The predictive effect of diffusion tensor imaging in the postoperative neurological recovery of cervical spondylotic myelopathy

English Keywords:Cervical spondylotic myelopathy  Diffusion tensor imaging  Prognosis

English Abstract:

  【Abstract】 Objectives: To investigate the usage of diffusion tensor imaging(DTI) in predicting postoperative neurologic recovery of cervical spondylotic myelopathy(CSM). Methods: The clinical data of fifty patients with CSM who underwent surgical treatment(43 to 86 years old, with mean age of 64.3±11.2 years old) were analyzed in this prospective study. Nineteen healthy volunteers(50 to 62 years old, with mean age of 54.6±4.2 years old) were recruited as control group. DTI parameters including fractional anisotropy(FA), mean diffusivity(MD), axial diffusivity(AD) and radial diffusivity(RD) in several regions of interest(ROIs) of cervical spinal cord(including the whole cord, ventral columns, lateral columns and dorsal columns) were derived from the images of a 3.0T magnetic resonance DTI sequence. The patients were divided into three groups according to postoperative neurologic recovery rate(RR) based on Japanese Orthopedic Association(JOA) scores, including patients with RR<0.4, patients with 0.40.7 respectively. The differences of DTI parameters between the control group and the three patient groups were compared. The correlations among FA, MD, AD and RD values in different regions of interest(ROI) and neurologic recovery rate were analyzed. The ability of predicting surgical outcome in CSM patients was compared among age, preoperative JOA scores, duration of symptoms, increased signal intensity on T2-weighted images and DTI parameters. Results: None of age, duration of symptoms and preoperative JOA scores showed significant difference among the three patient groups(P>0.05). Except for the FA value in the whole cord of RR>0.7 group, the FA, MD and RD values in the three patient groups showed significant differences from those in the control group respectively(P<0.01). In the three patient groups, only the FA values in the whole cord, lateral column and dorsal column of RR>0.7 group showed significant differences from those of RR<0.4 group. The FA, MD and RD values in all the investigated ROIs were significantly correlated with neurologic recovery rate(P<0.05), except for MD value in the dorsal column. A positive correlation existed between the FA value and the recovery rate in all the investigated ROIs(P<0.05), while MD and RD values showed negative associations with the recovery rate(P<0.05), in which FA value in the dorsal column showed the strongest correlation(r2=0.253, P<0.001). In the receiver operating characteristic curve(ROC) analysis, the areas under the curve(AUC) of the FA value in all the ROIs were significantly larger than those of age, preoperative JOA scores, duration of symptoms and increased signal intensity within cervical cord on the T2-weighted images. Among them, FA value of lateral column presented the best ability of predicting surgical outcome for CSM patients. Conclusions: Preoperative DTI parameters, especially the FA values in the dorsal and lateral columns, has superior capacity in predicting surgical outcome for patients with CSM compared to clinical and traditional MRI features. DTI may be a promising tool in the clinical practice of CSM.