Comparison of magnetic resonance imaging and computed tomography in predicting facet arthrosis in the cervical spine.

STUDY DESIGN: Retrospective review. OBJECTIVE: To determine the ability of magnetic resonance imaging (MRI) and computed tomography (CT) to predict the presence of cervical facet arthrosis. SUMMARY OF BACKGROUND DATA: In the Food and Drug Administration Investigational Device Exemption trials of cervical disc arthroplasty (CDA), the presence of facet arthrosis on CT was a contraindication to the insertion of a CDA. Most surgeons routinely obtain an MRI, but not necessarily a CT before performing surgery in the cervical spine. We sought to determine if the MRI alone is adequate to assess for the presence of facet arthrosis. METHODS: Three experienced spine surgeons retrospectively evaluated CT scans and MRIs of the same patients, obtained within 30 days of each other in a blinded, random fashion. Reviewers graded each of the MRI and CT scan as normal or abnormal on 3 separate occasions and if the facet was abnormal, each reviewer graded the degree of arthrosis. The radiologist's evaluation for each study was compared with our results. RESULTS: Of 594 facets analyzed, 43.1% were categorized as normal on CT, and of those, MRI concordance was only 63.7% with moderate/substantial intermethod agreement. Furthermore, MRI was concordant only 15.9% of the time in patients with ankylosed facet joints on CT. CT inter-rater reliability showed substantial agreement for diagnoses of both normal and ankylosis and fair agreement for lesser degrees of facet arthrosis. MRI inter-rater reliability showed fair/moderate agreement in normal and ankylosed segments and only slight agreement with lesser degrees of facet arthrosis. CT intrarater reliability showed substantial agreement in normal or ankylosed joints, but only fair agreement for all other categories; MRI showed only fair agreement. CONCLUSION: The ability of MRI to adequately determine the presence or amount of facet arthrosis is not reliable. Additionally, for abnormal facets, MRI was not reliable in adequately determining the degree of arthrosis. Our data suggest that computed tomography remains necessary in diagnosing facet arthrosis before CDA.