Imaging of glenohumeral osteoarthritis: Reliability and reproducibility of radiological classifications.

Abstract

Background: Glenohumeral degenerative joint disease may affect up to 20% of the population. There are several classification systems of this disease in the scientific literature.

Objective: The aim of this study is to determine the reliability and reproducibility of glenohumeral osteoarthritis classification systems.

Methods: We assessed glenohumeral plain radiographs performed in a University Hospital. These radiographs were graded into nine radiological classification systems by two observers on two evaluations. Patients who have performed CT/MRI scan were staged according to the Walch classification. The intra-observer and inter-observer reliability of the classification schemes were determined by using Cohen's weighted kappa coefficient.

Results: A total of 752 patients were included in the study (52.4% females and 47.6% males), mean aged 66.2 ± 16.3 years. The intra-observer and inter-observer reliabilities were 0.543 (observer 1), 0.630 (observer 2), and 0.760 (inter-observer) for the Weinstein grading system; 0.661, 0.706, and 0.761 for the Guyette grading system; 0.575, 0.679 and 0.704 for the Kellgren and Lawrence classification; 0.817, 0.816 e 0.871 for the Samilson and Prieto classification; 0.791, 0.811 and 0.847 for the Allain modification; 0.797, 0.842 and 0.860 for the Gerber modification; 0.773, 0.827 and 0.828 for the Buscayret modification; 0.584, 0.648 and 0.755 for the Hawkins and Angelo classification; 0.661, 0.749 and 0.764 for the Rosenberg classification. Intra-observer reliability for MRI was 0.757 (observer 1) and 0.675 (observer 2), while intra-observer reliability for CT was 0.811 (observer 1) and 0.653 (observer 2). Inter-observer reliabilities were 0.790 for MRI and 0.673 for CT.

Conclusion: The classification systems according to Weinstein, Guyette, Hawkins and Angelo, Rosenberg and the modifications of the Samilson and Prieto classification according to Allain, Gerber and Buscayret showed a comparable reliability with the commonly used glenohumeral osteoarthritis grading systems, Samilson and Prieto and Kellgren and Lawrence. Thus, they are recommended for clinical and especially scientific purposes.