ABSTRACT
PURPOSE: We designed this study to evaluate the interobserver reliability of magnetic resonance imaging (MRI), post-myelographic computed tomography (CT-myelogram) and myelography for the diagnosis of lumbar spinal stenosis. MATERIALS AND METHODS: Thirty two patients with lumbar spinal stenosis were evaluated preoperatively by MRI, CT-myelography, and myelography by four observers. For each set of scans, the examiners assessed the presence or absence of; 1) intervertebral disk protrusion, 2) facet arthrosis, 3) ligamentum flavum hypertrophy, and 4) nerve root impingement. The severities of 1) central, 2) lateral, 3) foraminal and 4) entire spinal stenosis were graded using a four-point scale range from 1 (no stenosis) to 4 (severe stenosis). RESULTS: Kappa statistical analysis revealed moderate interobserver agreement in terms of disk protrusion (0.49), nerve root impingement (0.42) and lateral stenosis grade based on MRI. Low levels of agreement were found for facet arthrosis and ligamentum flavum hypertrophy. Relatively higher levels of agreement were found for nerve root impingement, lateral stenosis and entire stenosis, graded by MRI and for central stenosis graded by myelography. CONCLUSION: The results of this study reflect that MRI showed relatively higher levels of interobserver reliability than the other diagnostic modalities in the evaluation of patients with lumbar spinal stenosis.