Comparison of ultrasonography with Doppler and MRI for assessment of disease activity in juvenile idiopathic arthritis: a pilot study

Considering juvenile idiopathic arthritis (JIA), the trend towards early therapeutic intervention and the development of new highly effective treatments have probably improved the outcome in many cases, but they have also increased the need for sensitive and specific methods to evaluate disease activity. According to recent investigations, clinical examination alone is inadequate to identify structures involved in JIA, and US often detects subclinical synovitis[1, 2], especially in the hands and feet[3, 4]. Clinical assessment of disease activity in the small joints of the hand is particularly prone to disagreement among clinicians[5] and the foot and knee are the most frequently involved joints in JIA.

In adult rheumatology, numerous studies have established the important role of MRI[6] and US[7] in assessment of disease activity, and MRI is considered the reference standard for advanced imaging. Nevertheless, due to differences in disease characteristics and the unique features of the growing skeleton, the findings of studies in adults are not directly applicable to children and adolescents[8] and thus far only a few MRI studies of JIA have been conducted, all of which have used different methodologies[9‐12]. Modern imaging techniques have not been used to their full potential in pediatric rheumatology[10, 13], and imaging studies in this context are still rare. Nonetheless, four recent investigations of JIA have shown that US and Doppler US are more sensitive in the detection of disease activity than clinical examination alone[1‐4].

US offers specific advantages over MRI in that it is non-invasive, does not require sedation or general anesthesia, is quickly accessible bedside, and is easy to combine with clinical assessment. Agitation of the patient is rarely a problem, which means that young children can be seated on a parent’s lap or play while being examined, and assessment of multiple locations is possible during a single session. Furthermore, modern high-frequency US transducers, in the hands of an experienced US examiner, provide unsurpassed resolution of the superficial musculoskeletal structures in children.

The aim of the present pilot study was to compare MRI and US with Doppler regarding their value in assessing the different aspects of disease activity in JIA, and to compare the results with images obtained in healthy controls.

A descriptive study with controls was conducted from 2007 to 2011 at the Department of Pediatrics, University of Copenhagen, Denmark. Consecutive patients, under the age of 18 years with clinically active arthritis (i.e., swelling or a limited range of joint movement with joint pain or tenderness) in the wrist region, knee, or ankle region, seen at the Pediatric Rheumatology Outpatient Clinic, Rigshospitalet, Copenhagen, Denmark, were invited to participate.

Eleven JIA patients with 12 clinically active joints consented to take part and were assessed by US and MRI. After the patient cohort had been established, age- and sex-matched controls were recruited among young healthy volunteers who had no history of arthritis or chronic pain; these individuals belonged to families of hospital staff members. The controls were assessed at the Department of Radiology, Skåne University Hospital, Malmö, Sweden, using the same US equipment, the same type of MRI equipment and the same protocols as for the patients in Denmark. One healthy 6-year-old control was excluded because she could not remain still during the MRI examination, and hence the matched patient was also excluded. Thus ten patients (median age 14 years, range 11–18 years) with 11 affected joints (six wrists, three knees, two ankles) and six healthy controls with eight joints (three wrists, three knees, two ankles) were left for evaluation. Patient characteristics and pharmacological treatments are summarized in Table1. Seven of the ten patients (70%) were female. Five had polyarticular (poly) JIA, two oligoarticular (oligo) JIA, one enthesitis-related arthritis, one psoriasis arthritis, and one systemic onset type JIA with polyarticular course.

The local research ethics committees approved the study. All parents gave informed consent for their children to participate, and oral assent was obtained from the children themselves.

The synovial membrane is extremely thin in healthy children and visualization by US or MRI imaging may be difficult. Various systems involving quantitative or semi-quantitative methods are used to grade the synovial perfusion visualized by Doppler flow. Here, we assessed synovial perfusion according to the presence or absence of color Doppler flow. In our present study, Doppler flow was detected in 86% of the 22 diseased areas with synovial hypertrophy on US. Those observations are comparable to our previous findings, using the same definitions of synovial hyperemia, the same US equipment and Doppler settings, in the ankle (89%,[18]) and wrist (88%) regions[19]. Other studies in JIA patients have shown hyperemia in 93% of symptomatic MCP joints[20] and 77% of symptomatic knees[21]. The MRI RAMRIS definition of synovitis[16] used in our study, includes both synovial hypertrophy and increased synovial perfusion[22, 23]. No synovial hypertrophy or synovial hyperemia was detected in any of the controls. This agrees with a recent US study of JIA patients showing that any presence of Doppler flow was significantly associated with clinical synovitis, and that Doppler flow was absent in all the healthy controls[17].

