MRI IN RHEUMATOID ARTHRITIS

Magnetic resonance imaging has been applied to the study of rheumatoid arthritis (RA) for more than 15 years, and a large body of clinical and research experience on its use has contributed to a growing understanding of the important role of this modality, both in helping to develop our basic understanding of the disease and in the ongoing clinical management of patients with RA and juvenile RA (JRA).

There are two principal areas in which MRI has been shown to have particular utility and for which there is considerable clinical experience: (1) in the detection and measurement (quantification) of erosions and pre-erosive changes, especially in early disease, and (2) in the evaluation (quantitative and semi-quantitative) of synovitis, primarily by using intravenous gadolinium-based contrast agents. As with Doppler ultrasonography, MRI can evaluate an individual patient’s response to treatment and help to guide day-to-day therapeutic decision making. Current research also suggests a developing role for MRI in depicting the basic cellular and molecular processes in RA, which may improve our basic understanding of the disease itself, as well as provide the highest degree of accuracy in evaluating the destructive activity of the disease and response to treatment that is possible with currently available methods. For each of these roles MRI overlaps with other modalities, including ultrasonography and molecular imaging.

EROSIONS IN RHEUMATOID ARTHRITIS: COMPARISON OF MRI VS. ULTRASONOGRAPHY VS. RADIOGRAPHY

Perhaps the most significant role for MRI in rheumatoid disease is in the early detection, measurement, and characterization of joint erosions and pre-erosive changes. The characteristic erosions of rheumatoid arthritis occur at the margins of joints, at synovial reflections, and at sites of insertion of interosseous ligaments. On MRI, these are best seen on coronal spin-echo images. As with ultrasonography, MRI is a cross-sectional imaging modality, and one with very high intrinsic tissue contrast. Despite having lower spatial resolution than plain radiographs, MRI has shown much greater sensitivity in the detection of joint destruction at an early stage and, in repeated studies, has been proven to detect and identify erosions, even very small ones, with greater sensitivity and efficiency than plain film radiography or, for that matter, any other available imaging modality. In terms of comparative sensitivity, evaluation with ultrasound has been shown to detect more erosions than are visible on plain radiographs, and MRI detects still more than ultrasound in the same patient populations. As such, MRI may be valuable even in patients who have normal findings on radiography. In a study involving 46 patients with early RA, Hoving and coworkers showed that MRI is significantly more sensitive for the detection of erosions than either ultrasonography or radiography and has high inter-reader reliability. More than twice the number of erosions were detected on MRI than by either ultrasonography or radiography in their series. In early RA, MRI represents an especially significant advance in the detection of erosive changes as compared with the previous gold standard, plain film radiography, even when optimized radiographic methods are employed.

In common clinical practice, most rheumatologists and radiologists over the past approximately 40 years have relied on conventional film-screen radiography using standard equipment and film for the detection and scoring of erosions. However, in many research studies and in some clinical practices, specialized high-resolution x-ray methods were also employed, such as fine-grain, single-emulsion screenless radiography and magnification radiography using small focal spot x-ray sources of 6 to 8 µm, typically using 5× geometric magnification. These uncommon radiographic methods served to optimize the radiographic technique, allowing earlier detection of smaller erosions and greater sensitivity overall. Presumably, this also led to more accurate scoring using established scoring systems, such as the Sharp and Larsen scoring systems, which involve counting as well as measuring rheumatoid erosions. As such, these optimized planar radiography techniques probably represent a fairer “gold standard” with which to compare the more advanced cross-sectional methods, such as MRI, despite their limited use in practice. It is clear, however, that MRI surpasses the ability of even the most optimal radiographic methods for the detection and characterization of erosions in rheumatoid disease—in fact, it has been firmly established that MRI is the single best modality for the detection and evaluation of erosions in RA, and, further, its use has revealed the existence of pre-erosive antecedent bony changes that are not apparent radiographically. Finally, MRI-based observations have also given substantial support to the prevailing hypothesis that erosive changes in RA are the result of rheumatoid synovitis, rather than there being an unrelated or separate disease process or epiphenomenon.

MRI VS. ULTRASONOGRAPHY IN CLINICAL PRACTICE

In comparison with ultrasonography, MRI takes significantly longer to perform, is generally more costly, and as a direct result of time and cost constraints, typically only one site (e.g., the dominant or most symptomatic hand and wrist) can be evaluated at a single session, whereas multiple sites can be readily imaged using ultrasound. With MRI, there is superior tissue contrast resolution and a more favorable signal-to-noise ratio (SNR). Images are visually pleasing and are generally limited to the conventional three anatomic planes. As such, they are more straightforward to interpret by the clinician and are easier to use as adjunctive “visual aids” in discussion with patients. Indeed, many rheumatologists prefer MRI over ultrasound images in practice, because MRI is the preferred modality in the detection of erosions, and there is a high clinical comfort level with the tracking of erosions as an indicator of disease progression in RA, and because the images are visually analogous to radiographs in that they depict anatomic structures in familiar ways. As an added advantage of MRI over ultrasonography, the comprehensive regional anatomic overview uniquely provided by MRI can often detect and characterize additional lesions that may have been unexpected or even unrelated to the patient’s arthritis in the course of the evaluation.

