Diagnosis of spinal dural arteriovenous fistula: a multimodal MRI assessment strategy

Objective To identify more specific screening indicators in magnetic resonance imaging (MRI) for the diagnosis of spinal dural arteriovenous fistulas (SDAVF) and to determine an efficient diagnostic strategy. Methods This retrospective study analysed clinical and imaging data of patients diagnosed with SDAVF and alternative myelopathy who underwent conventional MRI examinations. Additionally, 3D-T2-weighted-sampling perfection with application-optimised contrasts using different flip-angle evolutions (3D-T2-SPACE) and contrast-enhanced magnetic resonance angiography (CE-MRA) data from patients with SDAVF were compared with digital subtraction angiography (DSA) data. Results The age of onset, perimedullary flow voids (PFV), distribution of lesions, syringomyelia, degree of spinal oedema, and cauda equina disorder (CED) were factors that showed statistically significance in the identification of SDAVF with alternative myelopathy. After controlling for age, gender, PFV, degree of spinal cord swelling, and syringomyelia, the multivariable ordinal logistic regression model showed that CED sign (OR = 32.46; 95% CI: 2.47-427.15; P=0.008) was an independent predictor for SDAVF. The diagnostic model constructed using PFV sign and CED sign had better diagnostic performance, with an area under the curve of 0.957 (P<0.001), maximum Youden index of 0.844, sensitivity of 92.9%, and specificity of 91.5%. Both 3D-T2-SPACE (77.8%) and CE-MRA (83.3%) sequences had good localization values for SDAVF. Combining the two imaging examinations had better diagnostic accuracy than that of DSA. Conclusion CED and PFV on conventional MRI were specific indicators for the diagnosis of SDAVF. To make up for the lack of fistula localisation on conventional MRI, 3D-T2-SPACE and CE-MRA can be used. Together they complement each other and have good diagnostic potential.