A systematic review 1 included 20 studies on suspected meniscal tears (n=3,171), 18 studies on suspected rupture of the anterior cruciate ligament (n=2,475), 5 studies on the rupture of the posterior cruciate ligament (n=1,426) and 10 studies on articular cartilage lesions (n=452).
In meniscal tears the sensivity of MRI ranged from 71 to 100%, and specificity from 37 to 100%. At least 85% of lesions found at arthroscopy were also detected with MRI and 90% of menisci found healthy were also found healthy by MRI. In rupture of the anterior cruciate ligament the sensitivity of MRI ranged drom 90 to 100% with two double blind studies reporting 100% sensitivities. However, one study reported a sensitivity of 61%. Specificity was usually greater than 80%. The netative predictive value ranged from 70 to 100%, and the positive predictive value from 57 to 100%, averaging 75 to 80%. Data were limited for the posterior cruciate ligament, but MRI appeared to be a powerful diagnostic tool, with NPV and overall accuracy of 99% in one study and sensitivity and specificity of 100% in another. For articular cartilage lesions, sensitivity ranged from 18 to 100%, specificity from 50 to 100%, PPV from 72 to 88%, and NPV from 66 to 100%.
Another systematic review 3 abstracted in DARE included 22 studies (13 prospective studies with a total of 4,094 patients, and 9 retrospective studies with a total of 947 patients). MRI was compared to arthroscopy. The overall sensitivity of menisci and cruciate ligaments was 0.88 with a 95% CI of 0.87 to 0.90. The overall specificity was 0.94% (95% CI 0.93 to 0.94). The sampling error varied widely amongst studies.
Comment: The quality of evidence is upgraded by large magnitude of effect. There are limitations in review methdology, but there is marked consistency in the results of different studies.
Primary/Secondary Keywords