CLINICAL MRI OF OSTEOARTHRITIS AND OSTEOCHONDRAL INJURY

As discussed in Chapter 2 , the structural integrity and function of cartilage and underlying bone are intimately coupled. Chondral injury leads to increased stress on subchondral bone, and, conversely, failure of the subchondral osseous plate undermines integrity of overlying cartilage. Although unique imaging findings are observed within cartilage, bone, and the cartilage/bone interface, it is reasonable to consider these as injuries of the osteochondral unit. It can be difficult to visualize directly chondral injury in joints such as the ankle, hip, and shoulder where cartilage is thin and covering a curved surface. Often, altered signal intensity in subchondral bone marrow is the most conspicuous finding to indicate there has been an osteochondral injury. Certain features in combination with the clinical history may provide clues as to the type and duration of injury. Recognizing the pattern of injury can assist in focusing attention to specific areas of the articular surface that are likely to have been exposed to trauma. This information may suggest an underlying mechanism and time course of the injury. The contour of the chondral margins and marrow abnormalities may help differentiate acute traumatic insult from chronic degeneration. Pooling information gathered from evaluation of the entire joint will lead to greater confidence in the diagnosis of osteochondral injury.

CLINICAL MRI OF OSTEOARTHRITIS

Although MRI is a potentially valuable clinical research tool for longitudinal studies of osteoarthritis (OA), it plays a very limited role in the clinical evaluation of patients with established radiographic findings. There are occasions, such as when there is discrepancy between severity of clinical and radiographic findings, that MRI is a valuable clinical problem-solving tool. This most often occurs in patellofemoral and hip OA when radiographic evaluation of joint space narrowing can be unreliable and underestimate the degree of articular damage. MRI may exclude confounding pathologic processes such as meniscal tears that can mimic symptoms of OA and require different therapy. As illustrated in the case presented in Figure 3-1 , there is a role for MRI in OA patients who have acute worsening of pain that may indicate a new pathologic process that may impact patient management, such as an insufficiency fracture of subchondral bone.

Many factors, including biomechanics, genetics, and inflammation, affect the heterogeneous condition of OA. The combination of these factors contributes to patient symptoms of pain, stiffness, and joint dysfunction. An active area of clinical MRI research in the management of patients with OA is identifying sources of pain based on objective imaging findings. In the future, MRI may provide differentiation that could help guide and target treatment. For example, the presence of joint effusion and synovitis may indicate patients for whom the inflammatory component of OA is a primary source of symptoms. Such a patient may benefit from treatment with anti-inflammatory medications, in contrast to patients with focal subchondral marrow changes in whom dysfunctional biomechanics may be a more important factor. Another research goal in MRI is identification of imaging features that may identify patients at risk for rapid disease progression who would benefit from more aggressive intervention or alteration in lifestyle. Preliminary studies have identified subchondral bone marrow degenerative signal and focal osteochondral defects as two risk factors for knee OA progression.

OSTEOCHONDRAL INJURY

There is growing recognition that focal osteochondral injury represents a substantial risk factor for the development of OA ( Fig. 3-2 ). Particularly in the younger patient, MRI plays an important clinical role in the diagnosis of pre-radiographic OA and focal osteochondral injury. Increased awareness of osteochondral injuries, development of cartilage repair techniques, and advances in MRI technology have increased the diagnostic frequency of these lesions. The historical dependence on radiography has emphasized the osseous component of osteochondral injuries. Although radiography continues to be the first-line imaging modality in evaluation of articular injuries, it is insensitive to isolated cartilage injury. MRI has become a standard diagnostic tool in evaluation of osteochondral injuries because it is the only noninvasive imaging technique that provides direct assessment of articular cartilage and associated subchondral marrow pathologic processes.