In addition to these clinical factors driving the utilization of MRI for evaluation of patients with RA, there are also economic inducements that may have a significant impact as well. In the United States and elsewhere, ownership of an MRI system by a physician group may provide significant income to the practice, because third-party payer reimbursement for the technical component of MRI studies can be significant and is much higher than that for radiography or ultrasonography. This has created a significant emerging market for imaging equipment manufacturers, who have responded by producing relatively inexpensive imaging systems with reduced fields of view, tailored for the imaging of small joints. These so-called “niche” magnets, which offer relatively low acquisition costs, high-quality imaging output, ease of use, and low space requirements are ideally suited for an office-based rheumatology practice when there is a large enough patient base to support their use. Such units are heavily marketed to academic and private rheumatology practices, with both their clinical and financial benefits to the practice emphasized by the manufacturer’s representatives. It is undeniable that the combination of aggressive marketing of systems by the manufacturers and the substantial financial incentives of the health care marketplace contribute to a greater acceptance of MRI in the practice of rheumatology in the United States relative to ultrasonography.

As was discussed in the previous chapter in some detail, ultrasound has great utility in the evaluation of synovitis in RA ; however, MRI with gadopentetate dimeglumine (Gd-DTPA) contrast administration is also capable of this assessment. Indeed, there have been many published studies utilizing MRI primarily for the evaluation of synovitis in RA, although most have been studies of relatively few patients and there has been little comparison in the literature between high- and low-field MRI scanners or in comparative assessment of other technical factors. For the evaluation of synovitis by MRI in clinical practice, it is common to use conventional spin-echo T1-weighted and T2-weighted sequences, usually with spectral fat suppression and Gd-DTPA intravenous contrast, supplemented with the occasional use of short tau inversion recovery (STIR) sequences. Imaging sequences are obtained in coronal and axial planes and typically utilizing a high-resolution extremity surface coil. Because equipment manufacturers are constantly upgrading and improving available MRI systems and technologies, it is likely that future MRI systems will offer additional imaging options and ultimately may significantly outperform results published in the current literature.

RHEUMATOID EROSIONS AND PRE-EROSIVE CHANGES ON MRI

True RA erosions, that is , those involving frank cortical destruction and detected by T1-weighted, spin-echo MRI, can often also be demonstrated on radiographs. Gd-DTPA contrast enhancement may be a common feature of erosions in RA, seen in up to 72% of erosions to some degree. In addition, the use of MRI has also revealed the existence of reversible, “pre-erosive” lesions that are, in fact, seen only on MRI (i.e., erosion-like lesions that do not correlate with radiographic erosions and that are not visible with ultrasound or on CT) and that may completely vanish in response to effective chemotherapy. Ostergaard and coworkers have suggested that these early bone marrow signal changes seen on fluid-sensitive MRI sequences at the joint margins of patients with RA appear to represent a type of “pre-erosive” osteitis—an inflammatory rather than an erosive process occurring at the typical location of rheumatoid erosions—reflecting the presence of inflammatory or edematous changes in the bone marrow but with the overlying bony cortex still intact. These pre-erosive lesions are indeed distinguishable from true erosions on MR images by their intact cortical margins and internal signal characteristics. On MRI, true erosions show cortical interruption and have well-defined, rounded margins and contain only synovial fluid or synovial tissue signal. In contradistinction, these pre-erosive lesions have free water signal within the medullary space and tend to have somewhat ill-defined margins. Furthermore, unlike true erosions, they may also appear to contain interspersed trabecular bone or even fat.

Of particular note, these nonerosive inflammatory changes have been observed to be at least potentially reversible and heal in some cases, whereas in other cases they progress into true rheumatoid erosions that can be seen on plain films. In essence, then, MRI has demonstrated both the natural history of developing erosions in RA and also the repair or healing of what amounts to early erosion-precursor lesions previously not known to exist. The preponderance of evidence seems to suggest that when a lesion is seen only as a focus of relatively high signal on fluid-sensitive sequences, such as T2-weighted imaging, and is not evident on T1-weighted, spin-echo sequences, such a lesion generally may be thought of as a reversible osteitis or “pre-erosive change” that will not correlate with erosions on the radiograph; to the contrary, low-signal defects on T1-weighted, spin-echo images are generally seen with irreversible bone damage and correlate more completely with radiographically demonstrated rheumatoid erosions. Most importantly, it has been shown that the former often evolves into the latter with time.

