PURPOSE: To compare the diagnostic value of magnetic resonance arthrography (MRA) in different shoulder lesions using arthroscopy as gold standard. METHODS: We performed a comprehensive search in Cochrane, Scopus, PubMed, and Web of Science databases for articles that reported the diagnostic value of MRA in diagnosing labral tears, rotator cuff tears (RCTs), Hill-Sachs, and Bankart injuries. We used arthroscopic surgery as a reference standard for comparison. The quality of the studies was evaluated using the Quality Assessment of Diagnostic Accuracy Studies tool. We pooled the collected data and used Stata/MP17 to generate summary statistics. RESULTS: We identified a total of 53 articles compromising 5,487 patients. The sensitivity and specificity for Hill-Sachs lesions were 0.94 (95% confidence interval [CI] 0.80-0.99) and 0.89 (95% CI 0.77-0.95), Bankart lesions were 0.94 (95% CI 0.89-0.97) and 0.99 (95% CI 0.90-1.00), anterior labral tears were 0.91 (95% CI 0.82-0.96) and 0.96 (95% CI 0.90-0.99), posterior labral tears were 0.74 (95% CI 0.58-0.86) and 0.98 (95% CI 0.86-1.00), superior labral tears were 0.77 (95% CI 0.62-0.88) and 0.83 (95% CI 0.54-0.95), SLAP lesions were 0.86 (95% CI 0.80-0.90) and 0.91 (95% CI 0.83-0.96), partial-thickness RCTs were 0.67 (95% CI 0.45-0.83) and 0.92 (95% CI 0.78-0.97), whereas full-thickness RCTs were 0.97 (95% CI 0.91-0.99) and 0.99 (95% CI 0.93-1). CONCLUSIONS: In conclusion, MRA is a sensitive and specific imaging method for identifying anterior labrum lesions, full-thickness RCTs, Bankart, SLAP, and Hill-Sachs lesions, with lower accuracy in partial-thickness RCTs, both posterior and superior labrum. LEVEL OF EVIDENCE: Level III, meta-analysis of prospective and retrospective cohort and case-control studies.