Plain radiographs have limited sensitivity for glenohumeral cartilage lesions

Glenohumeral cartilage lesions are frequently encountered during shoulder arthroscopy, but are not always represented on radiographs. We intended to define the diagnostic value of plain shoulder radiographs in detecting glenohumeral cartilage lesions. The radiographs of 167 patients undergoing shoulder arthroscopy were analyzed for signs of degenerative joint disease. Formal criteria indicating osteoarthritis (osteophytes, joint space narrowing, and sclerosis) were registered and correlated to cartilage lesions diagnosed during shoulder arthroscopy. The intrarater reliabilities were .920 (T.K.) and .953 (M.E.) and the interrater reliabilities were .802 (T.K. 1, M.E. 1), .909 (T.K. 2, M.E. 1), .922 (T.K. 1, M.E. 2), and .953 (T.K. 2, M.E. 2), indicating excellent agreement. There were 35 correct positive, 91 correct negative, 34 false negative, and 7 false positive radiographs. The sensitivity and specificity for any degree of cartilage lesion were 50.7% and 92.9%, respectively, and the positive and negative predictive values were 83.3% and 72.8%, respectively. For third- and fourth-grade cartilage lesions, the sensitivity was 76.0% and the positive and negative predictive values were 73.1% and 93.8%, respectively. While plain radiographs can often detect severe cartilage lesions, the sensitivity and negative predictive value are too low to reliably exclude cartilage lesions overall. In case of doubt, we recommend further radiological workup.