Onur KAYA, Mert Erdem YEŞİLLİ, Nevzat GÖNDER, Bilgin BOZGEYİK, Tahir YAVUZ
Çukurova Anestezi ve Cerrahi Bilimler Dergisi - 2026;9(2):476-486
Objective: Magnetic resonance imaging (MRI) is the primary modality for preoperative evaluation of anterior cruciate ligament (ACL) injuries. However, interobserver variability in MRI interpretation may influence diagnostic outcomes. This study aimed to compare the diagnostic sensitivity of preoperative MRI radiology reports with independent orthopaedic surgeon evaluations for detecting complete ACL tears, and to assess interobserver agreement in meniscal assessment. Methods: This single-centre retrospective observational agreement study included 96 patients who underwent ACL reconstruction with arthroscopically confirmed complete ACL tears. Preoperative MRI images were independently evaluated by three orthopaedic surgeons blinded to clinical data. Agreement with the arthroscopic reference standard was analysed using the McNemar test. Interobserver agreement was assessed using Cohen's kappa, Fleiss' kappa, and Spearman correlation coefficients. Results: Radiology reports demonstrated a sensitivity of 28.1% for detecting complete ACL tears, compared with 82.3-87.5% for orthopaedic surgeons (p<0.001). Agreement between radiology reports and orthopaedic surgeons was poor (kappa=0.052-0.120). In meniscal assessment, interobserver agreement among orthopaedic surgeons ranged from poor to substantial for both the medial (kappa=0.099-0.585) and lateral meniscus (kappa=0.207-0.770), whereas agreement with radiology reports remained poor. Conclusion: In this selected patient cohort, preoperative MRI reports demonstrated substantially lower sensitivity than orthopaedic surgeon evaluations for detecting complete ACL tears. Interobserver agreement in meniscal assessment was variable. These findings emphasise the importance of integrating MRI findings with clinical evaluation and highlight the impact of observer-dependent variability in MRI interpretation.
