Background: MRI (magnetic resonance imaging) is widely used to diagnose meniscal pathology and ACL (anterior cruciate ligament) tears. Because of the enhanced signal-to-noise ratio and improved image quality at higher field strength, knee MRI equipment is shifting from 1.5 to 3.0 T. To date, objective evidence of improved diagnostic ability at 3.0 T is lacking. The purpose of this prospective study was to assess the accuracy of 1.5 and 3.0-T MRI of the knee, in the same individuals, for diagnosing meniscal pathology and ACL tears, utilizing arthroscopy as the reference standard.
Methods: Two hundred patients underwent MRI of the knee at 1.5 and 3.0 T. All MRI examinations consisted of multiplanar turbo spin-echo sequences. One hundred patients underwent subsequent knee arthroscopy. Two blinded independent radiologists assessed all MRI studies to identify meniscal pathology and ACL tears. In patients with MRI results indicating the need for surgical treatment, the sensitivity and specificity of the 1.5 and 3.0-T protocols for detecting these lesions were determined, utilizing arthroscopy as the reference standard, and compared with use of the McNemar test. The kappa statistic for inter-reader agreement in the 200 patients was calculated.
Results: For medial meniscal tears, the mean sensitivity and specificity for the two readers were 93% and 90%, respectively, at 1.5 T and 96% and 88%, respectively, at 3.0 T. For lateral meniscal tears, the mean sensitivity and specificity were 77% and 99%, respectively, at 1.5 T and 82% and 98%, respectively, at 3.0 T. For ACL tears, the mean sensitivity and specificity were 78% and 100%, respectively, at 1.5 T and 80% and 100%, respectively, at 3.0 T. None of the values for either reader differed significantly between the 1.5 and 3.0-T MRI protocols. Inter-reader agreement was almost perfect to perfect (kappa = 0.82 to 1.00).
Conclusions: Routine use of a 3.0-T MRI protocol did not significantly improve accuracy for evaluating the knee menisci and ACL compared with a similar 1.5-1 protocol.