Diagnostic accuracy of ultrasonography, MRI and MR arthrography in the characterisation of rotator cuff disorders: a systematic review and meta-analysis

• Published in: British Journal of Sports Medicine Vol. 49; no. 20; pp. 1316 - 1328
• Main Authors: Roy, Jean-Sébastien, Braën, Caroline, Leblond, Jean, Desmeules, François, Dionne, Clermont E, MacDermid, Joy C, Bureau, Nathalie J, Frémont, Pierre
• Format: Journal Article Book Review
• Language: English
• Published: England BMJ Publishing Group 01.10.2015; BMJ Publishing Group LTD
• Subjects: Accuracy; Arthrography - methods; Bias; Clinical trials; Diagnostic Imaging - methods; Diagnostic tests; Estimates; Humans; Joint Diseases - diagnosis; Magnetic Resonance Imaging - methods; Medical imaging; Meta-analysis; Methods; NMR; Nuclear magnetic resonance; Pain; Rehabilitation; Review; Rotator Cuff; Rupture - diagnosis; Sensitivity and Specificity; Shoulder; Shoulder Joint - injuries; Sports injuries; Studies; Surgery; Systematic review; Tendinopathy - diagnosis; Tendon Injuries - diagnosis; Tendons; Ultrasonic imaging; United States > US; MRI; Ultrasound; Shoulder; Meta-analysis
• ISSN: 0306-3674; 1473-0480
• DOI: 10.1136/bjsports-2014-094148

BackgroundDifferent diagnostic imaging modalities, such as ultrasonography (US), MRI, MR arthrography (MRA) are commonly used for the characterisation of rotator cuff (RC) disorders. Since the most recent systematic reviews on medical imaging, multiple diagnostic studies have been published, most using more advanced technological characteristics. The first objective was to perform a meta-analysis on the diagnostic accuracy of medical imaging for characterisation of RC disorders. Since US is used at the point of care in environments such as sports medicine, a secondary analysis assessed accuracy by radiologists and non-radiologists.MethodsA systematic search in three databases was conducted. Two raters performed data extraction and evaluation of risk of bias independently, and agreement was achieved by consensus. Hierarchical summary receiver-operating characteristic package was used to calculate pooled estimates of included diagnostic studies.ResultsDiagnostic accuracy of US, MRI and MRA in the characterisation of full-thickness RC tears was high with overall estimates of sensitivity and specificity over 0.90. As for partial RC tears and tendinopathy, overall estimates of specificity were also high (>0.90), while sensitivity was lower (0.67–0.83). Diagnostic accuracy of US was similar whether a trained radiologist, sonographer or orthopaedist performed it.ConclusionsOur results show the diagnostic accuracy of US, MRI and MRA in the characterisation of full-thickness RC tears. Since full thickness tear constitutes a key consideration for surgical repair, this is an important characteristic when selecting an imaging modality for RC disorder. When considering accuracy, cost, and safety, US is the best option.