Inter-examiner reliability of radiographic measurements from Open-mouth lateral bending cervical radiographs

BACKGROUND: Following head and neck trauma, the involvement of the cranio-cervical junction (CCJ) and its contribution to a patients transition to chronic pain, is poorly understood. The detection of hypermobility in this region is dependent on clinical examination and static imaging modalities such as x-ray, CT and MRI. Sagittal plane hypermobility of the CCJ is evaluated using saggital view, flexion-extension cervical radiographs. Frontal plane hypermobility is typically assessed using lateral bending and open mouth cervical radiographs. Unfortunately there is no established reliability surrounding the use of these measures. This study explores the reliability of radiographic measurements of lateral-bending open-mouth cervical radiographs. METHODS: Cervical open-mouth lateral-bending X-ray images were collected from 56 different patients between 18 and 60 years of age patients following cervical spine injury. These images were interpreted by two musculoskeletal radiologists and two clinicians (physiatrist and chiropractor), using a standard set of measurements. Measurements included qualitative and quantitative assessments of the amount of asymmetry noted between various osseous landmarks. Reliability statistics were calculated for levels of agreement using kappa coefficients (κ) and Intraclass Correlation Coefficients (ICC) for dichotomous and continuous variables, respectively. RESULTS: Reliability (κ) for qualitative assessments were moderate to substantial for asymmetry of neutral C2 spinous position, dens-lateral mass spacing, and “step off” between the lateral borders of the articular pillars of C2 and C1 lateral mass (κ range = .47–.78). ICC values for the quantitative measurements of dens-lateral mass spacing and “step off” between the lateral borders of the C2 articular pillars and C1 lateral mass were moderate to excellent (ICC range = .56–.97). CONCLUSIONS: The qualitative and quantitative measurements used in this study demonstrated good to excellent inter-examiner reliability. Correlation with clinical findings is necessary to establish the utility of these measurements in clinical practice.