Radiologically defining horizontal gaze using EOS imaging-a prospective study of healthy subjects and a retrospective audit.

BACKGROUND CONTEXT: As sagittal alignment of the cervical spine is important for maintaining horizontal gaze, it is important to determine the former for surgical correction. However, horizontal gaze remains poorly-defined from a radiological point of view.

PURPOSE: The objective of this study was to establish radiological criteria to define horizontal gaze.

STUDY DESIGN/SETTING: This study was conducted at a tertiary health-care institution over a 1-month period.

PATIENT SAMPLE: A prospective cohort of healthy patients was used to determine the best radiological criteria for defining horizontal gaze. A retrospective cohort of patients without rigid spinal deformities was used to audit the incidence of horizontal gaze.

OUTCOME MEASURES: Two categories of radiological parameters for determining horizontal gaze were tested: (1) the vertical offset distances of key identifiable structures from the horizontal gaze axis and (2) imaginary lines convergent with the horizontal gaze axis.

MATERIALS AND METHODS: Sixty-seven healthy subjects underwent whole-body EOS radiographs taken in a directed standing posture. Horizontal gaze was radiologically defined using each parameter, as represented by their means, 95% confidence intervals (CIs), and associated 2 standard deviations (SDs). Subsequently, applying the radiological criteria, we conducted a retrospective audit of such radiographs (before the implementation of a strict radioimaging standardization).

RESULTS: The mean age of our prospective cohort was 46.8 years, whereas that of our retrospective cohort was 37.2 years. Gender was evenly distributed across both cohorts. The four parameters with the lowest 95% CI and 2 SD were the distance offsets of the midpoint of the hard palate (A) and the base of the sella turcica (B), the horizontal convergents formed by the tangential line to the hard palate (C), and the line joining the center of the orbital orifice with the internal occipital protuberance (D). In the prospective cohort, good sensitivity (>98%) was attained when two or more parameters were used. Audit using Criterion B+D yielded compliance rates of 76.7%, a figure much closer to that of A+B+C+D (74.8%). From a practical viewpoint, Criterion B+D were most suitable for clinical use and could be simplified to the "3-6-12 rule" as a form of cursory assessment. Verbal instructions in the absence of stringent postural checks only ensured that ~75% of subjects achieved horizontal gaze.

CONCLUSIONS: Fulfillment of Criterion B+D is sufficient to evaluate for horizontal gaze. Further criteria can be added to increase sensitivity. Verbal instructions alone yield high rates of inaccuracy when attempting to image patients in horizontal gaze. Apart from improving methods for obtaining radiographs, a radiological definition of horizontal gaze should be routinely applied for better evaluation of sagittal spinal alignment.