Clinical and MRI findings associated with false-positive knee MR diagnoses of medial meniscal tears.

OBJECTIVE

The objective of our study was to determine if false-positive MR diagnoses of a medial meniscal tear are more common with specific clinical variables, tear type or location, or MRI findings of a longitudinal tear.

MATERIALS AND METHODS

We reviewed the records of 559 patients who underwent knee MR examinations and arthroscopy. We compared the positive predictive values (PPVs) of an MR diagnosis of a medial meniscal tear for differences in tear location or type, delay between knee injury and MRI, delay between MRI and arthroscopy, and the presence of an anterior cruciate ligament (ACL) tear. We also retrospectively reviewed the MR examinations of 50 longitudinal tears to compare the PPVs of various MRI findings of a longitudinal tear.

RESULTS

There was no association between either the delay between injury and MRI or the delay between MRI and arthroscopy and false-positive diagnoses. The PPV of 64% (32/50) for longitudinal tears was lower than the values of 83% (15/18) to 100% (116/116) for other types of medial meniscal tears. False-positive diagnoses of medial meniscal tears were more common in patients who had a prior episode of acute trauma (p = 0.004) or an ACL tear (p < 0.0001). Review of longitudinal tears revealed a decreased PPV when MRI showed signal contacting only the superior surface (p = 0.016) or when MRI showed signal contacting the surface at the meniscocapsular junction (p = 0.004). Four of the 18 menisci with a false-positive diagnosis of a longitudinal tear had a healed ACL tear noted at arthroscopy.

CONCLUSION

False-positive MR diagnoses of medial meniscal tears are more common for longitudinal tears than other tear types and are also more common with MR abnormalities at either the superior surface or the meniscocapsular junction. Spontaneous healing of longitudinal tears accounts for some false-positive MR diagnoses.