Interobserver and intraobserver variability in magnetic resonance imaging evaluation of patients with suspected disc herniation

Abstract

Objective: Magnetic resonance imaging (MRI) is usually the modality of choice to assess sciatica and intervertebral disc herniation. Despite remarkable progression in diagnostic imaging and surgical techniques, definite diagnosis based on imaging interpretation is still a great challenge. The aim of this study was to determine interobserver and intraobserver variability in reporting lumbar MRI between two neuroradiologists based on the new 2014 version of disc nomenclature. Patients and methods: The study population was composed of 134 irresponsive to conservative therapy patients with clinical presentations of disc herniation and lumbar radiculopathy. MRI was taken from all the participants using a 1.5 T MRI system. Two neuroradiologists evaluated the images, separately and one of them did it twice and interpreted the scans in sagittal and axial planes. Disc bulge, disc herniation and nerve root compression were evaluated at each level. Interobserver and interaobserver agreements between two neuroradiologists, and one neuroradiologist in two times of reporting were calculated for the evaluation of bulging and herniated discs and nerve root compression by applying the Kappa statistics. Results: Bulging disc, herniated disc, the type of disc, location of the discs, and nerve root compression diagnosis were significantly in excellent agreement (kappa 0.7, p-value<0.001) through intraobserver assessments, while interobserver assessments presented statistically significant with a fair agreement (kappa:0.4-0.7 and pvalue<0.05). Conclusion: Remarkable intraobserver agreement was found between diagnoses of disc-related pathologies of the lumbar spine while interobserver assessments revealed only fair concordance.