Abstract
In 1989, we proposed an arthroscopic classification for Bankart lesions. This classification is useful for the selection of a surgical technique for treating traumatic anterior shoulder instability. We studied whether it was possible to predict the type of lesion by CT arthrography before arthroscopy. Bankart lesions of 85 shoulder joints were arthroscopically classified as follows: type 1; labral detachment with a well-developed anterior inferior glenohumeral ligament (AIGHL) (27cases), type 2; labral detachment with a poorly-developed AIGHL (21 cases), type 3; ligamentous avulsion with a retracted AIGHL (22 cases), type 4; ligamentous disruption with an osseous defect (6 cases), and type 5; no Bankart lesion with a stretched AIGHL (9 cases). The discrimination of these five types was analyzed in terms of the CT arthrographic findings on bilateral shoulder joints using the quantification theory of Hayashi. CT arthrographic findings predominantly present in each type were minimal labral wear(types 1,2,5) a high attachment of AIGHL to the glenoid rim (type 1), a low attachment(type 2), a lax AIGHL(type 2), no Hill-Sachs lesions(type 5), a worn out labrum(type 3), and severe labral wear and a bony defect(type 4). The accuracy of the diagnosis from the CT arthrography was as follows: type 1(75.0%), type 2(73.8%), type 3(77.4%), type 4(96.4%), and type 5(85.7%), Thus, our five arthroscopic types of Bankart lesions can be discriminated by CT arthroscopic findings prior to arthroscopy.
