Three-Tesla MRI diagnosis of meniscal tears of the knee

Abstract

Objective: To study and diagnose meniscal tear of the knee using 3-Tesla (3-T) magnetic resonance imaging (MRI) compared with arthroscopy Material and Method: One hundred twenty eight consecutive patients who underwent MRI of the knee using a 3-T magnet between April 2007 and Nov 2008 were included in this study. The inclusion criteria were the patients who had (i) subsequent knee arthroscopy, (ii) available medical records, and (iii) no history of meniscal surgery. Their MR images were retrospectively reviewed by two radiologists with consensus agreement. The diagnostic values for diagnosing meniscal tears were evaluated and compared to the arthroscopic results Results: Thirty-two patients (64 menisci) were included; 26 males and 6 females, mean age was 36.4 years (range 19-62). The mean interval between MRI and arthroscopy was 93 days. To diagnose tear of medial, lateral, and both menisci; the sensitivity was 100%, 90%, 100%; the specificity was 77%, 73%, 50%; the accuracy was 91%, 78%, 84%; the positive predictive value (PPV) was 86%, 60%, 81%; and the negative predictive value (NPV) was 100%, 94%, 100%, respectively. False positive MR findings were found predominantly at the posterior horn and at the peripheral third of the menisci. By dividing the patients into 2 groups according to the mean MRI-arthroscopy interval (< 93 and > 93 days): increased sensitivity and NPV of detecting lateral meniscal tear; increased specificity, accuracy, and PPV of both meniscal tear was observed in the longer duration group, but there was no statistical significance in the present study Conclusion: The present results with 3-T MRI revealed high sensitivity and NPV comparable to the literature, thus supporting previous studies that if a meniscal tear is not seen on 3-T MRI, it is highly unlikely to be present. False positive MR findings found predominantly at the posterior horn and at the peripheral third of the menisci. The longer the MRI-arthroscopy interval yielded increased diagnosing values of meniscal tear but there is no statistical significance in the present study.