The thin, curved contour of articular cartilage places high technical demands on MRI. Compared with evaluation of meniscal or ligamentous injury, the accuracy and the sensitivity of MRI for cartilage injury are substantially lower and quite variable and depend on technique and experience. Over recent years, advances in MRI hardware and acquisition techniques have improved the abilities to visualize and quantitatively monitor morphologic changes in small cartilage lesions. In addition to clinical methods focused on changes in cartilage morphology there has been development of physiologic MRI techniques that are sensitive to changes in cartilage composition and organization of the extracellular matrix that precede loss of tissue. These techniques have the potential to provide novel information on the natural history of early cartilage injury, in-vivo cartilage biomechanics, and longitudinal evaluation of cartilage repair.

Osteochondral injuries represent a spectrum of articular conditions ranging from acute cartilage tears to chronic osteochondral defects, including osteochondritis dissecans (OCD). Although there is ongoing debate regarding the pathogenesis of OCD, there is a consensus supporting a traumatic/mechanical theory of injury to the osteochondral unit. The imaging appearance of chondral/osteochondral injury differs based on severity and acuity of the trauma and the reparative response of the tissue. A single episode of high-impact trauma may result in a chondral or osteochondral fracture. In certain locations with appropriate mechanisms of injury, the forces applied to the bone/cartilage unit result in a de-bonding or delamination of the cartilage from the underlying bone. In cases in which acute trauma does not produce a structural defect it may alter the biomechanical properties of the osteochondral unit, leading to progressive loss of articular cartilage and degenerative change in subchondral bone. In the absence of an acute traumatic insult, chronic repetitive microtrauma, such as overuse injuries, can produce focal microfracture, necrosis, and healing response of subchondral bone with localized degenerative changes in the overlying cartilage.

MRI Assessment of Chondral Injury

Current MRI grading systems of focal articular cartilage damage are based on modifications of the Outerbridge classification originally described for surgical grading of patellar lesions. The original surgical Outerbridge classification is based on size of surface fragmentation and fissuring. The MRI modification of the Outerbridge classification incorporates depth of the lesion from the articular surface. In addition, several MRI classifications have been proposed for grading osteochondral lesions.

GRADE 0: NORMAL CARTILAGE

By definition, grade 0 cartilage has both normal morphology and signal intensity; however, signal intensity of normal cartilage varies with depth from the articular surface, location in the joint, age of the patient, and the particular pulse sequence used to acquire the image. The appearance of normal cartilage was discussed in Chapter 2 .

GRADE I LESIONS: ISOLATED CHANGE IN CARTILAGE SIGNAL INTENSITY

In the surgical form of the Outerbridge classification, grade I lesions are identified by a subjective determination of cartilage softening or discoloration with an intact articular surface. Because there is no direct MRI finding that corresponds to cartilage softening, this has been modified to reflect isolated MRI signal changes without disruption of the articular surface. Early studies found poor correlation between grade I MRI lesions and arthroscopy, as well as low sensitivity in MRI detection of patellar cartilage softening found at arthroscopy. Recent studies indicate diagnostic accuracy is improved with 3.0T MRI technology.

With the lack of findings at arthroscopy there is a tendency to discount MRI findings of early cartilage injury. Studies performed on excised cartilage specimens suggest that poor correlation reflects inherent differences in the properties of cartilage evaluated with MRI and arthroscopy. It has been postulated that elevation in cartilage signal intensity on T2-weighted sequences is a result of alterations in the organization of the collagen matrix that reduce collagen fibril anisotropy and increase cartilage water content. In studies of enzymatically treated cartilage specimens, degradation of the type II collagen matrix was strongly correlated with elevation in cartilage T2, whereas removal of proteoglycans using either trypsin or interleukin-2 had minimal effect. Similarly, it has been shown that although removal of proteoglycan significantly decreases cartilage stiffness, degradation of the collagen matrix associated with elevation in cartilage T2 correlated poorly with Young’s modulus. Based on results of ex-vivo studies, it is likely that focal elevation in T2-weighted signal intensity reflects structural changes of the collagen matrix that do not substantially alter the visible appearance or compressibility of cartilage. Although such damage may not have an immediate impact on tissue biomechanics, because of the inability of cartilage to repair or replace type II collagen, it places cartilage at risk for further degeneration. Preliminary studies have shown grade I lesions frequently progress to higher grades. Long-term natural history trials such as the National Institutes of Health–funded Osteoarthritis Initiative (OAI) are needed to determine the clinical significance of these lesions.

