Utility of lateral scapular radiographs in initial evaluation of nontraumatic shoulder conditions

Abstract

Lateral scapular radiographs have been routinely included in the initial radiographic examination of both traumatic and nontraumatic shoulder conditions. With the advance of modern imaging modalities, the clinical utility of the lateral scapular view has become questionable. The purpose of the study was to assess the utilization of the lateral scapular view among the members of the American Shoulder and Elbow Surgeons (ASES) and to determine the clinical utility of the lateral scapular view in the initial evaluation of nontraumatic shoulder conditions. The study consisted of two parts. The first part involved an online survey of ASES members, which asked them 3 questions regarding their preference for radiographic evaluation of new patients with nontraumatic shoulder pain. The second part involved a clinical vignette-based survey, where 4 shoulder surgeons at our institution were given 50 clinical vignettes and asked to independently answer 4 questions regarding the most probable diagnosis, abnormal radiographic findings, further imaging studies, and treatment plan for each case. The survey was repeated twice; the first was given without a lateral scapular view, and the second given 4 weeks later with a lateral scapular view included. We obtained diagnostic accuracy and percent agreement of each surgeon over two surveys and intraobserver and interobserver reliability on each variable. Of a total of 235 ASES members who responded to the online survey, 193 (82.1%) indicated their routine use of a lateral scapular view. The most common reason for obtaining the view was better characterization of acromion morphology (75.4%). The clinical vignette-based survey showed substantial intrarater reliability (κ > 0.6) of the 4 surgeons between the two surveys for the most probable diagnosis, abnormal x-ray findings, and further imaging studies, while the intrarater reliability for treatment plan was moderate (κ= 0.548). The mean diagnostic accuracy of the 4 surgeons was almost equal (74% vs. 75%) between the surveys. Overall, each surgeon's percent agreement across the 2 surveys was over 70%. None of the 4 surgeons recommended a lateral scapular view for further imaging during the first survey; each wanted either advanced imaging (computed tomography, magnetic resonance imaging) or none. The addition of a lateral scapular radiograph in the presence of other orthogonal views does not appear to improve surgeons' diagnostic accuracy or affect their decision-making on the treatment plan in nontraumatic shoulder conditions. The clinical utility of the lateral scapular view may need to be reassessed in this setting.