MRI preferable to diagnostic arthroscopy in knee joint injuries. A double-blind comparison of 47 patients.

We compared the findings of low-field MRI of the knee with those of subsequent arthroscopy. In a double-blind set-up, 47 patients with knee joint injuries were enrolled. Two radiologists independently interpreted the MRI examinations and consensus was obtained in case of discrepancy. Arthroscopy was performed without knowledge of the MRI findings. The accuracy rates of MRI for evaluating the medial meniscus, lateral meniscus and anterior cruciate ligament were 77%, 91% and 96%, respectively, when arthroscopy was considered the "golden standard". When MRI was considered the standard, the figures for arthroscopy were 74%, 91% and 96%. MRI found the indication for treatment in 18 of 21 patients who were treated at the arthroscopy. In 17 patients, neither MRI nor arthroscopy detected any lesion. In the remaining 9 patients, MRI demonstrated a lesion, but no lesion was found at the subsequent arthroscopy. Our conclusion is that low-field MRI can be used as a first-line diagnostic examination in patients with suspected meniscus or cruciate ligament injuries and thus a substantial number of negative diagnostic arthroscopies can be avoided.