Abstract
Objective. To perform a systematic literature review of the diagnostic and prognostic value of magnetic resonance imaging (MRI) and ultrasound (US) in patients with undifferentiated peripheral inflammatory arthritis (UPIA), and to assess if MRI and US should be done at baseline and repeated, and if so, at what interval.
Methods. Medline, Embase, the Cochrane Library, and abstracts presented at the 2007 and 2008 meetings of the American College of Rheumatology and European League Against Rheumatism meetings were searched for diagnostic and prognostic studies of any duration examining the ability of MRI/US to predict outcome of patients with UPIA. Sensitivity, specificity, predictive values, and positive/negative likelihood ratios (LR+/LR−) were calculated. When available, odds ratios were extracted. Quality was appraised using validated scales.
Results. Regarding MRI, 11 out of 2595 screened references were included: 2 described pure undifferentiated arthritis (UA) populations and 9, mixed populations. Bone edema (LR+ 4.5) and combination of a distinct MRI synovitis and erosion pattern (LR+ 4.8) increased probability of developing rheumatoid arthritis (RA). Absence of MRI synovitis (LR− 0.2) and absence of a distinct synovitis pattern (LR− 0) decreased probability of developing RA. Regarding US, 2 out of 2111 references were included, both mixed populations; no data could be extrapolated for UPIA.
Conclusion. MRI bone edema and combined synovitis and erosion pattern seem useful in predicting development of RA from UPIA. The value of US in UPIA remains to be determined. The absence of MRI synovitis seems useful in excluding development of RA. No data were found about the value of repeating MRI/US. Studies evaluating MRI/US in UPIA are scarce, but current knowledge strongly encourages further testing in UA.
Within the field of imaging in rheumatic diseases, large and exciting advances have been made over the last decade. Although radiographs continue to be the most widely used tool, magnetic resonance imaging (MRI) and ultrasound (US) offer advantages through more sensitive depiction of inflammatory and destructive disease manifestations1.
In the context of undifferentiated peripheral inflammatory arthritis (UPIA), patients’ questions will focus on diagnosis and prognosis: the likelihood of developing a well defined rheumatic disease, what the future holds for disease progression, persistence, functional impairment, and quality of life. The answers to these questions are vital for clinical decision-making, including the choice of treatment2.
This article is part of the 3e Initiative (Evidence, Expertise, Exchange) in Rheumatology, 2008–20093,4,5. The resulting 10 recommendations on “How to investigate and follow-up UPIA” are described in more detail elsewhere5. The objective of this article was to systematically review the available literature about the following question: “What is the diagnostic and predictive value of MRI and US in patients with UPIA? Should they be done at baseline and repeated at what interval?”.
MATERIALS AND METHODS
Strategy and criteria for considering studies for this review
The clinical question was structured in the PIO format6 (Patients, participants or problem; Intervention or index test; Outcomes or target conditions), and the eligible types of study were defined.
Patients were defined as “adults with UPIA.” The definition of UPIA is controversial and there is no widely accepted classification criterion for this condition. During the 2008–2009 3e Initiative kickoff meeting, experts decided that only patients in whom clinically apparent joint swelling (synovial proliferation or synovial effusion) was observed by the rheumatologist should be included. This is in contrast to some reports that have included patients with inflammatory joint symptoms in the absence of clinically observable joint swelling (a state usually referred as “inflammatory arthralgia”). It was also emphasized that the terms “early arthritis” and “undifferentiated arthritis” should not be considered interchangeable or similar. For the current systematic review, participants were to be patients that, after initial visits and diagnostic investigations, did not fulfill the diagnostic/classification criteria for any rheumatologic disorder. Because we anticipated that very few studies would have included truly undifferentiated populations at baseline, we also kept a record of results from studies in mixed populations (e.g., UPIA + arthralgia, UPIA + early RA), as these could be useful for extrapolating results.
The index test was defined as a certain MRI feature (e.g., synovial fluid, synovitis, erosion, bone edema, and tenosynovitis) or US feature [e.g., US power-Doppler (PD) and US greyscale (GS) scores], as defined in the study.
