Reliability of Human Lumbar Facet Joint Degeneration Severity Assessed by Magnetic Resonance Imaging

Abstract Objective The purpose of this study was to determine the reliability of the assessment of lumbar facet joint degeneration severity by analyzing degeneration subscales using magnetic resonance imaging (MRI) in human participants. Methods The reliability of articular cartilage degeneration, subchondral bone sclerosis, and osteophyte formation subscales of lumbar facet joint degeneration severity was assessed in MRI images from n = 10 human participants. Each scale was applied to n = 20 lumbar facet joints (L4/5 level). Three examiners were trained. A first assessment of MRI images was provided by the examiners followed by a second assessment 30 days later. Intraobserver and interobserver reliability were determined using percent agreement, the weighted kappa coefficient κw for paired comparisons, and the overall weighted kappa κo. The minimum threshold for reliability was set at moderate levels of agreement, κw > 0.40, based upon previous recommendations. Results The articular cartilage subscale had acceptable intraobserver (κo = 0.51) and interobserver (κo = 0.41) reliability. Scales for subchondral bone sclerosis (intraobserver κo = 0.28; interobserver κo = 0.10) and osteophyte formation (intraobserver κo = 0.26; interobserver κo = 0.20) did not achieve acceptable reliability. Conclusion Of the 3 subcategories of lumbar facet joint degeneration, only articular cartilage degeneration demonstrated acceptable reliability. Subscales of lumbar facet joint degeneration should be considered independently for reliability before combining subscales for a global degeneration score. Owing to the inherent difficulty of assessing lumbar facet joint degeneration, the use of multiple examiners independently assessing degeneration with reliable scales and then coming to a consensus score upon any disagreements is recommended for future clinical studies.