[Comparison of the Arthroscopic Finding in the Knee Joint and the MRI - Retrospective Study].

: PURPOSE OF THE STUDY Our retrospective study presents the comparison of the preoperative magnetic resonance imaging of the knee joint - MRI - and the arthroscopic finding - ASC. Its aim is to find out how a positive or a negative finding of MRI corresponds with the operative finding and how much the experience of radiologist contributes to the conformity. MATERIAL AND METHODS The MRI findings of knee joints treated surgically at two departments in 2013 and 2014 were assessed. The MRI was performed in a total of 470 patients who subsequently underwent an arthroscopic surgery. A conformity or a non-conformity in anterior, posterior horn and complete rupture of both menisci and in partial or complete tear of anterior cruciate ligament - LCA was searched for. The sensitivity, specificity and accuracy of MRI were established. The difference between experienced and less experienced radiologists was evaluated. The analysis of the radiology report, surgical protocol of ASC and medical history in the documentation was performed. The cartilage was not subject to evaluation. The finding of Grade 1 meniscus tear on MRI was evaluated as negative. Grade 2 and Grade 3 were evaluated as positive. RESULTS Comparison of the preoperative MRI and the arthroscopic finding 1. The group with MRI reported 3 % of diagnostic arthroscopies. The control group without MRI (551 ASC) reported 15 % diagnostic arthroscopies. 2. Low sensitivity of MRI (0.67) in negative findings of ASC. It concerned 7 cases in which a pathological finding was identified on a MRI scan, but not by ASC. In two cases the repeated arthroscopy confirmed that a pathology inside the knee joint was overlooked by the arthroscopist. In the remaining five cases, the clinical finding improved without a repeated surgery. 3. High sensitivity of MRI is shown in the most frequent finding - posterior horn of medial meniscus (0.94). 4. Lower sensitivity (0.76) in partial and (0.83) in complete ACL tear. It increases to 0.93 if partial and complete tear are put together. Both MRI and ASC detect the pathology of ligament, but do not agree in terms of terminology. 5. Lower sensitivity (0.78) was seen in posterior horn of lateral meniscus, most likely due to its complicated anatomy. 6. Specificity of complete tear of medial meniscus tear is 0.99. Lower specificity in the posterior horn of medial meniscus (0.81) shows a higher number of positive MRI findings in negative ASC findings. Some posterior horn tear can be overlooked by an inexperienced surgeon. The MRD findings need to be studied. 7. High specificity (0.99) was described in negative findings. In three cases only, the surgeon discovered a pathological finding, not revealed by MRI scan. It always concerned a tear within the posterior horn of the medial meniscus. 8. Specificity (0.88, 0.93 or 0.86, respectively) in partial, complete and all ACL damages in total. 9. We concluded that contributing to the degree of agreement between MRI and ASC is also the experience of a radiologist. The most experienced radiologist evaluated 190 of 470 MRI scans, the remaining twelve radiologists assessed 280 scans. The posterior horn of the medial meniscus - sensitivity or specificity evaluated by an experienced radiologist (0.98 and 0.88, respectively) and inexperienced radiologist (0.91 and 0.79). The specificity and sensitivity in complete ACL tears - by experienced radiologist (0.91 and 0.94, respectively) and inexperienced radiologist (0.81 and 0.90, respectively). The test accuracy of the experienced radiologist in evaluating the most frequent injuries of soft knee structures was by 9-10 % higher than of the inexperienced radiologist. DISCUSSION The results obtained by the other authors show that the sensitivity and specificity range from 0.6 to 0.9. They agree that the MRI is unsuitable for assessing the cartilage. We confirm that the results are worse when evaluating the posterior horn of the lateral meniscus. We have also proven that the radiologist s experience does play an important role. It can be the reason for a high degree of difference between the results of various authors. Some of them give preference to a clinical examination or a diagnostic arthroscopy instead of the MRI. At our department, MRI is indicated if we are convinced it can help with the indication or where it will suggest what to focus on during the surgery. CONCLUSIONS A preoperative MRI scan can prevent an unnecessary arthroscopy. It displays structures to the surgeon which shall be reviewed in detail during the surgery. We recommend paying attention to Grade 2 MRI findings, positive MRI findings on the posterior horn of medial meniscus and to MRI findings on partial ACL tears. Such menisci and ligaments shall be carefully reviewed. Clinical preoperative examination and cooperation between the surgeon, the "arthroscopist", and the radiologist is essential. The experience of the radiologist also plays a role when evaluating the MRI scan. We have introduced MRI ward rounds. Key words: MRI, knee joint, knee arthroscopy, sensitivity, specificity, accuracy.