The outcomes were defined as the development of well defined rheumatic diseases (e.g., RA, psoriatic arthritis) or relevant disease outcomes (e.g., remission, radiographic progression). The use of internationally validated diagnostic/classification criteria [e.g., 1987 American College of Rheumatology (ACR) criteria for RA7] and validated outcome measures was given more value when appraising the definition of outcome.
Three types of studies were considered for inclusion: (1) cohort studies in which patients from a given UPIA population had MRI or US at baseline and in whom the outcome after a period of followup was recorded; (2) retrospective case-control studies in which patients had MRI or US at baseline and in whom it is known that they had UPIA when the baseline investigation was performed; and (3) randomized controlled trials of patients with UPIA that implicitly addressed the question of diagnostic or prognostic value, as each arm of a trial can be seen as a cohort study.
Methodology
Details of search methods for identification of studies, selection of articles, data extraction and analysis, and quality assessment used in the selection and appraisal of the articles can be found in Appendix 1, 2, and 3 of the online version (www.3eupia.com).
RESULTS
Magnetic resonance imaging
A total of 1734 articles and 861 meeting abstracts were found. After title and abstract screening, 15 articles8,9,10,11,12,13,14,15,16,17,18,19,20,21,22, 3 meeting abstracts (already published or later published in article format10,11,23, and one additional paper from hand searching24 remained for review. The inclusion criteria were fulfilled by 11 articles8,9,10,11,12,13,14,15,16,17,23, which were included in the systematic review. Two articles included truly undifferentiated populations8,23 while the other 9 included mixed populations9,10,11,12,13,14,15,16,17 at baseline. A detailed flowchart can be found in Appendix 4 and reasons for exclusion after full article review can be found in Appendix 6 of the online version (www.3eupia.com).
UPIA populations
Study characteristics and results for UPIA populations are summarized in Tables 1 and 2.
Tamai, et al23 evaluated 129 patients with UPIA; all the patients expressed rheumatic manifestations of the wrists and finger joints at study entry. At prospective followup after 1 year, 75 patients (58.1%) had disease progression that fulfilled 1987 ACR criteria for RA7. Contrast enhanced MRI images were evaluated for bone edema, bone erosion, and synovitis at 15 sites in each finger and wrist. Patients who were positive for at least 2 of 3 objective measures [anti-cyclic citrullinated peptide antibodies (anti-CCP) and/or IgM rheumatoid factor, MRI-proven symmetric synovitis, and MRI-proven bone edema and/or bone erosion] progressed to RA at 1 year with a positive likelihood ratio (LR+) of 2.8 and a negative likelihood ratio (LR−) of 0.4 [sensitivity (SE) 68%, specificity (SP) 76%]. Further, in 22 UPIA patients positive for both anti-CCP and MRI-proven bone edema who were considered to have progressed to RA at 1 year, the SP and positive predictive value (PPV) was increased to 100% (however, SE was 29%). Anti-CCP alone and bone edema alone had SP of 93% and 91%, respectively (SE 57% and 41%, respectively). MRI synovitis had a LR− = 0.2 regarding progression to RA (SE 91%, SP 44%).
Duer, et al8 investigated 41 patients with arthritis and subjective symptoms in the hands, who remained unclassified despite conventional clinical, biochemical, and radiographic examinations. Patients who fulfilled the 1987 ACR criteria for RA7 or had radiographic bone erosions were excluded. Contrast enhanced MRI of the wrist and 2nd–5th metacarpophalangeal joints of the most symptomatic hand was performed and the MRI pattern was compared with final diagnosis after a 2-year followup period (RA vs non-RA, according to 1987 ACR criteria). The combination of a distinct MRI synovitis and erosion pattern of RA (for definitions see Table 2) had a LR+ = 4.8 and a LR− = 0.4 (SE 64%, SP 87%) for development of RA. When the synovitis and erosion pattern of RA was combined with a scintigraphy pattern of RA, SP and PPV increased to 100%, but at the cost of a low SE (45%). MRI bone edema was not assessed in this study. That same MRI synovitis pattern alone had a LR− = 0 for progression to RA (SE 100%, SP 60%).