As noted earlier, the pattern of MRI changes observed across numerous experimental studies using differing techniques supports the view that synovitis is the antecedent change in RA, leading to osteitis and, in most if not all cases, ultimately producing bony erosions. Based on MRI data, some authorities have questioned whether the two (synovitis and erosive disease) occur out of proportion to each other, which would suggest, at the very least, that the linkage between synovitis and bone erosions may be a complex or indirect one. Nevertheless, it is clear that neither the osteitis seen on MRI nor the bony erosions seen by plain film, CT, MRI, or ultrasonography appear to occur in the absence of synovitis. Most experts are convinced of a temporal sequence or natural progression of changes as depicted on MRI, progressing from normal (no erosions or synovitis), to synovitis or tenosynovitis, demonstrable by MRI or ultrasound, to osteitis with signal changes on T2-weighted imaging, to finally frank erosions that appear dark on T1-weighted imaging and are visible on plain radiographs and CT and ultrasound images ( Fig. 6-1 ).

In adult RA, erosions occur as an early manifestation in the course of disease—usually within the first 6 months of symptom onset. As such, there may be a significant clinical advantage in their early detection. Because of its high sensitivity and predictive value, McQueen and associates have advocated the use of MRI in the patient’s dominant wrist in newly diagnosed RA to identify those patients who manifest early erosive and pre-erosive changes, because these patients would more likely benefit from early, aggressive chemotherapy than those who lacked evidence of early erosive (and pre-erosive) change.

MRI may be particularly useful in evaluating the hand and wrist, where the earliest and most frequent erosions are found.

MRI OF SYNOVITIS IN RHEUMATOID ARTHRITIS

The inflamed synovium in RA, as was noted in the previous chapter, is a highly vascularized, complex tissue that is richly cellular and whose overall volume correlates well with clinical disease severity and poor clinical outcome. On conventional spin-echo images, the thickened, inflamed synovium (pannus) in RA is of intermediate signal intensity on T1 weighting, appears bright on T2 weighting, and enhances robustly after intravenous administration of gadolinium-based contrast material. The intense signal change evident within the inflamed synovial tissues in RA patients is not seen to any significant degree in normal volunteers, although occasional minimal synovial enhancement has been reported in normal subjects. The phenomenon of contrast enhancement seen in RA synovitis occurs very quickly on intravenous administration of a contrast agent and thus reflects an unusually rapid increase in the concentration of gadolinium that occurs within the abnormal (but not within normal) synovium. This, in turn, reflects the abnormally increased synovial blood flow, vascularization, and capillary permeability present in active RA synovitis, features that are reversible with effective treatment. In the presence of active disease, the synovium become suffused with Gd-DTPA, which results in substantial T1 shortening within the inflamed synovial tissues—visualized as a bright region of enhancement on T1-weighted images. Indeed, bright and immediate gadolinium-induced T1 signal enhancement is seen fairly homogeneously throughout the inflamed synovium in patients with RA and JRA.

Because MRI is a cross-sectional “volumetric” method, it is possible to derive a quantitative estimate of the volume of the inflamed synovial tissue that is so well delineated by gadolinium. Measurements of this type must be done using fairly rapid imaging, however, because the contrast material diffuses rapidly from the synovium into the surrounding joint fluid, rendering it bright on T1-weighted images (sometimes referred to as an “arthrogram effect” because it mimics the injection of contrast agent directly into the fluid) and thus rendering the margins of the inflamed synovium itself relatively inconspicuous on the images. Diffusion of Gd-DTPA contrast into joint fluid occurs in patients without synovitis, albeit more slowly, and is the basis of the “indirect arthrogram” technique sometimes used to evaluate shoulder and knee tendons and other internal derangements of joints with MRI. The assessment of rheumatoid synovitis by Gd-DTPA–enhanced T1-weighted MRI, therefore, requires a rapid imaging technique that affords dynamic assessments of the rate and degree of signal enhancement in the first few minutes after intravenous gadolinium contrast administration ( Fig. 6-2 ). A rapid-bolus intravenous infusion of Gd-DTPA contrast, combined with a rapid imaging sequence protocol such as FLASH (fast low-angle shot), allows the accurate measurement of signal change as a function of time, generally referred to as time-enhancement or “time-intensity” curves. The slope of the time-intensity enhancement curve has been shown to be quantitatively predictive, at least in its early rising phase, of the severity of the underlying synovitis ( Figs. 6-3 and 6-4 ).