[Validity of ultrasound examinations of disorders of the shoulder joint].
OBJECTIVES: The objective of the present study was to assess the validity of ultrasound diagnosis of shoulder disorders in relation to examiner experience. METHODS: A total of 239 patients referred to us for shoulder arthroscopy from October 2001 to June 2004 were prospectively studied by ultrasound. The following ultrasound diagnoses were evaluated: total and partial rotator cuff tears, calcific tendinitis, biceps tendon injuries and subacromial bursitis. Examiner A established the ultrasound diagnoses and examiner B performed the surgery during week A, whereas examiner B did the examinations and examiner A operated in week B. The surgeon was blinded to the ultrasound results. Examiner A conducts ultrasound training seminars for DEGUM, the German Society of Ultrasound in Medicine, and has performed over 10,000 ultrasound examinations, with an average of roughly 150 examinations per year. Examiner B completed his ultrasound training some years ago and has performed roughly 1500 examinations (50 per year). The results were analysed in a time-independent and blinded manner by the co-author (H), who neither operated nor examined the patients. RESULTS: Ultrasound correctly identified 103 of 104 complete rotator cuff tears (sensitivity: 0.99--specificity: 0.99--accuracy: 98.7%). Both examiners achieved comparable results. Moreover, 41 of 52 partial rotator cuff tears were detected preoperatively (sensitivity: 0.79--specificity: 0.91--accuracy: 88.7%). Examiner A achieved a sensitivity of 0.92, a specificity of 0.95, and an accuracy of 94.7%. The corresponding rates for examiner B were: sensitivity 0.68, specificity 0.86, and accuracy 81.3%. 16 of 23 injuries of the long biceps tendon were identified correctly (4 of 8 dislocations and 12 of 15 tears: sensitivity 0.53; specificity 0.9; accuracy 95.3%). Examiner A achieved: sensitivity 0.58; specificity 0.99; accuracy 91.7% compared to examiner B: sensitivity 0.33; specificity 0.97; accuracy 95.3%. Both examiners correctly identified all 32 cases of calcific tendinitis (sensitivity 1.0; specificity 0.98; accuracy 98.3 % examiner A: 1.0--1.0--100% examiner B: 1.0--0.96--96.2%). 22 of 28 cases of subacromial bursitis were correctly diagnosed (sensitivity 0.79; specificity 0.98; accuracy 95.8% examiner A: 0.92--0.99--98.5% examiner B: 0.69--0.97--92.5%). CONCLUSIONS: Preoperative ultrasound examination of the shoulder permits a reliable diagnosis of complete rotator cuff tears and calcium deposits (calcific tendinitis). The method is less sensitive but sufficiently reliable for the diagnosis of partial rotator cuff tears and pathology of the long biceps tendon. Examiner experience plays an important role in these special cases. Permanent continuous training in the field of ultrasound diagnosis is a prerequisite for sufficient reliability of ultrasound diagnosis of shoulder disorders.