MRI versus arthroscopy in evaluating knee meniscal pathology.

The purpose of this project was to evaluate magnetic resonance imaging (MRI) accuracy for diagnosing and defining meniscal pathology. The charts of 45 patients who underwent both MRI and arthroscopy were viewed retrospectively. All cases were seen by one of two orthopedic surgeons. Magnetic resonance imaging scans were reviewed by staff radiologists at the University of Michigan Hospital Medical Center. Arthroscopy was used as the standard for comparing MRI results. Data were analyzed two ways with respect to questionable (code 3) MRI lesions. First, all code 3 lesions were interpreted as either true positives or negatives. Second, all code 3 lesions that corresponded with a negative arthroscopy finding were interpreted as false positives. The MRI accuracy was 87.8% when considering nonmatching code 3 lesions as true positives or negatives. The accuracy declined to 82.2% when considering code 3 lesions as false negatives. Magnetic resonance imaging sensitivity for medial meniscal and lateral meniscal tears was 84.6% and 87.5%, respectively, regardless of code 3 interpretation. The specificity for medial meniscal tears with and without the negative effect was 71.9% and 84.4%, respectively. The specificity with and without considering code 3 lesions as false positives was 89.2%, and 91.9%, respectively.