Magnetic resonance imaging of the knee

Magnetic resonance imaging has had an enormous impact on musculoskeletal imaging and in this area the knee is the most frequently imaged joint. The steadily increasing availability of magnetic resonance imaging is moving the investigation from the realms of the last resort of the hospital specialist to part of the diagnostic evaluation by the general practitioner.

Magnetic resonance imaging of the knee is most commonly indicated in patients with suspected injuries of the menisci and cruciate ligaments. Plain radiographs have little value unless there has been an injury due to direct impact. In teaching centres where dedicated musculoskeletal radiologists report on images, diagnostic accuracy of 90% can be achieved for damage to the medial meniscus and anterior cruciate ligaments, slightly less for the lateral meniscus and slightly more for the posterior cruciate ligament.^1–6

The contribution that this level of accuracy can make to therapeutic decisions has been shown in several studies. MacKenzie et al studied orthopaedic diagnoses before and after magnetic resonance imaging in 332 patients.⁷ Clinicians were asked to indicate their clinical diagnosis, level of confidence, and the proposal for management before imaging. In meniscal tears, 57 of 113 pre-imaging diagnoses were no longer considered after imaging, resulting in a change in management in 62% of patients. For confirmed diagnoses, confidence in the diagnosis improved significantly. The proportion of patients for whom arthroscopy was being considered changed considerably, with only 38% proceeding to arthroscopy after imaging.

Carmichael and Warwick have reported similar results in smaller studies.⁸,⁹ Weinstabl et al randomised patients with positive clinical tests for meniscal tears into two groups.¹⁰ In one group all patients underwent preliminary magnetic resonance imaging, which determined the need for arthroscopy. In this group only 2% of the patients who subsequently underwent arthroscopy had findings of importance at surgery. Patients in the second group were managed on the basis of their clinical findings alone. Of patients who had arthroscopy on the basis of positive clinical tests 30% were found to have no finding of importance. These findings confirm that magnetic resonance imaging is indicated even if clinical signs point to an important internal derangement.

Preliminary data from a multicentre trial in the Netherlands also show that management is changed in patients with persistent knee pain in whom clinical findings are minimal and important disease is not expected.¹¹ Even in the acutely locked knee, a condition where early arthroscopy has been the norm, preliminary magnetic resonance imaging can reduce the need for arthroscopy in the acute period by 45% (unpublished personal observation).

Magnetic resonance imaging of the knee may also be considered in patients with persistent pain, especially at night; a mass lesion; an acutely swollen joint; and osteoarthritis. In patients with nocturnal pain, a tumour needs to be considered. Plain radiography remains the mainstay in the diagnosis of bone tumours, followed by magnetic resonance imaging for staging if a lesion is detected.¹² If plain films are negative and symptoms persist, magnetic resonance imaging is a more sensitive investigation. Isotope bone scans are also a sensitive way of detecting tumours but do not provide the same anatomical detail as magnetic resonance imaging and carry a noteworthy load of radiation.

Similarly magnetic resonance imaging is more sensitive than plain films in detecting stress fractures, particularly in the early stages, and should replace isotope bone scans.¹³ In patients suspected of having soft tissue masses ultrasonography is recommended as an initial screening test, followed by magnetic resonance imaging should a mass be found.

Magnetic resonance imaging, however, usually has a limited role in patients in whom plain x rays show evidence of osteoarthritis. The extent of anatomical damage does not correlate with symptoms, which are the primary determinants in the timing of arthroplasty. An exception is when a unicompartmental (that is medial compartment) rather than a total knee replacement is proposed. Here magnetic resonance imaging can confirm that the other compartments are normal.

In inflammatory synovitis, magnetic resonance imaging can confirm the extent of involvement; distinguish between effusion and synovitis, particularly when intravascular enhancement agents are used; and determine synovial bulk where surgical synovectomy is being considered. It is less common for magnetic resonance imaging to yield a specific diagnosis, though on occasion entities such as pigmented villonodular synovitis can be diagnosed owing to their specific characteristics on imaging.¹⁴ Septic arthritis is usually associated with marked inflammatory reaction in the underlying bone, and a synovial biopsy is recommended to confirm it. Subsequent magnetic resonance imaging is needed to exclude osteomyelitis.