Abstract
Objective
We aimed to determine quantitative sacroiliac joint (SIJ) magnetic resonance imaging cut-offs for active and structural lesions that will be incorporated as imaging domains in classification criteria of axial disease in juvenile spondyloarthritis (JSpA).
Methods
MRI scans from an international cross-section of JSpA patients were reviewed by six musculoskeletal imaging experts blinded to clinical details. Raters globally assessed the presence/absence of lesions typical of axial SpA and performed SIJ quadrant or joint based scoring. Sensitivity and specificity of lesion cut-offs were calculated using rater majority (≥4/6 raters) on a global assessment of the presence/absence of active or structural lesions typical of axial SpA with high confidence as the reference standard. Cut-offs were validated in an independent cohort.
Results
Imaging from 243 subjects, 61% male, median age 14.9 years, had sequences available for detailed MRI scoring. Optimal cutoffs for defining lesions typical of axial disease in JSpA were: 1) inflammatory lesion: bone marrow edema in ≥3 SIJ quadrants across all SIJ MRI slices (sensitivity 98.6%, specificity 96.5%); 2) structural lesion(s): erosion in ≥3 quadrants or sclerosis or fat lesion in ≥2 SIJ quadrants or backfill or ankylosis in ≥2 joint halves across all SIJ MRI slices (sensitivity 98.6%, specificity 95.5%). Sensitivity and specificity of the optimal cut-offs in the validation cohort were excellent.
Conclusion
We propose data-driven cut-offs for active inflammatory and structural lesions on MRI typical of axial disease in JSpA that have high specificity and sensitivity using central imaging global assessment as the reference standard and excellent reliability.
Keywords: MAGNETIC RESONANCE IMAGING, PEDIATRIC RHEUMATOLOGY, JUVENILE INFLAMMATORY ARTHRITIS, SPONDYLOARTHRITIS
Magnetic resonance imaging (MRI) of the pelvis is a key part of the assessment of axial disease in adults and children with spondyloarthritis (SpA). MRI studies reported as consistent with or suggestive of sacroiliitis greatly influence therapeutic decision making for children, underscoring the importance of accurate interpretation. However, the anatomy of the maturing sacroiliac joint poses challenges because in healthy children the sacral cortex may be more irregular in appearance and more difficult to visualize than the iliac cortex. In fact, as many as 57% of healthy children have cortical irregularities (1, 2). Additionally, healthy children can have metaphyseal-equivalent signal intensity that can be mis-interpreted as subchondral bone marrow edema (BME) for those unfamiliar with the appearance of the normal maturing sacroiliac joint (SIJ) (1, 3, 4).
The Assessment of SpondyloArthritis international Society (ASAS) developed definitions of a positive MRI in adults in 2009, based upon consensus methodology (5). Subsequently, several cross-sectional studies have shown that lesions meeting the ASAS imaging criteria for classification can be observed in up to 40% of healthy adults (6–8). The ASAS MRI working group updated the definition of a positive MRI for axial disease in 2016 (9). Recently, a data-driven quantitative definition of ≥4 sacroiliac joint quadrants with BME at any location or at the same location on more than 3 consecutive slices was proposed for active lesions; this definition attained a specificity of ≥95% (10). Similarly, data-driven quantitative definitions of structural change typical with axial SpA in adults were proposed and included the presence of any of the following: ≥3 SIJ quadrants with erosion, ≥5 quadrants with fat lesions, erosion at the same location for ≥2 consecutive slices, fat lesions at the same location for ≥3 consecutive slices, or the presence of a deep fat lesion. These definitions for both active and structural lesions were derived from an external reference criterion comprised of majority expert reader decision that the global evaluation of the MRI scan was definitely indicative of active or structural lesions typical of axial SpA. The definitions also had high PPV (≥95%) for a clinical diagnosis of axial SpA in adults after a mean follow up of 4.4 years (10). These definitions have not been tested in the juvenile population.
Determining optimal cut-offs for active and structural imaging lesions of JSpA with axial disease is an important step in developing classification criteria, especially with the targeted therapies emerging on the market that need to be evaluated for efficacy and safety in the juvenile population. Since it is well established that the maturing SIJ looks different from the adult SIJ, it is likely the same criteria proposed for adults with SpA may not be applicable. As part of a larger study developing classification criteria for axial disease in children with SpA, the objective of this project phase was to define MRI findings typical of juvenile SpA with axial disease for use in the imaging domains of the classification criteria. Specifically, we aimed to determine quantitative SIJ imaging lesion thresholds to define a positive MRI for inflammatory and structural lesions typical of axial JSpA using majority imaging expert rater assessment as the reference criterion.