The differentiation of active synovial thickening from joint effusion may be difficult with non-enhanced MR imaging. Physiological joint effusion is common in healthy children, but there is no consensus concerning the normal amount of synovial effusion in healthy individuals. Indeed, a recent MRI study[24] detected fluid in the wrists of children at a relatively large volume that has previously been considered to be pathological in adults[25]. In our investigation, effusion was the only finding in the healthy controls, as shown by US in five joints and by MRI in six joints.

MRI proved to be the best method for identifying erosions in JIA patients (Figure5A + B), which agrees with the results of previous studies of patients diagnosed with rheumatoid arthritis (RA)[26] or JIA[27]. MRI detected erosions in five areas of three wrist regions (Table6). Three of these erosions were detected in anatomical locations that are not accessible to US waves. However, these areas are also difficult to evaluate with MRI and, due to their anatomic peculiarities, they are associated with an intrinsically higher risk of being scored as false positive[28]. Furthermore, a recent MRI study of healthy children demonstrated a high prevalence of bony, erosion-like carpal depressions that increased with age in the normal skeleton of the hand[23], and MRI investigations have revealed numerous changes of the same type in the wrists of healthy adults, primarily in the capitate and lunate bones[29, 30].

An MRI finding of bone marrow edema predicts erosive disease in RA patients[31‐33], whereas the prognostic value of such edema has not yet been established in JIA. A major advantage of MRI over US consists of detecting pathology in locations that are not accessible to US waves, such as the bone marrow. In our study, MRI revealed signs of juxta-articular bone marrow edema in six of the 10 JIA patients (four with poly JIA, one with systemic onset type JIA with polyarticular course, and one with psoriasis arthritis). Bone marrow edema is either rare or absent in healthy adults[30]. In a previous small MRI pilot study of the iliac crest (Laurell, unpublished data), we found that bone marrow edema associated with normal epiphyseal growth in healthy young individuals was difficult to distinguish from the pathological edema caused by enthesitis in JIA patients. Furthermore, other pediatric studies have reported similar MRI findings of physiological edema in healthy individuals at the iliac crest[34], in the wrist region[24], and in the ankle region[35, 36]. In the present investigation, MRI revealed multiple and patchy heterogeneous marrow signal changes on STIR images in five joints of four control subjects, suggestive of isolated foci of residual haematopoietic red marrow, or of focal bone marrow edema. Consequently, it may be difficult to use MRI to detect pathological bone marrow edema in children and adolescents.

There are no validated MRI or US scoring systems for the assessment of inflammatory and joint damage abnormalities in JIA, and there is little knowledge of the normal US and MRI reference values of each joint at different developmental stages in children. It is possible that a large part of the knowledge obtained in studies in adult rheumatology can be applied to children as well, but the imaging techniques must be further validated in all fields of pediatric rheumatology. In our study we adopted the OMERACT[15] definitions for US pathology, and the RAMRIS definitions[16] for MRI pathology. The MRI RAMRIS system, designed for use in adults, has been validated in the wrist and MCP joints of adult subjects, and was recently also tested on wrists in a pediatric population.

