Postoperative Articular Cartilage and Subchondral Bone Imaging

Issue: October 2016

Magnetic resonance imaging plays an important role in the postoperative evaluation of cartilage repair procedures in the knee. Magnetic resonance imaging is the most reliable noninvasive method to assess cartilage repair tissue and underlying subchondral bone.

Although the gold standard of evaluation is direct arthroscopic visualization and probing or histological examination of repair sites, MR affords an excellent noninvasive assessment of the repair site. It has the added advantage of visualizing the subchondral bone that provides valuable information when planning the treatment of symptomatic patients, especially in a revision surgery setting. This imaging also has the ability to demonstrate new joint pathology, such as new cartilage defects elsewhere in the knee or meniscoligamentous pathology away from the repair site.

Figure 1. Normal maturation of repair tissue following microfracture on coronal MR imaging of medial femoral condyle is shown. At 3 weeks (A), the repair site is completely filled with a congruent surface contour. The immature repair tissue has fluid-like signal intensity (bracket). At 3 years (B), the repair site remains completely filled. The mature repair tissue now appears darker than native cartilage (bracket).

To a degree, MR has reduced the need for second-look arthroscopy in clinical trials. The development of compositional MR techniques allows for evaluation of the microstructural and biochemical composition of the repair tissue and may serve as a surrogate for histological examination. CT arthrography is a complementary imaging test that provides excellent information about bone graft and subchondral bone. Although CT has relatively poor soft tissue contrast and cannot demonstrate bone marrow edema, CT arthrography can be a helpful alternative for patients who cannot undergo an MRI examination.

Arthroscopic scoring systems to quantify the quality of cartilage repair tissue have been developed and have served as models for MR scoring systems. Two commonly used arthroscopic systems are the International Cartilage Repair Society (ICRS) macroscopic evaluation of cartilage repair and the Oswestry Arthroscopy Scoring (OAS) system for cartilage repair. Currently, MR cannot demonstrate tissue stiffness, color and fine surface irregularities such as fibrillations, but it can assess subchondral bony changes and deep cartilage abnormalities (such as delamination) that may not be evident on visual examination.

MR scoring systems have been developed that aim to quantify the status of cartilage repair tissue. These include the Henderson scoring system, the Roberts scoring system, Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) and Osteochondral Allograft Magnetic Resonance Imaging Scoring System (OCAMRISS). These are primarily useful for clinical trials when quantification of findings is desired and serve as a guideline for clinical assessments. The common features defined in these scoring systems that coincide with the important arthroscopic features include defect fill, repair tissue character, repair tissue integration, subchondral bone status and nonrepair site features. It is important to note that MRI findings following cartilage repair may not correlate with symptoms but may have prognostic implications (eg, asymptomatic delamination). During the first 6 months after surgery, repair tissue may have a highly variable MR appearance dependent on the type of procedure and the status of maturing tissue. A study evaluating MRI correlation with clinical outcomes showed the best correlations between postoperative months 6 and 36.

Although many different cartilage repair techniques are performed, the postoperative MR evaluation of the repair sites is generally similar; some special considerations apply following certain procedures. In the following sections, the features of postoperative MR assessment common to all methods of repair will be discussed followed by specific emphasis on the findings observed after the most common cartilage repair procedures, including marrow-stimulation techniques (eg, microfracture, subchondral drilling, abrasion), osteochondral allograft and autograft transplantation and autologous chondrocyte implantation (ACI).

PAGE BREAK

Defect Fill

The degree of filling of the cartilage defect is perhaps the most important criterion for treatment success and has been correlated with symptomatic success. Both volume of defect fill and degree of surface contour restoration (ie, thickness of the repair) should be assessed. Because all repair procedures begin with complete removal of damaged tissue from the repair site, all tissue observed in a repair site represents repair tissue. While quantifying defect fill as greater or less than 50% is a quick metric, fill is often irregular and regions of exposed subchondral bone (empty defect) or overgrowth (graft “hypertrophy”) may cause clinically significant symptoms. The ideal repair site surface contour is smooth and congruent with adjacent native cartilage.

