Comparison of agreement of cervical spine degenerative pathology findings in magnetic resonance imaging studies

Abstract Background Context Magnetic resonance imaging (MRI) is often used in the evaluation of degenerative conditions of the cervical spine. However, the agreement of interpreting and reporting varying degenerative findings on cervical MRI has not been well assessed. Purpose This study aimed to compare the inter -rater and intra -rater agreement of MRI findings between common degenerative findings of the cervical spine. Study Design A retrospective diagnostic study was used as study design. Patient Sample The sample consisted of 48 patients who underwent routine cervical spine MRI at our institution between January 2011 and June 2012. Outcome Measures Reviewers evaluated each MRI study at each vertebral level for disc hydration, disc space height, central stenosis, foraminal stenosis, end plate changes, spondylolisthesis, and cord signal change. Methods A panel of two orthopedic spine surgeons and four musculoskeletal radiologists independently reviewed 48 sets of T2-weighted axial and sagittal MRI sequences for a series of preselected criteria, and their findings were compared with those of the other panelists to determine inter-rater agreement. Each panelist also re-reviewed the first 10 studies to determine intra-rater agreement. Absolute inter-rater and intra-rater agreements were then calculated and compared for different findings. A modified analysis ignored disagreements between the least severe grades of findings to determine the inter-rater and intra-rater agreements of the most clinically important severity grades. Results Absolute inter-rater agreement ranged from 54.6% to 95.0%. Disc hydration (54.6%), central stenosis (72.7%), and foraminal stenosis (73.1%) demonstrated the lowest inter-rater agreement, whereas spondylolisthesis (95.0%) and cord signal change (92.9%) demonstrated the highest agreement. The modified analysis found better inter-rater agreement, ranging from 80.9% to 95.0%. Absolute intra-rater agreement ranged from 74.2% to 94.7%. The modified analysis again found better agreement, ranging from 85.0% to 94.7%. As would be expected, overall intra-rater agreement (81.6%, 95% CI 78.9%–84.3%) was higher than inter-rater agreement (75.7%, 95% CI 74.4%–77.0%). The clinical specialty of the reviewer had no significant impact on inter- or intra-rater agreement. Conclusions MRI findings play an important role in the management of patients with cervical spine conditions. For this reason, consistent descriptions of these findings are essential and physicians should be aware of the relative reliability of these findings. This systematic study developed standardized grading criteria and nomenclature for common clinically significant MRI findings in the cervical spine. Even in this optimized research setting, we found significant ranges in agreement across these MRI findings. In the clinical setting, inter- and intra-rater agreements may be lower, and the range of agreements between findings may be greater. Physicians should be aware of inconsistencies inherent in the interpretation of cervical MRI findings and should be aware that some findings demonstrate lower agreement than others.