baseline preoperative modified Japanese Orthopaedic Association (mJOA) score and ∆mJOA score at the 3-, 6-, 12-, and 24-month follow-ups. Goodness-of-fit analysis was done using residual diagnostics. In addition, mixed-effects regression analysis was used to evaluate the impact of FA and NT2SI individually. A p value < 0.05 was selected to indicate statistical significance. Regression analysis showed a significant positive correlation between FA at the LMC and preoperative mJOA score (p = 0.041) but a significant negative correlation between FA at the LMC and the ΔmJOA score at the 12-month follow-up (p = 0.010). All other relationships between FA at the LMC and the baseline preoperative mJOA score or ∆mJOA score at the 3-, 6-, and 24-month follow-ups were not statistically significant. For NT2SI and the combination of FA and NT2SI, no significant relationships with preoperative mJOA score or ∆mJOA at 3, 6, and 24 months were seen on regression analysis. However, there was a significant correlation of combined FA and NT2SI with ∆mJOA score at the 12-month follow-up. Mixed-effects regression revealed that FA measured at the LMC was the only significant predictor of ΔmJOA score (p = 0.03), whereas NT2SI and time were not. Goodness-of-fit analysis did not show any evidence of lack of fit. In this large prospective study of CSM patients, FA at LMC appears to be a better biomarker for determining long-term outcomes following surgery in CSM patients than NT2SI or the combination values at LMC.