Diffusion Tensor Imaging Identifies Cervical Spondylosis, Myelitis, and Spinal Cord Tumors

TL;DR

This study shows that diffusion tensor imaging, especially fractional anisotropy, can help distinguish between cervical spondylosis, myelitis, and spinal tumors.

Contribution

The study demonstrates the diagnostic potential of DTI metrics, particularly FA, for differentiating spinal cord diseases.

Findings

FA values were significantly lower in tumor patients compared to spondylosis and myelitis.

FA showed strong discriminative performance with AUCs of 0.902 and 0.748 for differentiating tumors and myelitis from spondylosis.

FA correlated negatively with EDSS scores in myelitis and Ki-67 scores in tumors, indicating clinical relevance.

Abstract

Background: Diffusion tensor imaging (DTI) has been increasingly recognized for its capability to study microstructural changes in the neuropathology of brain diseases. However, the optimal DTI metric and its diagnostic utility for a variety of spinal cord diseases are still under investigation. Purpose: To evaluate the diagnostic efficacy of DTI metrics for differentiating between cervical spondylosis, myelitis, and spinal tumors. Methods: This retrospective study analyzed DTI scans from 68 patients (22 with cervical spondylosis, 23 with myelitis, and 23 with spinal tumors). DTI indicators, including fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD), were calculated. The Kruskal–Wallis test was used to compare these indicators, followed by Receiver Operating Characteristic (ROC) curve analysis, to evaluate the diagnostic efficacy of each indicator across disease pairs. Additionally, we explored the correlations of DTI indicators with specific clinical measurements. Results: FA values were significantly lower in tumor patients compared to those with cervical spondylosis (p < 0.0001) and myelitis (p < 0.05). Additionally, tumor patients exhibited significantly elevated MD and RD values relative to the spondylosis and myelitis groups. ROC curve analysis underscored FA’s superior discriminative performance, with an area under the curve (AUC) of 0.902 for differentiating tumors from cervical spondylosis, and an AUC of 0.748 for distinguishing cervical myelitis from spondylosis. Furthermore, a significant negative correlation was observed between FA values and Expanded Disability Status Scores (EDSSs) in myelitis patients (r = −0.62, p = 0.002), as well as between FA values and Ki-67 scores in tumor patients (r = −0.71, p = 0.0002). Conclusion: DTI indicators, especially FA, have the potential in distinguishing spondylosis, myelitis, and spinal cord tumors. The significant correlation between FA values and clinical indicators highlights the value of FA in the clinical assessment and prognosis of spinal diseases and may be applied in diagnostic protocols in the future.

Figures