Abstract
Purpose: Magnetic resonance imaging (MRI) and Computerized tomography (CT) are commonly used for the diagnosis and assessment of lumbar spinal stenosis. The available literature has not identified which modality is superior. We compared the reliability and accuracy of CT and MRI in the assessment of lumbar spinal stenosis.
Method: We performed a prospective review of CT and MRI scans of 54 patients referred for surgical consultation. One orthopaedic spine fellow and one neuro-radiologist reviewed the CTs and MRIs. A qualitative and quantitative analysis was performed. Intra-observer and inter-observer reliability was determined using Kappa coefficient. The patient’s official reports were correlated with analysis performed by the two reviewers. Owsestry and SF-36 data was correlated with the qualitative and qualitative assessment of stenosis on CT, MRI using the Pearson’s R coefficient.
Results: MRI – substantial inter-observer agreement was achieved between surgeon and neuro-radiologist as well as between surgeon and reporting radiologist (κ= 0.74 and κ=0.64 respectively). Moderate agreement was found between neuro-radiologist and reporting radiologist (κ=0.57). Almost perfect intra-observer reliability for MRI was achieved by the two expert reviewers (κ=0.91 for surgeon and κ=0.92 for neuro-radiologist). CT – moderate inter-observer agreement (κ=0.58) was found between surgeon and neuro-radiologist. Fair agreement was found between neuro-radiologist and reporting radiologist and between surgeon and reporting radiologist (κ=0.30 and 0.32 respectively). Substantial intra-observer agreement was found for the surgeon (κ=0.77) while the neuro-radiologist achieved almost perfect agreement (κ=0.96).
Conclusion: This study directly demonstrates that MRI is likely a more reliable tool than CT, but neither correlates with functional status.
Correspondence should be addressed to: COA, 4150 Ste. Catherine St. West Suite 360, Westmount, QC H3Z 2Y5, Canada. Email: meetings@canorth.org
