There exists a preclinical phase to the disease progression of rheumatoid arthritis, in which there is evidence of autoimmunity but no overt clinical arthritis. Identifying patients in this phase would allow for early treatment, to potentially halt manifestation of the disease. Imaging, because it is noninvasive, provides an appealing alternative to gold-standard synovial biopsies for identification of these preclinical patients. Ultrasonography, magnetic resonance imaging, and positron emission tomography all have their advantages and disadvantages as imaging modalities in this regard. Further research into alternative imaging modalities with larger cohorts is required to determine the most effective technique.
Imaging can identify evidence of synovitis in patients with autoimmunity without clinical arthritis.
Imaging could be used to risk-stratify patients most at risk of progressing to rheumatoid arthritis.
More work needs to be done in evaluating and comparing different imaging modalities.
Rheumatoid arthritis (RA) is an autoimmune disease affecting up to 1% of the developed world’s population. Because early treatment can limit joint erosion and impede disease progression, it has been important to detect clinical RA at its earliest presentation and to treat with disease-modifying therapy (DMARDs) or biological therapy early, as it has been convincingly shown that delaying therapy in RA patients leads to worse outcomes.
This aspect has led investigators to hypothesize that there may be an even earlier point in RA pathogenesis, when there is evidence of autoimmunity but not yet any evidence of clinical arthritis, and when treatment may be able to abort overt manifestations of the disease altogether ( Fig. 1 ). Imaging, therefore, would play a crucial role in identifying preclinical RA patients from those with noninflammatory arthralgias and myalgias.
Does the Preclinical Phase of Rheumatoid Arthritis Exist?
There is clear evidence that preclinical RA exists. Kraan and colleagues have studied the synovial tissue from patients with RA and have found that synovial tissue from clinically uninvolved knee joints shows infiltration with macrophages and expresses macrophage-derived cytokines. In addition, the uninvolved knee had significantly higher CD68 + macrophages when compared with tissue biopsies from control patients without RA ( P <.005). Animal studies have similarly demonstrated that pathologic inflammation can exist in joints before the onset of clinical arthritis. Rhesus monkeys were immunized with type II collagen from bovine hyaline cartilage to model the development of arthritis. Serial synovial arthroscopic biopsies were then performed on their knee joints at 3 time points: before immunization with collagen, 2 weeks after immunization but before the onset of clinical disease, and after the development of clinical arthritis. An influx of macrophages was observed in all knees during the second time course, before the occurrence of arthritis. These findings confirm that there is objective synovial abnormality present in joints before the development of clinically evident arthritis.
Is Imaging an Effective Noninvasive Method of Identifying Preclinical Inflammatory Arthritis?
The question then becomes: is it possible to accurately identify patients with preclinical inflammatory arthritis (IA)? The gold standard would be a synovial biopsy, performed at the first sign of arthralgia. Although this may be feasible under research protocols, it is neither practical nor ethical to routinely perform such biopsies on patients as part of routine clinical care. Imaging, however, is noninvasive and could avoid the risks associated with biopsies, such as infection, bleeding, nerve and tissue damage, or prolonged pain. Because it would be unreasonable to perform gold-standard synovial biopsies on every arthralgia patient, it will be important to evaluate and validate different imaging techniques to discern which, if any, can accurately identify patients with autoimmunity that have imaging evidence of synovitis but no overt clinical synovitis. It will also be important to compare modalities and determine which are most sensitive. Once validated, investigators will be able to ascertain the additive benefit of imaging above and beyond a history and clinical examination in predicting which preclinical patients progress to IA. Determining whether imaging can effectively risk-stratify patients with preclinical inflammatory arthritis would have important implications for both clinical trials and routine clinical care.
Ultrasonography and power Doppler
Van de Stadt and colleagues assessed the utility of ultrasonography (US) and power Doppler (PD) in the setting of preclinical RA. In a cohort of 192 arthralgia patients without clinical arthritis but with a positive anti-citrullinated peptide antibody (ACPA) and/or immunoglobulin M rheumatoid factor (IgM-RF), joints that were tender on examination or painful by history were scanned in 2 planes with both US and PD. Two blinded radiologists scored synovitis, joint effusions, and tenosynovitis on a semiquantitative scale. These patients were then followed for a median of 11 months. Joint effusions, synovitis, and positive PD findings were seen in up to 13% of patients, and 23% of the cohort developed arthritis in 1 or more joints, most commonly in the wrist, metacarpophalangeal (MCP), proximal interphalangeal (PIP), or metatarsophalangeal (MTP) joints. The presence of pain in a joint at the baseline physical examination was associated with arthritis development in that joint, with an odds ratio (OR) of 3.2 (95% confidence interval [CI] 1.8–5.9; P <.001). The positive predictive value (PPV) of pain predicting arthritis was 11%, and the negative predictive value (NPV) was 96%. However, when positive findings of synovitis on US and PD were both present, the PPV increased to 35%. This study suggests that in fact when US/PD are used together they are better than clinical examination findings alone at predicting the progression to arthritis in a joint. Although this is a promising result, these findings were only found to have statistical significance at the joint level, not at the patient level. There was also only moderate interobserver reliability (κ = 0.46, 0.56, 0.23 for synovitis, PD, and joint effusions, respectively).
The same group went on to evaluate 374 IgM-RF–positive and/or ACPA-positive arthralgia patients. After a median follow-up of 12 months, 131 had developed clinical arthritis. Using a multivariate Cox regression, the group developed a prediction rule using the following 9 baseline variables to stratify patients into low-, intermediate-, and high-risk tiers. Increased risk of developing RA was associated with having a first-degree relative with RA, abstaining from alcohol, symptom duration of less than 12 months, intermittent symptoms, symptoms in both upper and lower extremities, a visual analog scale pain score greater than or equal to 50, morning stiffness duration of at least 1 hour, reported joint swelling, and a positive IgM-RF or ACPA ( Fig. 2 ). The area under the curve of this model was 0.82 (95% CI 0.75–0.89). Using this risk-calculation rule, 41% had a low risk, 27% had an intermediate risk, and 31% had a high risk of developing arthritis. Compared with the low-risk group, the medium-risk and high-risk groups were both more likely to develop IA, with hazard ratios (HR) of 4.52 (95% CI 2.42–8.77) and 14.86 (95% CI 8.40–28.32), respectively. After 5 years, 12% of the low-risk group developed arthritis, compared with 43% of the intermediate-risk group and 81% of the high-risk group.
This prediction rule was good at identifying low-risk and high-risk patients, suggesting that information routinely available as part of clinical care can effectively risk-stratify patients at the extremes of risk. However, for the 27% in the intermediate-risk group there was still a high level of uncertainty, as 43% developed arthritis within 5 years. These data suggest that perhaps it is in these intermediate-risk patients that incorporating imaging as part of a prediction rule might be particularly informative, and that additional precision is needed to accurately predict progression to clinical RA. Other groups that might also benefit from the addition of imaging to aid with risk stratification include seronegative patients or first-degree relatives of RA patients. Because much of the work in early IA has been done in Europe in ethnically homogeneous cohorts, it will also be important to evaluate the benefit of imaging in patients from a variety of ethnic or racial backgrounds.
Ultrasonography Pros and Cons
US is attractive because it is inexpensive and easy to learn, and individual joints can be scanned quickly. PD adds information on real-time blood flow, and the technique can be easily incorporated into routine care ( Table 1 ). However, results are user-dependent and equipment-dependent, and visualizing multiple joints can be very time-consuming.