Radiographs and computed tomography scans show similar observer agreement when classifying glenoid morphology in glenohumeral arthritis.

Background: Glenohumeral subluxation and glenoid morphology are commonly evaluated in primary osteoarthritis by use of the Walch classification. The reliability of this classification system has been analyzed only by computed tomography (CT). The purpose of this study was to determine the reliability of plain axillary radiographs compared with CT scans.
Methods: Three shoulder surgeons blindly and independently evaluated the radiographs and CT scans of 75 consecutive shoulders with primary glenohumeral osteoarthritis. Each observer classified all shoulders according to Walch in 4 separate sessions, each 6 weeks apart. There were 2 sessions using only radiographs and 2 using only CT scans. The order of shoulders evaluated was randomized.
Results: The first reading by the most senior observer based on CT was arbitrarily used as the "gold standard" (A1, 21; A2, 13; B1, 12; B2, 28; C, 1). The average intraobserver agreement for radiographs was 0.66 (substantial; 0.66, 0.59, and 0.74 for each observer). The average intraobserver agreement for CT scans was 0.60 (moderate; 0.53, 0.61, and 0.65). Pairwise comparisons between observers showed higher agreement for radiographs than for CT scans (0.48 vs. 0.39). The average agreement for observations on radiographs and CT scans was 0.42 (moderate; 0.40, 0.37, and 0.50).
Conclusion: In this study, intraobserver agreement using the Walch classification based on axillary radiographs was substantial and compared favorably with agreement based on CT scans. The Walch classification provides a useful frame of reference when assessing subluxation and glenoid morphology in primary glenohumeral osteoarthritis, but not unlike other classification systems, it does not allow perfect agreement among observers.
(Copyright © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. Published by Elsevier Inc. All rights reserved.)