A new radiographic classification of greater tuberosity irregularities in the detection of rotator cuff tears

Abstract

INTRODUCTION: We propose a new radiographic classification of greater tuberosity changes seen in patients with atraumatic rotator cuff dysfunction. The aim of this study is to determine the predictive values and accuracy of this classification for the detection of rotator cuff tears using arthroscopic findings as the gold standard.

STUDY DESIGN & METHODS: This was a retrospective study of 91 consecutive patients (47 male, 44 female; mean age 57 year old) who underwent shoulder arthroscopy for a clinical diagnosis of rotator cuff dysfunction. The presence and size of a rotator cuff tear for each patient had been documented at the time of surgery. The radiographs were anonymised and randomised for interpretation by 2 experienced shoulder surgeons and of 2 orthopaedic residents newly-educated on the use of the classification. The predicative values of the new classification for the diagnosis of rotator cuff tear were calculated. Inter-rater agreement values were obtained using the Fliess’ kappa method. Statistical significance was set at p<0.05.

RESULTS: Comparative statistics showed that Stage 0 or 1 were associated with the absence of a cuff tear. Stage 2 and 3 were strongly predictive of a rotator cuff tear. The mean predicative values of the new classification stage 2 and 3 for the detection of a rotator cuff tear were: sensitivity 93%, specificity 83%, positive predictive value 87%, and negative predictive value 92%. The accuracy of the experienced observer group was significantly higher than the inexperienced group (97% v 83%, p-value <0.01) and therefore the overall group accuracy was 89%. Inter-rater agreement for stages 1, 2 and 3 changes was substantial overall (Kappa value of 0.63) and almost perfect between the experienced assessors (Kappa=0.95).

CONCLUSIONS: The new classification is a useful adjunct for the diagnosis of rotator cuff tears where adequate radiographs are available. It may be used as a screening test in a large number of patients with shoulder pain and select stage 2 and 3 patients for further rotator cuff imaging to assist preoperative planning.