PATIENTS AND METHODS
IRB
This study was reviewed by the Children’s Hospital of Philadelphia (CHOP) Institutional Review Board (IRB) and the IRB determined the procedures met the exemption criteria per 45 CFR 46.104(d) 4(iii).
Patient and Public Involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Subjects and Data
This was an international cross-sectional study of children with clinically diagnosed SpA who met the provisional Pediatric Rheumatology International Trials Organization (PRINTO) criteria for enthesitis/spondylitis-related juvenile idiopathic arthritis or had a rheumatologist juvenile SpA diagnosis between 2006 and 2021 from 6 institutions across Europe (Istanbul, Turkey and Sankt Augustin, Germany) and the United States (Bethesda, MD, Birmingham, AL, Madison, WI, and Philadelphia, PA). Subjects were identified through medical record query or through physician referral and had symptom onset prior to age 18 years and underwent MRI as part of a diagnostic evaluation for axial disease. Clinical data were abstracted from subjects’ medical records and collected and managed using REDCap electronic data capture tools hosted at CHOP (11). Digital Imaging and Communications in Medicine (DICOM) MRI files were transferred using a secure filesharing platform. The validation cohort was an independent cohort of youth collected from 11 institutions (United States, Germany, Turkey, Brazil, India, and Belgium) using the same inclusion criteria. The study size was calculated for the larger project developing classification criteria for axial disease in children with SpA. Research reporting follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (12).
Central Imaging Review
To enable SIJ detailed scoring, all MRI scans had dedicated views of the SIJ and included semicoronal slices through the cartilaginous part of the joint on fluid sensitive sequences (STIR or T2 fat-saturated) for assessment of inflammation and semicoronal slices through the cartilaginous part of the joint on T1-weighted sequences for structural lesions. All studies were performed on 1.5 or 3.0T MRI and, since imaging was not protocolized, technical parameters such as field of view, repetition time (TR), echo time (TE), and inversion time (TI) varied between institutions.
Evaluation of the SIJ by MRI was based on the detection of inflammatory and structural lesions meeting standardized definitions for each type of lesion. These definitions were decided by consensus of the central imaging team and represented a mix of definitions recently developed or updated by ASAS and the Juvenile Idiopathic Arthritis MRI Score (JAMRIS) Outcome Measures in Rheumatology (OMERACT) working group (5, 13, 14). The definitions and their origin (ASAS or JAMRIS) are listed in Table 1. All MRI scans were reviewed by the six central raters blinded to clinical details using an online eCRF (www.carearthritis.com/service/mri-scoring-modules/). For the validation exercise, all MRI scans were reviewed by the three central raters blinded to clinical details. All central raters are experts in the evaluation of musculoskeletal imaging. The eCRF comprised two sections. In the first section data was entered into the ASAS MRImagine module, a global scoring form where readers recorded the presence/absence of each type of active and structural lesion in iliac and sacral portions of each SIJ. For each lesion, a link was provided to a consensus-based reference image with the wording of the lesion definition. Readers provided a yes/no response to two standardized questions as follows: 1) “Are there typical acute/active inflammatory lesions compatible with axial SpA present in SI joints or at entheseal sites outside the SI joint?” 2) “Are typical structural chronic lesions present in the SI joints?”. If inflammatory lesions were reported, readers were asked if the MRI fulfilled the ASAS definition of a positive MRI (15). In the second eCRF section and after the ASAS global assessments were completed, data was entered onto a granular scoring web-based interface where inflammatory and structural lesions were recorded in each SIJ quadrant or half on consecutive semicoronal slices using the Spondyloarthritis Research Consortium of Canada (SPARCC) SIJ inflammation (SIS) and structural (SSS) scoring system. All slices with a minimum 1cm vertical height of the SIJ cartilaginous compartment were scored and SIJ quadrants/halves were defined according to established rules (16). Lesions were recorded in either SIJ quadrants (subchondral BME, inflammation in erosion cavity, erosion, fat, sclerosis) or upper and lower SIJ halves (fat metaplasia in an erosion cavity, ankylosis) as previously defined (16, 17).
Table 1.