An early manifestation of OA is failure of the type II collagen matrix. This may occur as a result of a single traumatic insult in which tensile strain in the collagen network exceeds the biomaterial properties or through repetitive stress that can lead to fatigue fracture. Age is an important factor in modifying the material properties of the collagen matrix. Because of the slow rate of turnover of type II collagen in cartilage there is accumulation of advanced glycation end products and crosslinking of collagen. These age-related factors increase the stiffness and friability of collagen, making it more brittle and prone to fracture.

As originally describe by Maroudas and colleagues, fragmentation of the collagen network may be one of the earliest manifestations of osteochondral injuries leading to OA. As illustrated in Figure 3-3 , fractures in the collagen network prevent constraint on the hydrated aggrecan, allowing the tissue to swell. As the tissue swells there is a measurable increase in water content and loss of the normally high interstitial swelling pressure. This reduces the ability of cartilage to restrict the flow of water through the tissue as it undergoes compression. With loss of restricted water movement, more energy deposited in the tissue from compressive force is transferred to the solid extracellular matrix. This leads to progressive degradation and fragmentation of the extracellular matrix. Concurrently, high magnitude compressive strain in cartilage leads to chondrocyte injury and apoptosis, limiting ability to synthesize new proteoglycan. This results in a mismatch where degradation from excessive biomechanical loads exceeds the biomaterial properties of the tissue. With time, this imbalance results in visible cartilage damage and, ultimately, loss of tissue.

Many of these earliest changes in cartilage injury increase the T2 relaxation time of cartilage water and produce changes in MRI signal intensity that can be visualized with standard T2-weighted, turbo spin-echo MRI techniques. Breaks in the collagen matrix, particularly those occurring in regions of high structural anisotropy such as the radial zone, reduce the residual dipolar coupling with water, leading to prolongation of the cartilage T2 relaxation time. The concurrent elevation in cartilage water content and increased mobility of water are also factors that lead to longer T2 relaxation in early cartilage injury.

Focal areas of T2 hyperintensity are frequently found in patients without a discrete cartilage defect. Although the clinical significance of this finding is unknown, as illustrated in Figure 3-4 , these lesions often progress to sites of morphologic damage. A more diffuse heterogeneous pattern of high T2-weighted signal can be observed after acute trauma, frequently in association with hyperintensity in the adjacent subchondral bone marrow. Isolated areas of T2 hyperintensity may be observed in the cartilage of patients with no reported history of trauma. As demonstrated in Figure 3-4 , this can be associated with a focal blister or smooth contour abnormality of the overlying articular surface. Similar findings of focal swelling and alterations in the fibril density in the superficial zone of patellar cartilage have been reported in the electron microscopy literature, supporting the hypothesis that these lesions represent structural reorganization/degeneration of the collagen matrix.

In addition to T2 hyperintensity, which is frequently present in the acute setting, focal areas of decreased T2-weighted cartilage signal are frequently observed adjacent to sites of cartilage injury. Decreased T2-weighted signal is generally not observed immediately after trauma but, as demonstrated by the case illustrated in Figure 3-5 , is observed with subacute and chronic osteochondral injuries. The etiology of the decreased T2 signal intensity has not been determined but may reflect fragmentation of collagen fibrils leading to a greater number of exposed water-binding sites. In addition to providing more efficient T2 relaxation, collagen fragmentation may decrease overall signal intensity through magnetization transfer between collagen fragments and cartilage water. Alternatively, T2 hypointensity can result from formation of fibrocartilage repair tissue. Areas of low T2-weighted signal intensity are also observed with sites of chondrocalcinosis, particularly with gradient-echo techniques or with high magnetic field strengths.

