Objective To compare whole-body MRI (WB-MRI) with clinical examination in the assessment of disease activity in juvenile dermatomyositis (JDM).
Methods WB-MR images were obtained from 41 JDM patients and 41 controls using a 1.5 T MRI scanner and short τ inversion recovery sequences. 18 patients had follow-up WB-MRI. Muscle, subcutaneous tissue and myofascial signal abnormalities were scored in 36 muscular groups and on proximal and distal extremities. WB-MRI and clinical assessments were performed concurrently and results compared. Validation procedures included analysis of feasibility, reliability, construct validity, discriminative ability and responsiveness.
Results WB-MRI revealed distal legs (26/41 patients) and forearm (19/41 patients) muscle inflammation undetected during clinical examination and allowed an accurate assessment of subcutaneous (23/41 patients) and myofascial involvement (13/41 patients). 27 patients showed a patchy distribution of muscle inflammation while in seven the abnormal hyperintense areas tended to be homogeneously distributed. The inter-reader agreement for muscular, subcutaneous and myofascial WB-MRI scores was excellent. Correlations between WB-MRI muscle score and disease activity measures were excellent (Manual Muscle Test: rs=−0.84, Childhood Myositis Assessment Scale: rs=−0.81). WB-MRI score was higher in JDM active patients when compared with the control group (pB<0.0001) and the inactive patients (pB=0.004), and showed an excellent responsiveness (standardised response mean=1.65). Follow-up WB-MRI showed resolution of inflammation in nine patients whereas clinical criteria for remission were satisfied in five.
Conclusions WB-MRI provides additional information to clinical evaluation and represents a promising tool to estimate total inflammatory burden, tailor treatment and monitor its efficacy.
- Magnetic Resonance Imaging
- Disease Activity
Statistics from Altmetric.com
Juvenile dermatomyositis (JDM) is a rare multisystem connective tissues disorder of unknown aetiology, characterised by chronic inflammation that primarily affects the skin and skeletal muscles. It represents the most common idiopathic inflammatory myopathy (IIM) in childhood with an estimated incidence of 2–3 cases per million children per year.1 ,2 Typical cutaneous manifestations and symmetrical proximal muscle weakness are the clinical hallmarks, while perivascular inflammatory infiltrate and structural changes in the muscle fibres are the histological counterparts.3 Although during the past decade the prognosis of JDM has markedly improved, there are still patients with persistent active disease leading to permanent disability.4 ,5 One of the major challenges in the clinical management of JDM is to accurately assess disease activity to optimise therapeutic strategies; in this perspective, a core set of disease activity clinical measures have been proposed.6 ,7 However, a major limit of these scales concerns their inaccuracy in discriminating active disease from chronic damage, especially in patients with longstanding disease, whose muscle weakness reflects a combination of ongoing muscle inflammation, necrosis, muscle atrophy and fatty replacement. The potential of musculoskeletal MRI as a non-invasive imaging procedure for defining the contribution of each of these components to the patient's weakness is well established.8–12 Water-sensitive sequences (ie, T2-weighted or short τ inversion recovery (STIR) sequences) are in fact very sensitive to the presence of inflammation13 and have the potential to support diagnosis, guide muscle biopsy in an area of active disease14 ,15 and monitor treatment response.10 So far, all MRI studies in JDM focused on pelvic and thigh musculature. Whole-body (WB)-MRI screens the entire body with the advantage of evaluating much larger areas of muscles as well as subcutaneous fat tissue, thus providing a complete assessment of total inflammatory burden in IIMs patients. Even though its value has been suggested in a small series of adult IIMs16 ,17 and in a JDM case report,18 its real potential in childhood dermatomyositis has not yet been explored.
In the present study, we investigated the contribution of WB-MRI to clinical evaluation in JDM disease activity assessment. We also devised a WB-MRI score for the assessment of disease activity and provided preliminary evidence of its validity in a cohort of JDM patients followed at a tertiary level care centre.
