8.3 Magnetic resonance imaging of the knee

Magnetic resonance imaging (MRI) has revolutionized the investigation and treatment of the painful knee. It is non-invasive and avoids patient exposure to ionizing radiation. MRI has the advantage of establishing diagnoses in a painful knee without the morbidity of surgical intervention. It is now widely available and has moved from a simple diagnostic adjunct into a key planning tool. It offers improved management of theatre resources and it allows for more accurate planning of postoperative rehabilitation.

The role of MRI in management of the injured knee is determined by its cost-effectiveness and its ability to augment the diagnostic accuracy of clinical examination. Accuracy of clinical examination by specialist orthopaedic surgeons is comparable to MRI when interpreted by specialist radiologists (Table 8.3.1). Increasingly, MRI has been shown to be cost neutral. Whilst costs are high, diagnostic information reduces the need for unnecessary surgery.

MRI protocols will differ slightly between units (Box 8.3.1). MRI T₁-weighted sequences are used to define local anatomy and can be used to identify trabecular microfracture. T₂-weighted sequences (with fat suppression) are most commonly employed to demonstrate oedema of both soft tissues and bone marrow. Proton density and T₂ gradient echo sequences are utilized to highlight meniscal pathology. Viewing images on a workstation is now preferred to hardcopy films with enhancement of brightness, contrast, and magnification with linking of orthogonal views. Comparison with plain radiographs complements interpretation of calcific lesions and assists identification of small avulsion injuries.

Bone bruises are occult bony injuries and represent the local sequelae of chondral and microtrabecular trauma (Figure 8.3.1). Patterns of bone bruising can be pathognomonic for specific ligamentous injuries. Bone bruises have been described as reticular (oedema not extending to the subarticular cortical bone), or geographic (oedema which is continuous with the subarticular cortical bone). The long-term significance of these lesions has yet to be determined. However, histopathological studies of geographic bone bruises have demonstrated local chondrocyte and osteocyte death and have suggested that severe bone bruising may represent a precursor of early degenerative change. This could explain why degenerative changes develop despite successful anterior cruciate ligament (ACL) reconstruction and may provide justification for less aggressive rehabilitation programmes in patients with large subcortical lesions.

MRI is useful in the diagnosis of ACL rupture (Figure 8.3.1) and associated meniscal or chondral lesions. ACL rupture can be inferred from absence, discontinuity, or oedematous replacement of the normal ligamentous band that parallels Blumensaat’s line. Less accurate indirect radiological signs such as buckling of either the posterior cruciate ligament (PCL) or patella tendons are evidence of anterior tibial translation (usually more than 5–7mm referenced from posterior margins of the tibia and femoral condyles). Difficulty arises from evaluation of partial tears, which should be correlated with clinical findings to determine appropriate management. Transient subluxation and impaction of the femur into the tibia at time of injury can lead to a ‘kissing contusion’ seen typically as bony oedema in the midportion of the lateral femoral condyle and posterior lip of the lateral tibial plateau. This ‘traumatic pivot shift’ is also responsible for the associated chondral, capsuloligamentous, and meniscal injuries.

Clinical assessment of meniscal injuries (Table 8.3.1) is difficult and a combination of history and positive provocation tests are required for accurate diagnosis. In a systematic review of 32 papers, Ryzewicz and colleagues found that routine MRI was not indicated, as clinical assessment in experienced hands was of equal or higher reliability. However, MRI is able to assess the type of meniscal injury, aids surgical planning, and assists diagnosis with concomitant ligamentous injury where clinical examination becomes less reliable.

Detection of meniscal pathology involves identification of abnormal meniscal morphology and signal. Care should be taken as many asymptomatic individuals will have areas of high intrameniscal signal which is generally considered to be a normal consequence of aging. Abnormal signal, extending to the free edge of the meniscus, is accepted as representing a tear. However, a degenerative tear may not always represent the problematic pathology, and MRI review should not cease once a tear is identified. Meniscal fragments can become displaced into the intercondylar notch or settle into synovial recesses. It should also be noted that aberrant meniscal anatomy may mask injury. Remnants of a torn discoid meniscus can appear misleadingly normal.

Posterolateral corner (PLC) injuries pose a difficult clinical problem. They often occur in conjunction with other ligamentous injuries and physical examination may be impossible in the acute setting. MRI allows evaluation of the PLC and can influence surgical management. Injuries to the PLC structures can be assessed with a variety of coronal, sagittal, and oblique views. Oblique coronal slices through the fibular head can more accurately demonstrate the popliteofibular ligament and inclusion of the fibular head in all sequences assists in diagnosis. Soft tissue oedema in the posterolateral aspect of the knee should always be viewed with suspicion. Osseous injuries such as occult arcuate fracture, Segond fracture, medial femoral condyle bone bruising, anterior rim tibial plateau fracture, and avulsion of Gerdy’s tubercle should all be assessed. Each component should be reviewed and evidence of thickening, oedema, tear, disruption, or avulsion should be sought. However, the anatomical definition and individual variation of the PLC mean that defining all structures can be difficult. The popliteus, popliteofibular ligament, arcuate ligament, and lateral collateral ligament are primary structures considered for repair and thus their evaluation will provide the most useful information for surgical decision making.

Postoperative MRI is more difficult to interpret, although it is useful for the assessment of graft integrity, tunnel position, and tunnel widening following cruciate reconstruction, monitoring osteochondral defects or for missed meniscal pathology following arthroscopy. MR arthography using gadolinium has been employed to assess the postoperative meniscus and reconstructive grafts. Although this technique is invasive, it provides joint distension which facilitates imbibition of fluid into latent tears, potentially improving accuracy. However, at this time there are no convincing results to justify its routine use. Artefacts from modern metallic implants can now be minimized by various MRI techniques and do not preclude evaluation.

MRI should be considered in the evaluation of other lesions in the knee joint. It can be used to provide prognostic data to patients with spontaneous osteonecrosis of the knee (SONK), examine the extent of spread of bony and soft tissue tumours (e.g. pigmented villonodular synovitis, PVNS), and evaluate postoperative cartilage repair after procedures such as chondrocyte implantation or marrow stimulating techniques.

Dramatic improvements in the quality of images produced by MRI over the past decade have provided clinicians with invaluable data that has undoubtedly saved patients from unnecessary surgery. It has also improved our understanding of pathology and the postoperative healing processes. As new technology improves resolution, and experimental techniques such as real-time MRI become more accessible, our understanding of individual patients’ pathokinomatics will improve. The increasing quality of MRI will continue to augment clinical orthopaedic practice, and should ultimately result in improved outcomes for patients.