Summary

X-ray is inferior to cross-sectional imaging in assessing glenoid wear, and advanced imaging is recommended for preoperative planning in shoulder arthroplasty.

Abstract

Introduction

Preoperative assessment of glenoid wear is crucial for patients with severe glenohumeral osteoarthritis undergoing shoulder arthroplasty. The Walch classification is the most commonly employed means of assessing glenoid wear, and has been shown to have good inter and intra-observer reliability. This classification relies on an advanced imaging modality, which may be expensive, radiation-intensive, and often requires outside referral. For these reasons, some surgeons advocate the use of plain radiographs (x-rays) alone for preoperative planning. The purpose of this study was to determine whether the Walch classification could be accurately applied to plain radiographs (x-rays) when compared to magnetic resonance imaging (MRI) as the gold standard. We hypothesize that x-rays cannot adequately replace advanced imaging for the evaluation of glenoid wear.

Methods

The pre-operative axillary x-rays of 50 patients undergoing shoulder arthroplasty between 2002 and 2013 were independently reviewed by five raters – one fellowship-trained shoulder surgeon, two sports medicine fellows, one junior orthopaedic surgery resident, and one musculoskeletal radiologist. Each rater individually classified the glenoid wear on x-ray and MRI according to the Walch classification. Subsequently, MRI scans were reviewed collectively and assigned a consensus classification to serve as the gold standard. The kappa coefficient was used to determine inter-observer agreement for x-rays and independent MRI reads, as well as the agreement between x-ray and consensus MRI.

Results

Analyses were performed for the complete 5-category (A1, A2, B1, B2, C) and 3-category (A, B, C) Walch classification. Inter-rater agreement for x-ray and MRI was “moderate” for both the 5-category (?=0.42, ?=0.47) and 3-category (?=0.54, ?=0.59) Walch classification. The agreement between x-ray and consensus MRI was only “fair-to-moderate” (?=0.21-0.51) for the 5-category and “moderate” (?=0.36-0.60) for the 3-category Walch classification.

Discussion

The inter-rater agreement for the Walch classification when applied to x-rays was “moderate,” with lower agreement between x-ray and MRI which was “fair to moderate.” These results are inferior to those reported previously for the reliability of the Walch classification in assessing glenoid wear using CT scans. The difficulty in accurate interpretation of axillary x-rays may be secondary to inherent variability with x-rays that is challenging to control. These factors include variations in the direction of the x-ray beam, body habitus of the patient, varying levels of shoulder stiffness, and experience of the x-ray technician. Accordingly, x-rays cannot be considered equivalent to cross-sectional imaging when applying the Walch classification. Advanced imaging is necessary when assessing glenoid wear for preoperative planning.