Comprehensive computed tomography assessment of the upper cervical anatomy: what is normal?

High-speed computed tomography (CT) exams have replaced traditional radiographs for assessment of cervical spine injuries in many emergency departments. Recent evidence demonstrates that even subtle displacements can indicate significant upper cervical spine injuries. Many different anatomical measurements have been described in the upper cervical spine to date, most of them based on X-ray. The range of anatomical relationships that exist in an uninjured population must be known to reliably detect abnormal relations. The measurements with the lowest normal variation are likely to be most useful in detecting injuries.

The purpose of this study was to describe the normal quantitative anatomical relationships as well as the threshold measurements most likely to detect injury in the upper cervical spine.

Retrospective anatomical case review.

Seventy-six thin-sliced cervical CT scans randomly selected from a trauma population, all negative for injury in the cervical spine.

Forty-two different anatomical measurements were made of the upper cervical spine. These included traditional historical measurements and other detailed dimensions to characterize occipitocervical (OC) and atlantoaxial (AA) joint relationships.

After review of all the anatomical measurements performed in the upper cervical spine, direct measurements of the joint space had the least variation. The mean OC joint space was 0.6mm, with an upper 95% confidence interval (CI) of 1mm at the most anterior or posterior aspects of the joints. This was true for both sagittal and coronal measurements. The mean AA joint space was 0.6mm, with an upper 95% CI of 1.2mm at the lateral aspect of the joint on the coronal image only. The midsagittal structures demonstrated significantly higher standard deviation and variability.

These results revealed consistently narrow joint spaces and left-right symmetry in the upper cervical spine joints that do not vary according to demographics. There was distinctly greater consistency in the coronal plane, which enabled more precise diagnostic measurement and side-to-side comparison of measurements. This precision will enable more accurate identification of abnormal scans, which should prompt consideration for additional workup. Thus, better understanding of these relationships may enable earlier detection of subtle craniocervical dissociative injuries based on CT scan data. This is important, because the only evidence of a severe injury on CT can be subtle misalignment.