Patients and methods
A total of 41 consecutive patients with JDM according to criteria of Bohan and Peter 19 ,20 and seen at our Department between March of 2010 and January 2013 were enrolled. Patients requiring sedation or with contraindication to MRI were not included. Clinical examination was concurrent to WB-MRI and included: physician's global assessment of the disease activity;7 muscle strength assessment using both the Manual Muscle Test (MMT)21 ,22 and the Childhood Myositis Assessment Scale (CMAS);23 ,24 functional ability using the Childhood Health Assessment Questionnaire;25 ,26 the health-related quality of life using the Paediatric Rheumatology Quality of Life Scale;27 the muscular and cutaneous components of the Disease Activity Score;28 the physician's global disease damage assessment;29 and the Myositis Damage Index.30 Laboratory assessments included: serum creatine phosphokinase, lactate dehydrogenase, aspartate aminotransferase, erythrocyte sedimentation rate (Westergren method) and C reactive protein (nephelometry).
Eighteen patients had a follow-up WB-MRI (mean time interval 9.4 months). Treatment efficacy was assessed according to Paediatric Rheumatology INternational Trials Organisation (PRINTO) response criteria for JDM,31 while remission was defined according to PRINTO criteria for clinically inactive disease.32 Clinical measures were assessed simultaneously by a paediatric rheumatologist (AR, SV, AB) along with a paediatric physical therapist (AP) who had considerable experience with clinical measures of JDM and who were blinded to the imaging results.
WB-MRIs that were performed for diagnostic purposes in 41 patients without clinical evidence of muscle disease who were visited in our Department were chosen as controls. This control group consisted of 15 patients with fever of unknown origin, four with juvenile idiopathic arthritis and the remaining affected by chronic recurrent multifocal osteomyelitis. The study protocol was approved by the local institutional review board; informed consents were obtained from parents and, when age appropriate, subjects.
WB-MRI was performed on a 1.5 T MRI scanner (Achieva Intera; Philips Medical Systems, Best, The Netherlands) using the body coil with a field of view covering from the head to toes. WB-MR images were acquired (and displayed) in coronal plane using a turbo spin echo STIR sequence (TR 2800 ms, TE 165 ms, TI 64 ms, slice thickness 5 mm, gap 0.5 mm, NSA 3). Total scan time was 15–20 min. A WB-MRI scoring system was devised by a consensus between paediatric radiologists with a special interest in musculoskeletal MRI (MBD, CM, GMM) and paediatric rheumatologists (CM, AR, NR, AM) with extensive expertise in JDM clinical assessment and scoring system development.
The degree of muscle inflammation was evaluated on 42 muscular groups (anterior neck muscles, posterior neck muscles, cervical paraspinal muscles, deltoid, supraspinatus, infraspinatus and teres minor, biceps brachii, triceps brachii, forearm flexors, forearm extensors, gluteal muscles, iliopsoas, sartorius, adductors, quadriceps femoris, hamstrings, tensor fasciae latae, anterior leg muscles, lateral leg muscles, posterior leg muscles, abdominal muscles, thoracic paraspinal muscles), all easily identifiable on coronal WB-STIR-MRI. Muscle inflammation was scored using a 0–2 point scale (0=no muscle signal abnormalities; 1=mild to moderate signal abnormalities; 2=high degree of signal abnormalities) considering both the degree of visual signal intensity abnormality and its extent within the entire muscles (more or less than 50% of muscle area affected). A score of 2 reflects the presence of areas (even if circumscribed) with high signal intensity and/or signal abnormalities involving more than 50% of the muscle groups. The symmetrical muscles were scored separately. Final score was obtained as an arithmetical sum of all muscular group scores. Perifascicular and subcutaneous tissue inflammation were evaluated using a binary scale (0=absent, 1=increase signal intensity) on eight sites (arm, forearm, thigh and lower leg bilaterally), with a total score ranging from 0 to 8.
A paediatric radiologist (MBD) and a paediatric rheumatologist (CM) blinded to child's identity and disease status (JDM or control patients) independently assigned the WB-MRI scores to all studies. All images were read on large-screen (21″) radiological workstation monitors using a standard PACS software program, after a training session. The follow-up WB-MRI scans were scored without reference to baseline. Time required for assigning the score varied from 5 to 10 min.