For assessment of children, US offers specific advantages over MRI in that it is non-invasive, does not require sedation or general anesthesia (which facilitates repeated examinations for follow-up), is quickly accessible bedside, and is easy to combine with clinical assessment (interactivity). On the other hand, MRI gives an overview and also detects pathology in locations that are not accessible to US waves. This last mentioned factor might explain some of the differences between MRI and US results in our study evaluating complex joint structures like the wrist regions and ankle regions. Advantages and disadvantages of MRI and US, respectively, are summarized in Table7. In our study the positioning of joints also differed slightly between the MRI and US examinations, which might have influenced the distribution of effusion on MRI and US images, respectively. On MRI images intravenous contrast is required to differentiate synovial tissue from effusion; the differentiation is easily performed on the fat-saturated post-contrast MRI sequences. The need for intravenous contrast injection is a disadvantage in dealing with pediatric patients. On US images, synovial tissue is hypoechoic and effusion is generally anechoic, but the differentiation is not always easy and confusion might occur as fluid may be very hypoechoic, nearly anechoic, and effusion containing particles may be hypoechoic. In anatomical areas where compression with the transducer is easily performed, e.g. the suprapatellar recess, differentiation between synovial tissue and effusion is facilitated (Figure2A). Doppler US examination, displaying vascularization, and dynamic US examination, during active or passive mobilization of the soft tissues examined, may also facilitate such differentiation. In children, the hypoechoic synovial tissue might also be difficult to distinguish from the hypoechoic cartilage of the epiphyses. Doppler examination is generally not a solution to this problem, since vascularization can be present in both hypertrophic synovial membranes and cartilaginous epiphyses during growth. Therefore, to avoid diagnostic errors, it is important to have good knowledge of the normal US appearance of each joint at different stages of development, and it is also imperative to use a meticulous scanning technique that allows clear interpretation of possible anisotropic artifacts[37] (Laurell, 2012, Clinical and Experimental Rheumatology, accepted for publication).

In our study two small and flat Baker’s cysts with a thin synovial wall were visualized on fat-saturated T1w post-contrast MRI sequences, but were not detected on US. US is more sensitive than MRI in detecting small volumes of effusion, but in these cases no fluid content, or hyperemia of the synovial tissue, could be visualized on Doppler US. The reason for the discrepancy between the two imaging modalities in detecting tendon sheath involvement is unclear. It can be essential to apply as little pressure as possible with the US transducer during the examination of superficial tendons in order to visualize vascularization on Doppler examination, and synovial hypertrophy might be difficult to distinguish from effusion within a tendon sheath.

A main disadvantages of MRI, is that it requires sedation of young children, an age group with a high prevalence of JIA[38]. In our study, a 6-year-old patient had to be excluded, because the age-matched control failed to remain still during the MRI procedure, while US examinations could be performed on both individuals without any objections.

The present investigation was descriptive in nature and was not designed to compare the results of clinical assessment and imaging, but rather to compare two imaging modalities. Four different anatomical areas were assessed in the knee and ankle region, respectively, and three different areas in the wrist region. The objective was to investigate whether the US and MRI findings, respectively, occur in the same or in different anatomical locations. The inclusion criterion was being a child with clinically active arthritis. Patients without focal clinical symptoms were not presented to the US examiner, who was also blinded to other aspects of the children, such as clinical status and subtype. As the examinations on patients and controls were performed at different occasions and at different locations the examiner was not blinded to whether he was evaluating JIA patients or controls. Another experienced musculoskeletal radiologist analyzed all MRI images. He was blinded to the results of US examinations but, as the dates of examination were differing between the two groups, not to the respective subject category (patients or controls), which might have constituted a bias.

Another study has previously examined the role of MRI versus US in assessing knee inflammation in JIA[9], but to our knowledge, our pilot study is the first to compare results of MRI and Doppler US for the overall assessment of all aspects of JIA (i.e., considering synovial hypertrophy, synovial perfusion/enhancement, effusion, bone erosions, and bone edema). A weakness of the present study is that the US assessments and the MRI examinations, respectively, were evaluated for accuracy by only one experienced musculoskeletal radiologist. Furthermore, the discrepancies we observed between the US and MRI findings might have been related to the time span between the two examinations, as well as to the fact that most of the patients had received pharmacological treatments with a potential effect on synovitis. The US examination protocol used in our study did not include the radio-ulnar joint and hence we cannot rule out any synovitis in this area of the wrist, but MRI showed involvement of this joint in 5 out of 6 wrists examined (Table3). Accordingly, in future investigations of US examination of the wrist in JIA, we will use a revised and more appropriate scanning protocol that also includes the radio-ulnar joint.

Although small, this study has yielded results indicating that both MRI and US provide valuable imaging data on disease activity in various joints of children with JIA. Importantly, the two techniques seem to complement each other and give partly different information on the patients who are assessed. In JIA, as well as in adult rheumatology, US seems to provide useful imaging information that could make it an option, on many occasions, in daily clinical practice. However, it should be pointed out that the current work represents a pilot study, and thus our results need to be confirmed in a larger prospective clinical investigation.