Purpose: To assess the inter- and intra-reader agreement of commonly used quantitative and qualitative image parameters for the assessment of degenerative lumbar spinal canal stenosis (LSS) by magnetic resonance imaging (MRI).
Methods: In this ethical board approved cross-sectional multicenter study, MRI of 100 randomly selected patients (median age 72.5 years, 48 % female) of the prospective Lumbar Stenosis Outcome Study (LSOS) were evaluated by two independent readers. A set of five previously published core imaging parameters as well as nine qualitative and five quantitative additional parameters regarding LSS and degenerative changes of the lumbar spine were assessed to calculate κ and intraclass correlation coefficients (ICC) for the inter-reader agreement. Additional repeated image evaluations were performed by one reader to calculate the intra-reader agreement.
Results: κ values for the core image parameters ranged between 0.42 (compromise of the foraminal zone) and 0.77 (relation between fluid and cauda equina) for inter-reader agreement and between 0.59 (compromise of the foraminal zone) and 0.8 (compromise of the central zone) for intra-reader agreement. The inter-reader agreement for the non-core parameters showed κ values of 0.27-0.69 and ICC values of 0.46-0.85. The intra-reader agreement showed κ values of 0.53-0.69 and ICC values of 0.81-0.88.
Conclusions: The inter- and intra-reader agreement of commonly used quantitative and qualitative image parameters for the assessment of LSS showed quite a variability with previously defined core parameters having good to excellent inter- and intra-reader agreements.
Keywords: Low back pain; Lumbar; Magnetic resonance imaging; Reproducibility of results; Spinal canal.