GRADE II/III LESIONS: PARTIAL-THICKNESS CARTILAGE INJURY

Grade II lesions represent fissures, erosion, ulceration, or fibrillation involving the superficial 50% of cartilage thickness. In the modified Outerbridge classification, lesions that extend to the deep 50% of cartilage are considered grade III. These injuries can occur in the setting of acute trauma or chronic cartilage wear; however, imaging features are different. Acute osteochondral injuries are suggested by the presence of adjacent soft tissue edema, joint effusion, and subchondral bone marrow T2 hyperintensity with ill-defined margins. The presence of well-demarcated or cystic lesions in subchondral bone is suggestive of chronicity. As discussed previously, chronic lesions are frequently heterogeneous, with areas of increased and decreased signal intensity in cartilage on proton density (PD)–weighted or T2-weighted fast spin-echo sequences.

There is no consensus in the MRI literature regarding terminology used to describe morphology of focal cartilage lesions. Fissures represent linear clefts of the articular surface ( Fig. 3-6 ). They are most frequently observed acutely after joint trauma, particularly in patellar cartilage. These clefts frequently follow the preferential orientation of the collagen matrix. The rate and magnitude of loading of cartilage influence the location of injury. When shear force is applied at high speed but with low energy, cracks are produced along the articular cartilage surface. At low speed and low energy, splits initially occur in the deeper layers.

As seen in Figure 3-7 , obliquely oriented fissures or flap tears can be seen as linear defects extending from the articular surface. Such tears may result from excessive shear strain, resulting in failure of the collagen network at the junction of the transitional and radial zones. Because chondral flap tears may produce symptoms of locking, patients can present with clinical symptoms mimicking a meniscal tear.

Ulceration of superficial cartilage blisters results in a small focal irregular crater. Erosion refers to a smoothly marginated area of thinned cartilage and is frequently seen in older patients. Cartilage erosion is often identified in the posterior tibial plateau and femoral condyle, particularly in patients with chronic tears of the anterior cruciate ligament ( Fig. 3-8 ). Fibrillation or fraying of the articular surface appears visually as a fine velvety surface and is a common finding in subjects with OA and in asymptomatic older individuals. MRI has insufficient spatial resolution to resolve the individual fibrillations and generally appears as an indistinct articular margin. Although MRI has poor correlation with arthroscopy for Outerbridge grade I lesions, sensitivity and specificity are greater than 85% for grade II lesions and higher.

GRADE IV LESIONS: FULL-THICKNESS CARTILAGE LESIONS

Full-thickness lesions with exposure of the underlying subchondral bone are classified as grade IV lesions. The margin of the lesion can suggest the mechanism of cartilage injury. As shown in Figure 3-9 , sharply marginated borders are characteristic of traumatic cartilage injuries, whereas shallow or irregular margins are features more characteristic of chronic degeneration. Abnormal signal from the underlying bone marrow and central osteophytes is frequently associated with grade IV lesions but is also observed in lower-grade chondral lesions. MRI has demonstrated high specificity and sensitivity for detection of grade IV defects.

MRI Findings in Injury of the Cartilage/Bone Interface

Excessive shear forces to the cartilage/bone interface can injure the tidemark zone and disrupt collagen fibers that bind cartilage to subchondral bone. In addition to shear force applied directly to the articular surface, high shear strain at the bone/cartilage interface develops with axial compression. Cleavage of collagen fibers leads to delamination or de-bonding of cartilage from the underlying bone. As illustrated in Figure 3-10 , this can result in a displaced chondral fragment. In less severe cases, cartilage delamination may not be readily apparent at arthroscopy, because the articular surface is often intact. In addition to biomechanical factors, recent evidence demonstrates that genetic factors influence the risk of cartilage delamination.