The Association between Knee MRI Interpretations vs. Arthroscopic Examination: a Retrospective Analysis of 250 Cases

Introduction: Knee MRI is valuable in preoperative planning of arthroscopic surgery. However, disagreement between MRI interpretations and arthroscopic findings may affect clinical decisions and treatment strategy, but its extent is not well-established for specific knee pathologies. The purpose of this study was therefore to determine the association between preoperative MRI interpretations versus knee arthroscopic findings.

Materials and Methods: Operative reports of all knee arthroscopies performed in a single arthroscopy unit between 2013 and 2016 and their respective preoperative MRI interpretations were reviewed. Inclusion criteria were: (1) Adult population; (2) primary arthroscopy; (3) MRI performed in similar institute; and (4) interval MRI-surgery shorter than 6 months. Specific pathologies reviewed included: ACL tears, meniscal tears, discoid menisci, cartilage lesions, osteochondral lesions, loose bodies, benign tumors, and medial plicas. Kappa coefficients were calculated for analyzing the association of each lesion between the modalities. Statistical significance was set at p=0.05.

Results: Inclusion criteria were met by 250 cases. Preliminary analysis demonstrated that the highest association among all lesions was related to ACL tear (93% agreement, Kappa=0.85). Among meniscus lesions, the highest association was related to posterior horn of medial meniscus (84% agreement, Kappa=0.64), and the lowest was related to anterior horn of lateral meniscus (81% agreement, Kappa=0.15). Chondral lesions as a group showed a relatively low association (Kappa=0.1), of which the LTC had the lowest one. Medial plicas also had low association, which was the case for loose bodies as well.

Conclusions: Preoperative MRI examination is effective for identifying ACL and medial meniscus posterior horn lesions. However, when anterior horn meniscal lesions, chondral lesions, medial plicas, or loose bodies are the main pathology suspected for which surgery is indicated, additional tools should be considered to improve preoperative diagnostic accuracy, such as the addition of intra-articular gadolinium injections or other tissue-specific sensitive modalities.