Abstract
Rotator cuff tears involving the supraspinatus tendon are a common cause of shoulder pain. Ultrasound (US) is accessible, dynamic, and less costly than MRI, but its diagnostic performance may vary with tear severity and operator expertise. This systematic review and meta-analysis evaluated the diagnostic performance of US for supraspinatus tears using MRI as the non-invasive reference imaging standard. MEDLINE, Embase, Scopus, and CINAHL Plus were searched for studies published from January 2015 to June 2026, and study quality was assessed using QUADAS-2. A total of 23 studies involving 1452 participants were included. Mean sensitivity, specificity, and accuracy were 89.7%, 97.3%, and 95.4%, respectively, for full-thickness tears, compared with 71.6%, 93.1%, and 80.2% for partial-thickness tears. HSROC analysis excluding studies with high overall risk of bias showed pooled sensitivity and specificity of 83% and 88% for any tear, 67% and 91% for partial-thickness tears, and 72% and 97% for full-thickness tears. Agreement was generally more consistent for full-thickness tears, while operator expertise and study quality varied. Overall, US demonstrated stronger diagnostic performance for full-thickness than partial-thickness supraspinatus tears and may be considered an accessible initial imaging option in appropriate clinical settings, with MRI remaining complementary for equivocal or subtle partial-thickness tears and preoperative assessment.