Comparison of quantitative variables between two groups was made by the Mann–Whitney U test and among more than two groups by the non-parametric analysis of variance (Kruskal–Wallis test). Bonferroni's correction (PB) was applied in post hoc comparisons. Comparison of dependent quantitative data has been evaluated by the Wilcoxon's test for paired data. Validation procedure included the analysis of reliability, construct validity, discriminant validity and responsiveness. Inter-reader agreement of the WB-MRI total scores was analysed using the method described by Bland and Altman.33 The concordance for each muscle group was analysed by computing-weighted Cohen's k coefficients (weights were calculated as follows: 1-|i-j|/(k-1); i=rows, j=columns of the ratings by the two raters; k=maximum number of possible ratings); for perifascicular and subcutaneous binary items, the simple k coefficient was calculated34 (0–0.2 poor, ≥0.21–0.40 fair, ≥0.41–0.60 moderate, ≥0.61–0.80 substantial, ≥0.81–1 almost perfect).35 All the MRIs were rescored a second time, in a blinded fashion, 12 weeks after the previous review to assess the intraobserver reliability. Given that WB-MRI score could be considered a measure of disease activity, it was predicted that correlations of its value with disease activity parameters would be in the moderate to high range. Correlations with measures of damage were predicted to be poor. All correlations were assessed using Spearman's rank order correlation coefficient (rs). Responsiveness was evaluated by the standardised response mean (SRM), calculated by dividing the absolute mean change by the SD of individual's change in score. The threshold levels for SRM were defined as follows: ≥0.20=small, ≥0.50=moderate and ≥0.80=strong.36 Statistical analysis was performed with Statistica (StatSoft, Tulsa, Oklahoma, USA, release 9.0).
A total of 41 JDM patients (19 boys, 22 girls; median age 8.8) and 41 controls (16 boys, 25 girls; median age 11.9) were included. One patient was diagnosed clinically as amyopathic JDM. Demographic, clinical and imaging findings are reported in table 1. Nine patients newly diagnosed were untreated at time of WB-MRI, while the others were receiving either prednisone alone (N=7) or in different combinations with methotrexate (N=14) and/or cyclosporine (N=14). In the treated group, 15 patients were also receiving hydroxychloroquine and five patients monthly intravenous immunoglobulin.
Clinical examination of muscle strength (MMT) revealed a predominant involvement of proximal compared with distal muscles of the upper (shoulder abduction vs wrist extensors; p=0.001) and lower (gluteus maximus and medius vs ankle dorsiflexors; p<0.0001) extremities.
All patients completed the WB-MRI without problems. Seven patients showed normal muscle signal intensity. The distribution and frequency of the abnormal hyperintense signal in the muscles, subcutaneous tissue and myofascia are reported in figure 1. In addition to the proximal distribution of signal abnormalities, 26 (63.4%) and 19 (46.3%) out of 41 patients showed high signal intensity in distal muscles legs and in forearm muscles respectively that went undetected during clinical examination. No differences in signal intensity were found between proximal and distal muscles. WB-MRI obtained from the amyopathic JDM patient revealed an increased signal intensity within proximal and distal muscles. In all, 27 patients showed the typical, focal and patchy distribution of muscle signal abnormalities while in seven patients the hyperintense areas tended to be diffusely and homogeneously distributed within the muscle, affecting all the muscles throughout the body (figure 2). Subcutaneous tissue and myofascia signal abnormalities were found in 23 (56%) and 13 (32%) out of 41 patients, respectively (figure 1). Patients with subcutaneous and myofascial involvement showed a significantly higher WB-MRI muscle score (median value: 43.5) compared with patients with normal subcutaneous fat and myofascia signals (median value: 7.5; p=0.007).
Bland and Altman plots showed a substantial inter-observer agreement for the muscle, subcutaneous and myofascial WB-MRI total scores (table 2; online supplementary figure S1). Cohen's κ coefficient showed an inter-observer concordance ranging from substantial to excellent for each muscle group with the only exception of anterior neck, abdominal and paraspinal muscles (table 2). Only the muscle groups with Cohen's k coefficient ≥0.7 were included in the final muscle score (resulting in 36 instead of 42 muscle groups).