Definitions of Lesions for central imaging review.
| Definitions of Lesions for central imaging review | Origin | Fleiss’ kappa (95% CI) | |
|---|---|---|---|
| Subchondral inflammation/bone marrow edema | An ill-defined area of high bone marrow signal intensity on fluid-sensitive sequences within the subchondral bone of the ilium or sacrum compared to the signal intensity of the iliac crest, edges of the vertebrae, and triradiate cartilage and in comparison to physiological changes normally seen on MRI examinations of age- and sex-matched children, and visible in 2 planes | JAMRIS | 0.70 (0.65-0.76) |
| Inflammation at the site of erosion | Increased signal on STIR and/or T1FS post-Gd at the site of erosion | ASAS | 0.56 (0.49-0.62) |
| Erosion | A defect in subchondral bone associated with full-thickness loss of the dark appearance of the subchondral cortex at its expected location, with loss of signal on a T1-weighted non-fat-suppressed sequence compared with the normal bright appearance of adjacent bone marrow | ASAS | 0.72 (0.67-0.77) |
| Fat lesion/ fat metaplasia | Homogeneous increased signal intensity within the subchondral bone on T1-weighted non-FS image presenting with a distinct border of the lesion | JAMRIS | 0.43 (0.30-0.57) |
| Fat metaplasia in an erosion cavity (Backfill) | Bright signal on a T1-weighted sequence in a typical location for an erosion or confluent erosions, with signal intensity greater than normal bone marrow, which meets the following requirements: a) associated with complete loss of the dark appearance of the subchondral cortex at its expected location. B) clearly demarcated from adjacent bone marrow by an irregular band of dark signal reflecting sclerosis at the border of the original erosion. | ASAS | 0.41 (0.25-0.58) |
| Sclerosis | A substantially wider than normal (of ≥ 5 mm in adolescents) area of low subarticular bone signal on T1-weighted and fluid sensitive images | JAMRIS | 0.53 (0.45-0.60) |
| Ankylosis | Presence of signal equivalent to regional bone marrow continuously bridging a portion of the joint space between the iliac and sacral bones | JAMRIS | 0.50 (0.15-0.86) |
Analysis
To identify the optimal MRI definitions for active and structural lesions typical of juvenile axial disease, the reference standard was a central imaging majority (≥4/6 raters) designation of high confidence (≥+3 on confidence scale from −5, “Definitely No”, to +5, “Definitely Yes”) in a global inflammation assessment (“are there typical acute/active inflammatory lesions compatible with axial SpA present in the SI joints?”) or a global structural lesion assessment (“are typical structural chronic lesions present in the SI joints?”). As with most classification criteria, we aimed for specificity of detection over sensitivity.
Interrater agreement for the presence/absence of active and structural lesions typical of juvenile axial disease, each of the lesions listed in Table 1, and ASAS definition of positive MRI was assessed using Fleiss kappa coefficient. Kappa coefficients were interpreted as poor ≤0.40, fair 0.41-0.59, good 0.60-0.74, and ≥0.75 excellent.
The top performing inflammation and structural definitions for definite lesions typical of axial JSpA were assessed in an independent validation cohort using diagnostic test statistics, and the reference standard was a central imaging majority (≥2/3 raters) designation of high confidence (≥+3 on confidence scale from −5, “Definitely No”, to +5, “Definitely Yes”) in a global inflammation assessment (“are there typical acute/active inflammatory lesions compatible with axial SpA present in the SI joints?”) or a global structural lesion assessment (“are typical structural chronic lesions present in the SI joints?”).
RESULTS
Subjects were an international cross-sectional convenience sample of children with SpA who were evaluated for suspected axial disease. Imaging from 243 patients, 61% male, median age 14.9 years (range 4.7 years to 20.3 years) had sequences available for detailed MRI scoring. MRI sequences were available for detailed scoring of active inflammation-only (N=17), structural lesions-only (N=4), or both inflammation and structural lesions (N=222). Active inflammatory and structural lesions typical of axial disease (ASAS MRImagine global assessments) were observed on MRI by a majority of central imaging raters with high confidence in 30% of the 239 cases assessed for inflammatory lesions and 31% of the 226 cases assessed for structural lesions. Of the 243 studies, 65% of cases had no imaging findings of inflammatory or structural lesions on global assessment by ≥4 raters with high confidence. Fleiss’ kappa for interrater agreement for the designation of the presence/absence of active and structural lesions typical of juvenile axial disease on the ASAS MRImagine global assessments at any confidence level was excellent at 0.81 (95% CI 0.76-0.86) and 0.80 (95% CI: 0.75-0.86), respectively; when restricted to only cases where at least 4 of the raters reported presence or absence of lesions typical of SpA with high confidence (stronger than a confidence rating of ±3), agreement was 0.88 (95% CI 0.84-0.93) for inflammatory lesions (N=231) and 0.88 (95% CI 0.84-0.92) for structural lesions (N=215). Fleiss’ kappa for the ASAS positive MRI definition at any confidence level was 0.63 (95% CI: 0.50-0.76). Table 1 shows interrater agreement, as measured by Fleiss’ kappa for the presence/absence of individual inflammatory and structural lesions assessed.