Ultrasound
A total of 1250 articles and 861 meeting abstracts were found. After title and abstract screening, 3 articles19,20,25 and 3 meeting abstracts (already or later published in article format26,27) were retained for full article review. Inclusion criteria were fulfilled by 2 articles (mixed populations only). A detailed flowchart can be found in Appendix 5 (www.3eupia.com) and reasons for exclusion after full article review can be found in Appendix 6 of the online version (www.3eupia.com). Study characteristics and results for the 2 mixed populations26,27 are summarized in Tables 5 and 6.
DISCUSSION
This systematic review summarizes and evaluates available evidence on the value of MRI and US in UPIA.
Our results show that MRI bone edema (LR+ 4.5) is more likely to be seen in patients with UPIA who develop RA than in those who do not, and that the combination of MRI bone edema and anti-CCP positivity is highly specific for development of RA (LR+ infinite, i.e., SP = 100%)23. However, the absence of both these features does not allow excluding development of RA23. On the other hand, results also showed that patients without MRI synovitis have decreased probability of developing RA (LR− 0.2)23.
In another study, the combination of a distinct MRI synovitis and erosion pattern with involvement of several hand joints, but not the first carpometacarpal joint, was more likely to be seen in UPIA patients who developed RA (LR+ 4.8) than in patients who did not8. The combination of MRI pattern plus scintigraphy pattern with involvement of several joints (but not distal interphalangeal joints or first carpometacarpal joint) was even more specific for development of RA (LR+ infinite, i.e., SP = 100%)8. However, again, none of these features ruled out development of RA8. On the other hand, results also showed that patients without the above MRI synovitis pattern had decreased probability of developing RA (LR− 0, i.e., SE for RA = 100%)8.
Results based on MRI studies in mixed populations9,10,12,13,14,15,16,17 must be viewed with caution due to the heterogeneity of the study populations and the different measurements and outcomes that were used and that made the pooling of data impossible. Overall, they provide some evidence for the usefulness of MRI (bone edema, synovitis, and erosions) in predicting RA, but direct extrapolation of results to UPIA cannot be done.
Regarding US, no studies were found in UPIA. We describe one study in a cohort of patients with very early inflammatory hand symptoms26 and another in a population with mainly (very) early RA27. Again, extrapolation of results to UPIA cannot be done, although they suggest that US-PD signal and US-GS synovitis can be regarded as potential candidates for future studies in UPIA. However, their usefulness in this population remains undetermined.
Definite answers about the diagnostic and prognostic value of MRI and US in UPIA can be achieved only through well conducted longitudinal studies of patients with UPIA. Studies of this kind are scarce, particularly in truly undifferentiated populations. The value of MRI and US should be compared with other potentially useful variables; this should be done not only by assessing the performance of the single variables alone, but also using multivariate logistic regression analysis with the aim of developing the best possible predictive model, which has never been done taking into account MRI and US28. The definition of a positive index test is also of great importance; ideally this should be done using validated and reproducible scoring systems. For the clinician, US may have some advantages due to low operating costs and easy accessibility; however, extremity MRI has potential to address the question of high costs of MRI. Lastly, no data were found about the value of repeating MRI or US in UPIA, and this should also be a matter of future study. Recent new ACR/EULAR criteria for RA29 should also be taken into account in the future, as several of the patients we describe as having UPIA will likely be labelled as RA.
In conclusion, a distinct MRI pattern of erosion and synovitis and presence of MRI bone edema increased the probability of developing RA from UPIA; however, some patients with UPIA presenting these MRI features may remain undifferentiated, or develop other diseases, or have a self-limited course. The absence of MRI synovitis decreased the probability of developing RA; however, some patients without MRI synovitis may still develop RA. Regarding US assessment, US-PD signal and US-GS synovitis are potential candidates for future studies in UPIA. Current knowledge already provides evidence for the usefulness of MRI in UPIA and strongly encourages further testing of both MRI and US in undifferentiated arthritis.
Acknowledgments
We thank J.W. Schoones, Walaeus Library, Leiden University Medical Center, The Netherlands, who participated in the elaboration of the systematic search strategy; and all participants of the 3e, especially the bibliographic team, who participated in developing the search strategy and planning the analyses.
Footnotes
-
Supported by an unrestricted educational grant from Abbott. Abbott had no role in the design, literature search, data collection, data analysis, data interpretation, or writing of this report. Dr. Bombardier holds a Pfizer Chair and Canada Research Chair in Knowledge Transfer for Musculoskeletal Care.