Objective: The purpose of this study was to investigate the ability of ultrasound to accurately identify long head of biceps brachii (LHB) tendon abnormalities.
Materials and methods: The surgical findings for 66 consecutive patients who underwent arthroscopic shoulder surgery were retrospectively compared with the findings of preoperative ultrasound examination. Patients were excluded if more than 200 days elapsed between ultrasound and surgery. Ultrasound images were obtained using high-frequency transducers and were compared with findings at arthroscopy for the presence of LHB tendon abnormalities, including full-thickness tears, partial-thickness tears, or nontear abnormalities, including tendinosis and tenosynovitis. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated for full- and partial-thickness tears.
Results: Ultrasound correctly identified 90% (35/39) of normal LHB tendons, 88% (7/8) of full-thickness tears, 27% (3/11) of partial-thickness tears, and 22% (2/9) of tendons with nontear abnormality in those patients who underwent surgery. In the ultrasound diagnosis of full-thickness tear versus other findings, sensitivity was 0.88, specificity was 0.98, PPV was 0.88, NPV was 0.98, and accuracy was 0.97; in the ultrasound diagnosis of partial-thickness tear versus other findings, sensitivity was 0.27, specificity was 1.00, PPV was 1.00, NPV was 0.88, and accuracy was 0.88, in those patients who underwent surgery.
Conclusion: Shoulder ultrasound is an accurate method to confirm a normal biceps tendon or full-thickness tear, but is less accurate in the diagnosis of partial-thickness tear and nontear abnormalities.