Abstract

Purpose: Subscapularis (SSC) tendon tears significantly impact shoulder stability and function. While conventional MRI serves as the preferred noninvasive diagnostictool, its accuracy shows considerable variation across different populations. This study aimed to comprehensively evaluate the diagnostic accuracy of conventional 1.5-Tesla MRI for detecting SSC tears in a distinct Iranian population, using arthroscopy as the gold standard, and to identify key demographic and technical factors influencing its diagnostic performance through multivariate analysis.

Methods: A multicenter cross-sectional study was conducted across three university hospitals from 2023 to 2025. We enrolled 185 patients scheduled for shoulder arthroscopy who had undergone preoperative conventional 1.5-Tesla MRI following a standardized imaging protocol. MRI interpretations were independently evaluated by two musculoskeletal radiologists and one orthopedic shoulder surgeon, all blinded to clinical data. Arthroscopic findings served as the reference standard. We calculated sensitivity, specificity, positive and negative predictive values, and performed multivariatelogistic regression.

Results: For detecting any SSC tear, conventional MRI demonstrated an overall sensitivity of 52.3% and specificity of 85.4% across all readers. Diagnostic accuracy wassignificantly superior in older patients (>65 years) and when the dominant arm was affected. Multivariate analysis confirmed that patient age, limb dominance, and specific MRI sequence parameters independently influenced diagnostic performance.

Conclusion: Conventional MRI shows moderate sensitivity but high specificity for SSC tendon tears. Its diagnostic performance is significantly affected by specific patient factors and imaging parameters. These important findings highlight the clinical necessity for optimized imaging protocols and alternative diagnostic approaches when strong clinical suspicion persists despite negative MRI results.