Utility and validity of neurite orientation dispersion and density imaging with diffusion tensor imaging to quantify the severity of cervical spondylotic myelopathy and assess postoperative neurological recovery

Abstract

Background context: Predicting postoperative prognosis with preoperative diagnostic imaging has clinical importance. Recent studies have indicated the utility of diffusion tensor imaging (DTI) to quantify the severity of cervical spondylotic myelopathy (CSM) and assess the prognosis of surgical outcomes. However, how to apply DTI to evaluate CSM in a clinical setting is not fully elucidated. Neurite orientation dispersion and density imaging (NODDI) is a model-based practical diffusion-weighted magnetic resonance imaging analysis for estimating specific microstructural features related directly to neuronal morphology. In a prior study, we indicated preoperative NODDI parameters are a promising tool with which to predict neuronal recovery after decompression surgery in patients with CSM with 2 years follow-up. However, the correlation between NODDI parameters and postoperative long-term outcomes and change of parameters over time postoperatively has remained largely unknown.

Study design: Retrospective cohort study.

Purpose: To determine the change of parameters of NODDI and conventional DTI over time, and the relationship between parameters and neurological recovery 2 years after surgery.

Patient sample: We included 28 consecutive patients with nontraumatic cervical lesions from CSM who underwent laminoplasty and were followed up for >2 years. Patients underwent magnetic resonance imaging before and approximately 2 weeks, 6 months, and 1 year after surgery.

Outcome measures: In addition to conventional DTI metrics, we evaluated intracellular volume fraction (ICVF) and orientation dispersion index, which are metrics derived from NODDI. The Japanese Orthopedic Association (JOA) scoring system was used before and 2 years after surgery to assess neurological outcome (JOA recovery rate).

Methods: NODDI and conventional DTI values were measured at the C2-C3 intervertebral level (control value) and the most compressed levels (C3-C7 intervertebral levels) were measured by 3 observers. The changes of these values from preoperatively, 2 weeks after surgery, 6 months after surgery, and 1 year after surgery, were determined. The correlations between preoperative neurological severity, postoperative neuronal recovery, and preoperative DTI or NODDI metrics were determined. No financial or material support was obtained for this study. There is no conflict of interest.

Results: The preoperative ICVF and fractional anisotropy at the most compressed level were significantly less than the preoperative values at the control C2-C3 intervertebral level and fractional anisotropy at the most compressed level was increased in the immediate postoperative period. By contrast, ICVF at the most compressed level was not increased in the immediate postoperative period and a significant increase was observed at 6 months after surgery. Preoperative ICVF was significantly correlated with JOA recovery rate at 2 years after surgery.

Conclusions: NODDI is a reproducible and reliable method for evaluation of CSM. ICVF improved after surgery and recovery of physical findings accompanied this change. ICVF may be applied clinically to predict postoperative recovery.

Keywords: Cervical spondylotic myelopathy; Diffusion tensor imaging; Fractional anisotropy; Intracellular volume fraction; Neurite orientation dispersion and density imaging; Neurological recovery.