False-negative meniscal tear in MR imaging using non fat-suppressed techniques.

ABSTRACT

Magnetic resonance (MR) imaging of menisci has a diagnostic accuracy of more than 90 per cent when the images are properly obtained. Errors in the interpretation of MR images may be related to many problems, resulting in false positive or false negative readings. We conducted this study to evaluate the prevalence of false negative results, using arthroscopy as gold standard. Matched MR imaging (performed before July 1997) and arthroscopic findings of the menisci in 63 consecutive patients were retrospectively reviewed separately by imaging planes and sequences used, by the site of tear, and by the site of meniscus (medial or lateral meniscus). The number and percentage of false-negative results were recorded. We found that with non fat-suppressed MR techniques, missed tear (false negative reading) of both menisci occurred predominantly in the meniscal body. The prevalence depends on imaging plane and sequence used. Among the techniques reviewed, sagittal T2-weighted (T2W) sequence had the highest prevalence of undetected meniscal tear.