Arthroscopy and MRI for the knee.

In its early days arthroscopy was used mainly for diagnosis; it saved many knees from unnecessary arthrotomy. Today, MRI is available to save knees from unnecessary arthro-scopy; internal derangements can then be treated by arthro-scopic surgery. Arthrotomy for internal derangement is now obsolete and the debate has moved on to the relative roles of arthroscopy and MRI in the management of knee disorders. MRI is non-invasive, free from known morbidity and is safer and less expensive than arthroscopy. The sensitivity of MRI for meniscal lesions may exceed 90%, 1-3 but despite this, diagnostic arthroscopy is often advised too readily. One study showed that 51% of patients on a waiting-list for arthroscopy for a suspected meniscal lesion were removed after review and MRI. 4 Should all knees be examined by MRI before arthroscopic surgery? Sensitivity is not the same as accuracy. Some authors 5,6 have reported that clinical assessment is more accurate than MRI; others have found no difference. 7 The accuracy of the clinical diagnosis of meniscal tears is about 75% to 80%, 5,6 compared with 88% to 90% for MRI. 2 For lesions of articular cartilage, both the sensitivity and accuracy of MRI are low. 8,9 Symptomatic meniscal lesions demonstrated by MRI will still need arthroscopic surgery and it can be argued that the MRI was unnecessary. Many needless arthroscopies will be performed if every MRI report is taken at face value. The clinical problem is to try to avoid MRI for patients who definitely need therapeutic arthroscopy and yet to prevent invasive arthroscopy when there is no surgically treatable lesion. Carmichael et al review the cost-effectiveness of MRI in a paper on p. 624. They compared the proportions of negative arthroscopies and positive MRIs for a surgeon with a special interest in the knee with those for a group of general orthopaedic surgeons. The specialist knee surgeon performed fewer negative arthroscopies and had more positive MRI scans than his colleagues. This suggests that better clinical skills allowed a more economical use of resources. The study did not show how many patients were managed without either arthroscopy or MRI; a detailed review to demonstrate the value of this clinical judgement would be instructive. Only an experienced clinician can decide if a lesion demonstrated by MRI is the likely cause of symptoms. Reports by radiologists may present particular difficulties: a bright signal due to myxoid degeneration of the posterior third …