DIAGNOSIS (IMAGING)
MRI, MRA, CT and ultrasound are useful adjuncts to a clinical exam and radiographs for identifying rotator cuff tears.

Rationale

Ultrasound
Evidence from seven high and 25 moderate quality studies evaluating the use of ultrasound for diagnosing rotator cuff tears indicates that ultrasound can be a useful adjunct to clinical exam and radiographs for identifying rotator cuff tears (Yazigi Junior, 2018; Porta, 2023; Day, 2016; Chauhan, 2024; Frere, 2021; Cheng, 2015; Saragaglia, 2021; Wang, 2022, Sabharwal, 2019; Zhu, 2022; Gill, 2023; Gormeli, 2014; Moosmayer, 2007; Aminzadeh, 2020; El-Shewi, 2019; Duan, 2021; Agarwal, 2021; Gulshan, 2019; Wang, 2023a; Yoon, 2020; Shapla, 2024; Ahmad, 2018; Reddy, 2022; Prakash, 2024; Reddy, 2024; van der Kraats, 2023; Dirkx, 2020; Wang, 2023b; Tang, 2019; Li, 2023a; Li, 2023b, Waldt 2007). The positive likelihood ratios (LR+) of the studies were generally found to be in the strong and moderate range with the negative likelihood ratios demonstrating a greater spread from poor to strong.

Magnetic Resonance Arthrography (MRA)
Evidence from 2 high and 12 moderate quality studies evaluating MRA for the diagnosis of rotator cuff tears indicates that MRA can be a useful adjunct to clinical exams and radiographs for identifying rotator cuff tears. (Duc 2006, Hahn, 2021, Magee, 2016; Groarke, 2021; Magee, 2018; Dirkx, 2020; Schreinemachers, 2009; Lee, 2018; Pfirrmann, 1999; Waldt, 2007; Herold, 2006; Probyn, 2007; Singer, 2020; Khil, 2021). A meta-analysis of 6 studies (appendix 2) analyzing the use of MRA for diagnosing all size RC tears, reported the pooled LR+ to be 20.87 (95% CI 6.65-65.47; i2 = 51.26) with the majority of studies analyzed reporting LR+ values greater than 10, suggesting that MRA is a strong rule in test. Furthermore, the studies also reported LR- values in the weak to strong range with the majority being moderate or strong.

Magnetic Resonance Imaging (MRI)
Evidence from eight high and thirty-three moderate quality studies evaluating MRI for the diagnosis of rotator cuff tears indicates that MRI can be a useful adjunct to clinical exam and radiographs for identifying rotator cuff tears (Bullock, 2018; Lenz, 2021; Sill, 2022; Nigues, 2022; Haider, 2024; Hahn, 2022; Zappia, 2021; Ercan, 2021, Mohtadi, 2004; Lee, 2021; Apostolopoulos, 2019; Beer, 2021; Wang, 2022; Incesoy, 2021; Ahn, 2022; Hasegawa, 2023; Jeong, 2018; Kim, 2021; Sugimori, 2022; Atinga, 2021; Kilic, 2024a; Kilic, 2024b; Ringshawl, 2020; Siriwanarangsun, 2023; Fazal Gafoor, 2023; Jung, 2024; Mi, 2024; Sabharwal, 2019; Zhu, 2022; Gill, 2023; Li, 2023a; Magee, 2016; Groarke, 2021; Magee, 2018; Dirkx, 2020; Wagner, 2021; Chowdhury, 2023; Tang, 2019; Li, 2023b; Thiagarajan, 2021; Gowda, 2024). A meta-analysis of 8 studies (appendix 2) analyzing the use of MRI for diagnosing all size RC tears reported the pooled LR- to be 0.10 (95% CI 0.06-0.14; i2 = 8.15) suggesting that MRI is a strong rule out test. Furthermore, the studies also reported LR+ values in the poor to strong range with the majority being in the weak to strong range.

CT
Evidence from five moderate quality studies evaluating CT for the diagnosis of rotator cuff tears indicates that CT can be a useful adjunct to clinical exam and radiographs for identifying rotator cuff tears (Allam, 2019; Dirkx, 2020; Magee, 2018; Ma, 2022; Gomez-Vieira, 2019). This evidence was upgraded to be included with this strong recommendation because it is widely available and may be demanded clinically where MRI is not available. Furthermore, the use of CT in diagnosing RC tears may be useful for patients with cardiac restrictions.