肱二头肌长头腱撕裂的肩关节MRI评价

目的 评价肩关节MRI对肱二头肌长头腱撕裂的诊断价值,并对比分析肩关节MR造影和常规肩关节MRI的诊断性能.方法 由2名影像诊断医师独立回顾分析215例肩关节MR图像,包括107例肩关节MR造影和108例常规肩关节MRI,分析结果与肩关节镜手术相比较.计算MRI评价肱二头肌长头腱撕裂的诊断敏感性、特异性和准确性.采用Kappa分析2名影像诊断医师评价的一致性.结果 215例中,肱二头肌长头腱完伞撕裂7例,部分撕裂29例,无撕裂179例.医师甲用肩关节MRI评估肱二头肌长头腱撕裂(包括完全和部分)的敏感性、特异性和准确性分别为72.2%(26/36)、91.6%(164/179)和88.4%(190/215),医师乙分别为80.6%(29/36)、93.8%(168/179)和91.6%(197/215),2名医师的评估结果间一致性好(Kappa=0.681).肱二头肌长头腱完全撕裂,2名医师的诊断敏感性、特异性和准确性均为100%(7/7)、100%(208/208)和100%(215/215).利用肩关节MR造影评价肱二头肌长头腱撕裂时,医师甲的准确性为93.4%(100/107),医师乙为96.3%(103/107),均明显高于常规肩关节MRI的结果医师甲为83.3%(90/108),医师乙为87.0%(94/108),P值均<0.05].结论 肩关节MRI是评价肱二头肌长头腱撕裂比较可靠的方法,而且肩关节MR造影的准确性优于常规肩关节MRI.

Tear of the long head of the biceps brachii tendon: evaluation with shoulder MRI

Objective To assess the diagnostic value of MRI for the long head of the biceps brachii tendon tear,and to compare the diagnostic efficiency between routing MRI and MR arthrography. Methods A retrospective study was conducted in 215 cases with shoulder MR examination (107 with MR arthrography, and 108 with routing MRI) and subsequent shoulder arthrescopy and surgery. Two radiologists analyzed all MR examinations independently, and the results were compared with those of arthroscopy and surgery. Sensitivity, specificity, and accuracy were calculated. Kappa values were used to quantify the interobserver agreement. Results Based on the results of arthroacopy and surgery, 215 patients comprised 7 cases of complete tear of the long head of the biceps brachii tendon, 29 cases of partial tear, and 179 cases without tear. The overall sensitivity, specificity, and accuracy for the diagnosis of tears (complete and partial tear) of the long head of the biceps brachii tendon were 72.2% (26/36), 91.6% (164/179), and 88.4% (190/215) respectively for observer 1, 80.6% (29/36), 93.8% (168/179), and 91.6% (197/215) respectively for observer 2. The interobserver agreement was good (Kappa value=0.681). For the complete tear of the biceps brachii tendon, the sensitivity, specificity, and accuracy were all 100% (7/7), 100% (208/208), and 100% (215/215) for both observers. For the tear of the long head of the biceps brachii tendon, the accuracy of MR arthrography were 93.4% (100/107) for observer 1 and 96.3% (103/107) for observer 2. They were higher than the Accuracy of routing MRI, which were 83.3% (90/108)and 87.0% (94/108) respectively for two observers (P<0.05). Conclusion Shoulder MRI is a moderate reliable method for evaluating the tear of the long head of the biceps brachii tendon, and the accuracy of MR arthrography is found to be superior to that of routine MRI.