Figure 2. Incongruent surface following multiple osteochondral autograft transfer (mosaicplasty) is shown. Three months postoperatively, the transplanted osteochondral cylinders (asterisks) do not restore the articular surface. This may have resulted from initial placement or postoperative plug subsidence. An area of fibrocartilaginous overgrowth is present.

Following marrow stimulation procedures such as microfracture, repair sites may initially appear filled, especially when augmented by a scaffold, usually by bright signal tissue. However, more commonly, complete fill may occur more slowly, taking up to 2 years to achieve the final degree of fill (Figure 1). Despite this, many recent publications have demonstrated significant improvements in clinical outcomes for the first 2 years following surgery. Although the evidence is currently inconclusive beyond 2 years, several studies have illustrated a gradual functional decline with higher failure rates, and lower rates of return to sport.

After osteochondral allograft or autograft procedures, defect fill should be complete immediately because a pristine donor segment is utilized. However, the articular surface contour of the repair site should be carefully assessed to determine if the graft is flush, raised (“proud”) or recessed (“depressed”) with respect to the adjacent native cartilage. When depressed, subsequent fill by fibrocartilage may occur with time (Figure 2). It is important to note that the subchondral bone plate of osteochondral grafts is often not at the same level as adjacent bone because the graft usually does not have the same cartilage thickness as the host site (different proportions within the osteochondral unit).

In the early postoperative setting following first-generation ACI (periosteal patch [pACI]) and second-generation ACI (collagen-cover [cACI]), the defect fill is usually complete or slightly overfilled. For third-generation ACI (matrix-associated [MACI]), the defect may initially appear underfilled with maturation to the final level of fill with time (up to 2 years) (Figure 3). Patients with defect fill of less than 50% are more often symptomatic. Hypertrophy has been frequently reported with pACI and much less commonly with cACI and MACI. Hypertrophy greater than 150% of adjacent cartilage thickness was more commonly associated with symptoms than less severe hypertrophy (Figure 4). Mild hypertrophy, less than 125%, may regress spontaneously.

Repair Tissue Character

Repair tissue character refers to the signal intensity and homogeneity of the cartilage repair tissue. The normal trilaminar appearance of hyaline articular cartilage is a reflection of the microscopic orientation of type II collagen. Osteochondral allograft and autograft procedures transfer intact mature articular cartilage to the repair site and thus appear similar to normal cartilage on MRI postoperatively. For the other types of cartilage repair, including marrow stimulation, ACI, acellular scaffolds, and particulated cartilage transplantation repair tissue that grows within the site does not have a normal collagen structure and typically appears different from normal articular cartilage, even if it has a hyaline-like character histologically. Compositional MR techniques are able to quantify glycosaminoglycan content and/or discern gross collagen orientation patterns.

Figure 3. Variation in the normal appearance following MACI is shown. On coronal MRI images of patient 1 (A, B), the MACI repair site on the medial femoral condyle appears underfilled in the initial postoperative period with a high-signal immature interface between the repair tissue and native cartilage (A, arrow). By 12 months, the fill and cartilage integration have improved with the repair tissue appearing similar to adjacent cartilage (B, arrow). For patient 2 (C, D), 3 months after MACI on the medial femoral condyle, there are bright lines at both the repair-bone and repair-cartilage interfaces; the appearance may be confused with delamination in situ (C, arrow). At 12 months, the repair site appears completely filled with well-integrated tissue (D, arrow).

PAGE BREAK

Following marrow stimulation, repair tissue is initially high in signal intensity, sometimes even close to fluid signal intensity (see Figure 1). With time, the repair tissue (primarily fibrocartilage, with a greater concentration of type I collagen than normal hyaline articular cartilage) becomes more similar to normal cartilage in signal intensity, but some heterogeneity commonly persists. As marrow stimulation repair tissue fails, the signal intensity may increase with development of fraying and fissures.

