Comparison of the critical shoulder angle in radiographs and computed tomography

The critical shoulder angle (CSA) is an indicator of degenerative shoulder pathologies. CSAs above 35 are associated with degenerative rotator cuff disease, whereas values below 30 are common in osteoarthritis of the glenohumeral joint. Measurements are usually performed on radiographs; however, no data have been reported regarding the reliability of CT scan measurements between different readers or the reproducibility of measurements from radiographs to CT scans. The purpose of our study was to clarify whether CSA measurements on radiographs and CT scans of the same patients show similar values. CSA measurements of 60 shoulders (59 patients) were performed on radiographs and multiplanar reconstructions of corresponding CT scans. Inter-reader reliability and inter-method correlation were calculated. The mean discrepancy between readers was only 0.2 (SD ±1.0 ) on radiographs. CT scan measurements showed a mean discrepancy of 0.3 (SD ±1.2 ). The inter-reader reliability was 0.993 for radiographs and 0.989 for CT scans. There was a very strong inter-method correlation between the CSA measured on radiographs and CT scans (Spearman's rho = 0.974). The mean differences between angles on radiographs and CT measurements were -0.05 (SD ±1.2 ) and 0.1 (SD ±1.2 ), respectively. Measurements of the CSA on anterior-posterior radiographs and CT scans are highly correlated, and inter-modality differences are negligible. (orig.)