Meniscal tears diagnosed with MR imaging versus arthroscopy: how reliable a standard is arthroscopy?

Arthroscopy of the menisci is considered the standard by which all noninvasive imaging procedures of the knee are measured. To better recognize the accuracy and relevance of arthroscopic correlation, a study was done in 254 consecutive patients who underwent both magnetic resonance (MR) imaging of the knee and arthroscopy. If interpretation of arthroscopic findings disagreed with MR findings, the arthroscopic videotapes were reviewed. Seventeen errors occurred in interpretation, resulting in false-positive MR images. Nine of these errors (53%) were associated with lines of high signal intensity that reached the articular surface of the posterior horn, an area incompletely seen on the arthroscopic videotapes. Two errors (12%) occurred in postmenisectomy abnormalities that the arthroscopist believed were not clinically significant; the six other errors (35%) occurred in high-signal-intensity alterations in either the free edge or the entire meniscal substance that did not correlate with clinically significant tears. Thirty-two errors occurred in interpretation, resulting in false-negative MR images. Fifteen of these errors (47%) occurred in small tears that did not require meniscectomy.