Osteochondral allografts and autografts are expected to have a normal hyaline cartilage appearance that does not evolve with time barring degradation of the tissue. There may be subtle differences in the signal intensity of the grafts with respect to adjacent native cartilage, but this may be because the graft tissue was harvested from a different location and may therefore have a difference in collagen orientation.

Figure 4. Hypertrophy of repair tissue following pACI is shown. Six months after an ACI of the medial femoral condyle using a periosteal patch, the repair tissue in the anterior portion of the repair protrudes above the expected articular contour by more then twice the normal cartilage thickness (ie, greater than 200% [arrows; arrowheads = ACI margins]). As with this patient, hypertrophy of more than 150% of normal cartilage is often symptomatic requiring surgical intervention.

Repair tissue formed following ACI and MACI progresses in a predictable manner as the repair matures (up to 2 years). Initially, the repair appears brighter, sometimes nearly fluid signal intensity. As the cells expand and form extracellular matrix, the signal intensity becomes more similar to articular cartilage, though the tissue often remains slightly increased in signal intensity and heterogeneous when compared to native cartilage (see Figure 3).

Integration

Integration refers to the incorporation of the repair tissue into adjacent native cartilage and underlying subchondral bone. Ideally, these interfaces should be imperceptible on MR. On early postoperative images, the interface may be abnormal with a bright signal line representing normal immature integration (see Figures 1 and Figure 3). A new or persistent fluid intensity slit or measurable gap between repair tissue and native cartilage is a good indicator of incomplete or failure of repair tissue integration. The repair-cartilage interface should be evaluated throughout the entire circumference of the repair site. Focally failed integration may be symptomatic. Focal cyst formation and new or persistent edema-like marrow signal beneath an interface may indicate poor integration. It should be noted, however, that MRI is not perfect at demonstrating repair-cartilage integration when compared to microscopic evaluation. An indiscernible or low-signal linear interface may appear following osteochondral allografts and autografts even though no microscopic integration has occurred.

In general, successful integration between repair tissue and subchondral bone demonstrates an interface with signal equivalent or darker than the repair tissue on MRI. However, on early postoperative images, techniques utilizing fixation of a scaffold are expected to have a thin line of brighter signal since the integration is immature (see Figure 3). Failure of integration at the repair-bone interface appears as a line of new or persistent brighter, often fluid-like, signal intensity at the interface similar to that seen with delamination of native cartilage. Delaminated tissue may remain in situ appearing similar to a cartilage flap, or become displaced creating a full-thickness articular defect. Often, delamination of the repair tissue from bone involves only a portion of the repair site.

Microfracture repair tissue typically integrates with adjacent tissue over the course of 2 years. Edema-like marrow signal is common at the repair-bone interface initially, but typically resolves. Osteochondral autografts and allografts are unique in that the transferred mature cartilage is naturally integrated to bone; mature articular cartilage is not felt to integrate with adjacent native cartilage. Therefore, fluid gaps can be seen between osteochondral graft cartilage and native cartilage. However, intermediate signal is commonly seen at this interface either from the fibrocartilage “grout” filling the gaps or because the interface is thin and below the resolution of MRI. For ACI and MACI, fluid signal intensity is commonly seen at the repair-cartilage and repair cartilage-bone interface in the early postoperative period. This typically resolves with time as the tissue interfaces mature.

References:
Alparslan L, et al. Semin Ultrasound CT MR. 2001; 22(4):341-351.
Blackman AJ, et al. Am J Sports Med. 2013;doi;10.1177/0363546513485931.
Chang EY, et al. Cartilage. 2014;5(1):16-27.
Henderson IJ, et al. J Bone Joint Surg Br. 2003;85(7):1060-1066.
Marlovits S, et al. Eur J Radiol. 2004;52(3):310-319.
Mithoefer K, et al. J Bone Joint Surg Am. 2005;87(9):1911-1920.
van den Borne MP, et al. Osteoarthritis Cartilage. 2007;15(12):1397-1402.