Supine oblique views of the cervical spine: A poor proxy for the lateral view

Abstract

To determine whether supine 45° oblique radiographs of the cervical spine can accurately detect and quantify anteroposterior intervertebral plane displacements, an observer performance study was performed. A normally aligned dry cervicothoracic vertebral specimen and the same specimen with varying degrees of subluxation at one or more levels were radiographed in anteroposterior, lateral, and 45° oblique projections, in a simulated supine position. Twenty-five sets of radiographs were obtained, of which three were normal. In the remaining 22 sets, there were 43 intervertebral segments with an abnormal anteroposterior displacement. Since each intervertebral level (C2-T1) was individually evaluated, the study sample consisted of 150 intervertebral segments. Blinded to experimental conditions, six observers evaluated every intervertebral level on each lateral and oblique image in isolation. The data for each projection were compared using alternative free-response operator characteristic and free-response forced-error methodologies.

Significant differences in diagnostic accuracy were detected between the horizontal lateral and the supine oblique views to detect malalignment. Although supine oblique views detected many subluxations, they produced numerous false-negative and false-positive results. The findings of this study suggest that supine oblique views may be a useful part of the standard evaluation of the acutely injured cervical spine; however, they may not reliably portray clinically important anteroposterior displacements.