As reported in table 3, WB-MRI muscle score showed moderate to excellent correlations with clinical indicators of disease activity, such as the MMT (rs=−0.84) and CMAS (rs=−0.81); as expected, they correlated poorly with damage indicators. The WB-MRI subcutaneous and myofascial scores showed moderate to high correlation with physician's global assessment of the disease activity, the Disease Activity Score cutaneous component and Childhood Health Assessment Questionnaire.
Sensitivity to change
Overall, 11 out of 18 patients with imaging follow-up met the PRINTO criteria for improvement. Six out of seven patients who did not improve showed the diffuse and homogeneous pattern of distribution of muscle inflammation. Changes in WB-MRI muscle score paralleled those in clinical parameters concerning muscle assessments; they, in fact, were significantly higher in improved patients (median value: 49) compared with those worsened or stable in their muscle strength (median value: 15, p=0.03). Follow-up WB-MRI revealed substantial reduction of the extent of muscle signal abnormalities resulting in an SRM of 1.65; subcutaneous inflammatory changes disappeared in seven out of nine patients and five out of six recovered from myofascial oedema. The responsiveness of WB-MRI muscle score (SRM of 1.65) appeared higher (with no statistical significance) compared with those of muscle tests (SRM-CMAS=0.56, SRM-MMT=0.74).
All but two patients in the control group showed normal signal intensity in the muscle, subcutaneous tissue and myofascia. WB-MRI muscle score was significantly increased in active JDM patients when compared with either the control group (pB<0.0001) or the inactive patients (pB=0.004), thus indicating an excellent discriminant validity of the score (figure 3). Follow-up WB-MRI showed complete resolution of signal abnormalities in nine patients whereas the clinical criteria for remission were satisfied in five patients.
The present study represents the first application of the WB-MRI to a series of JDM patients demonstrating that this technique, by establishing the distribution and magnitude of inflammatory process throughout the entire body, allows a more accurate assessment of total inflammatory burden compared with clinical examination. Changes in WB-MRI signalling, in fact, were not limited to symptomatic proximal muscles but also shown-up in clinically asymptomatic distal muscles of the limbs. Backing our results is the histological evidence of inflammatory infiltrates and expression of MHC class I and II antigens and of IL-1α on the muscle fibres of patients with DM and polymiositis, present to an equal degree in both proximal and distal muscles, independently from clinical symptoms.37 Although MRI detectable inflammation in muscles not traditionally involved and with clinically apparent normal muscle strength have been anecdotally reported in DM,16 ,18 ,38 the surprising finding emerging from our study was the high frequency of this phenomenon in JDM patients. The potential of this technique to reveal clinically undetectable muscle inflammation was further highlighted by the results obtained from the amyopathic patient. Overall, these findings suggest that a significant degree of disease activity remains clinically unrecognised. The long-term relevance of subclinical inflammation needs to be investigated in additional studies to determine its prognostic meaning and role in therapeutic management.
Another advantage of WB-MRI was its ability to visualise signal abnormalities in the subcutaneous tissue and myofascia, often undetected by clinical assessment. Notably, our results suggest subcutaneous and myofascia signal intensity changes as a potential markers of disease severity, being associated with a higher load of muscle inflammation. In line, Ladd et al have recently demonstrated that abnormal subcutaneous signals in the pelvis and thighs were associated with a more aggressive disease course in a cohort of 45 JDM patients.39 Moreover, subcutaneous changes were indicated as potential precursor to the development of dystrophic calcification in 26 JDM patients.40 In this perspective, WB-MRI, by mapping subcutaneous changes throughout the entire body, becomes a promising tool for verifying the impact of subcutaneous inflammation in the onset of calcification, one of the most debilitating aspects of JDM.
Another interesting finding was the detection of different patterns of muscle inflammation (patchy vs diffuse). An homogeneous distribution of muscle inflammation was previously reported in polymiositis and inclusion body myositis, in contrast to the more focal oedema identified in DM.17 To the best of our knowledge, this is the first description of different patterns of muscle inflammation in JDM. Our results suggest the potential heterogeneity of JDM and the value of WB-MRI in discriminating different disease-subtypes. Even if patients with homogeneous and diffuse distribution of muscle inflammation experienced a worse disease course, the relatively small number of patients and the short observation window might limit the significance of the results. Longitudinal analyses of outcome are required to clarify the prognostic meaning of different patterns of muscle involvement.
