Correlation of magnetic resonance imaging and arthroscopic findings in patients with soft tissue knee injuries

The knee is indispensable in everyday life and injuries to it can be debilitating with significant loss of earnings incurred. Clinical diagnosis may not always be made with certainty and Magnetic Resonance Imaging (MRI) helps further elucidate intra-articular injuries. MRI reporting has its shortcomings and may provide spurious results according to the interpretor’s level of experience. This study aims to test the diagnostic reliability of MRI done in a teaching hospital for the evaluation of anterior cruciate ligament and meniscal injuries using arthroscopy as the baseline for comparison. Due to the long waiting times to have surgery we also determined if there was a change in the reliability of an MRI result as time elapsed. A retrospective review of records of patients who had knee arthroscopies at Chris Hani Baragwanath Academic Hospital (Johannesburg, South Africa) from May 2009 to May 2015 was done. Adults (16 − 60 years) with one major episode of trauma to the knee and had MRI done prior to surgery at the above institution were included. Arthroscopy was performed by 2 senior surgeons or by residents under their direct supervision. Arthroscopic findings of anterior cruciate ligament (ACL) and medial (MM) or lateral meniscal (LM) injuries were compared to MRI findings. Data was analysed by STATA version 13.1 to determine injury demographics, sensitivity, specificity and diagnostic accuracy of MRI. The effect of time interval from MRI to surgery on the diagnostic accuracy was determined. A total of 72 patients (74 knees) qualified for review. The median age was 35 years (IQR 26 − 43) with a significant difference between males and females (28 vs 41 years, p = 0.0019). Leading causes of injury were traffic accidents (32.4%), falls (27.0%) and sports injuries (17.6%). Median interval from MRI to surgery was 71.5 days (IQR 29 − 143). The sensitivity of MRI for ACL, MM and LM injuries was (63.6%, 58.8% and 52.6%), specificity (92.7%, 86.0% and 80.0%) and diagnostic accuracy (79.7%, 79.7% and 73.0%) respectively. The patients were divided into subgroups of early (< 6 weeks), intermediate (6 − 16 weeks) and late intervention (> 16 weeks) post-MRI. There were marked differences in the diagnostic accuracy in the three groups for the ACL (70.8% vs 92.6% vs 73.9%) and LM (62.5% vs 81.5% vs 73.9%). This was unremarkable for the MM (75.0% vs 81.5% vs 82.6%). MRI findings correlate well with arthroscopic findings making it a reliable preoperative screening tool for ACL and meniscal injuries. However its diagnostic accuracy appears to change with time. It is apparent that the diagnostic accuracy is higher between 6 − 16 weeks post MRI. A bigger cohort would help determine an ideal waiting time interval without significant depreciation in diagnostic accuracy.
A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Medicine Johannesburg, 2018.