The present study also provides evidence for the validity of WB-MRI in assessing disease activity. Due to technical advances, WB imaging of the entire musculoskeletal system is now possible in a reasonable time-frame, with acceptable costs (the cost of WB-MRI is almost superimposable to the cost of localised MRI), without compromising image quality compared with dedicate exams.41 WB-MRI was well tolerated and allowed a reliable characterisation of inflammation, without exposing young patients to ionising radiation and contrast medium agent. The inter-reader agreement ranged between substantial to excellent for each muscle group, with the only exception being the anterior neck, abdominal and paraspinal muscles. Field inhomogeneity artefacts occurring at the edge of the field of view, as for these muscles, may justify the moderate agreement. The close proximity of cervical muscles to anatomical structures (ie, the vascular tree) characterised by high signal intensity in STIR sequences may further contribute to the less than satisfactory agreement.
WB-MRI muscle score correlated closely with measures of muscle strength and function as well as with other disease activity indicators, providing evidence for its construct validity. In line with previous studies using localised MRI,40 changes in subcutaneous tissue were not correlated with most of disease activity measurements, further emphasising the value of WB-MRI in revealing inflammatory signs that are undetected by clinical examination. The fact that myofascial and subcutaneous involvement were limited to the limbs (being the fascia of the thoraco-abdominal wall muscles more susceptible to partial volume effect) could also have contributed to these results.
WB-MRI muscle score was very sensitive in detecting changes due to the treatment and showed an excellent discriminant validity since it was able to differentiate not only between JDM and control group, but also patients with active disease from inactive patients and responders from non-responders. Finally, WB-MRI was more accurate than clinical assessment in detecting remission status. WB-MRI, in fact, revealed no signal abnormalities in four patients who were classified as having active disease due to changes in muscle strength probably due to chronic damage. Discriminating ongoing inflammation from structural damage is essential in guiding treatment decision, especially in patients with longstanding disease for whom the contribution of these components to muscle weakness is not easily assessable by clinical examination alone. In this perspective, the lack of information on muscle structural aberration might be considered a limitation of the present study. The inclusion of T1 sequences, very sensitive in detecting the hallmark of chronicity,42 ,43 would have allowed to be more accurate in chronic damage assessment, increasing however the length procedure and affecting its feasibility. Another possible limitation is the lack of histological data. Although the presence of muscle oedema is not exclusive to IIMs, previous studies have demonstrated that changes on water-sensitive sequences correlated with histological evidence of inflammatory infiltrates.14 ,44 ,45 Finally, this pilot study is limited by the low number of patients with imaging follow-up which was however sufficient to demonstrate the higher responsiveness of WB-MRI compared with clinical measures.
To summarise, our results indicate that WB-MRI enables a reliable analysis of the pattern and the site of inflammation (skeletal muscle, subcutaneous tissue, myofascia), thus providing an accurate estimation of the total disease activity load. By providing additional information with respect to clinical examination, this technique could be useful to tailor treatment according to disease severity.
The authors wish to acknowledge the technical staff of the Radiological Department of the G. Gaslini Institute, in particular Mrs Francesca Maiuri, Mr Paolo Del Mirto, Mr Stefano Franceschi and Mr Marco Ciccone, who supervised the whole-body MRI scans.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online figure
Handling editor Tore K Kvien
Contributors The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors an exclusive licence (or non-exclusive for government employees) on a worldwide basis to the BMJ Publishing Group to permit this article (if accepted) to be published in ARD and any other BMJPGL products and sublicences such use and exploit all subsidiary rights, as set out in our licence (http://group.bmj.com/products/journals/instructions-for-authors/licence-forms).
Competing interests None.
Patient consent Obtained.
Ethics approval Istituto Giannina Gaslini Ethical Board.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The present study represents the first application of the WB-MRI to a series of JDM patients demonstrating that this technique, by establishing the distribution and magnitude of inflammatory process throughout the entire body, allows a more accurate assessment of total inflammatory burden and provides additional information to clinical examination.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.