Table 2 shows the performance of cut-offs for presence of definite inflammatory lesions typical of axial disease. The lesion-based cut-offs achieving specificity ≥95% were BME in ≥3 quadrants across all SIJ slices, BME in ≥4 quadrants across all SIJ slices, BME lesion in the same location on ≥3 consecutive slices, and ≥1 quadrant with inflammation in an erosion cavity across all SIJ MRI slices. Of these definitions, the sensitivity of BME in ≥3 SIJ quadrants across all slices at 98% was the highest. Figure 1 shows two demonstrative cases that meet this definition.
Table 2.
Performance of cut-offs for presence of definite inflammatory lesions typical of axial disease
| Cut-offs for number of SIJ quadrants (any location) with majority (≥4/6) rater agreement for definite active lesion | Sensitivity (95% CI) | Specificity (95% CI) |
|---|---|---|
| BME score ≥1 | 100 (95-100) | 85.9 (79.7-90.7) |
| BME score ≥2 | 100 (95.0-100) | 93.5 (88.7-96.7) |
| BME score ≥3 | 98.6 (92.5-100) | 96.5 (92.5-98.7) |
| BME score ≥4 | 95.8 (88.3-99.1) | 97.1 (93.3-99.0) |
| BME lesion, same location on ≥3 consecutive slices | 88.6 (78.7-94.9) | 98.8 (95.8-99.9) |
| BME score ≥4 or BME lesion at same location on ≥3 consecutive slices* | 97.3 (90.5-99.7) | 94.8 (90.3-97.6) |
| Inflammation in erosion cavity (any) | 56.2 (44.1-67.8) | 98.3 (95.0-99.6) |
| BME score ≥3 or inflammation in erosion cavity (any) | 98.6 (92.5-100) | 95.3 (90.9-97.9) |
Sacroiliac joint (SIJ) quadrants = upper ilium, lower ilium, upper sacrum, lower sacrum for left and right SIJ. Number of quadrants = total # of quadrants with detectable lesion across all slices of study.
Criteria adopted for positive MRI in adults.(10) BME = bone marrow edema; SIJ = sacroiliac joint; CI = confidence interval.
Figure 1. Examples of Positive MRI for Active Inflammation in Juvenile SpA.
A) There is increased subchondral signal on coronal oblique short tau inversion recovery (STIR) sequences in the left and right upper and lower sacral quadrants; left and right lower iliac quadrants (total: 6 quadrants with BME). B) There is increased subchondral signal on coronal oblique T2 fluid sensitive sequence in the left upper and lower sacral quadrants and upper and lower left iliac quadrants (total: 4 quadrants with BME).
Table 3 shows the performance of cut-offs for presence of definite structural lesions typical of axial disease. Using the central imaging majority assessment with high confidence of definite structural lesion as the reference standard, several lesion cut-offs achieved ≥95% specificity including erosion ≥2 quadrants, sclerosis in ≥1 quadrant, fat lesion in ≥1 quadrant, and backfill in ≥1 SIJ half. Several combinations of lesions also achieved ≥95% specificity (Table 3). The optimal definition was erosion in at least 3 quadrants or sclerosis or fat lesion in at least 2 SIJ quadrants or backfill or ankylosis in at least 2 joint halves across all SIJ MRI slices (sensitivity 98.6%, specificity 95.5%). Cases that meet this definition are show in Figure 2. Table 4 shows the frequency with which the active and structural lesions at thresholds used in positive MRI definitions were designated by at least 2 and at least 4 of the 6 central raters.
Table 3.
Performance of cut-offs for presence of definite structural lesions typical of axial disease
| Cut-offs for number of SIJ quadrants/halves (any location) with majority (≥4/6) rater agreement for definite structural lesion | Sensitivity (95% CI) | Specificity (95% CI) |
|---|---|---|
| Erosion lesion ≥1 | 97.1 (90.1-99.7) | 91 (85.4-95) |
| Erosion lesion ≥2 | 95.7 (88-99.1) | 96.8 (92.7-99) |
| Erosion, same location on ≥2 consecutive slices | 94.3 (86-98.4) | 98.1 (94.5-99.6) |
| Erosion lesion ≥3 | 91.4 (82.3-96.8) | 98.7 (95.4-99.8) |
| Sclerosis lesion ≥1 | 74.3 (62.4-84) | 95.5 (91-98.2) |
| Sclerosis lesion ≥2 | 62.9 (50.5-74.1) | 98.1 (94.5-99.6) |
| Fat lesion ≥1 | 31.4 (20.9-43.6%) | 96.8 (92.7-99%) |
| Deep fat lesion^ | 2.8 (0.3-9.7) | 100 (97.7-100) |
| Fat lesion ≥2 | 22.9 (13.7-34.4) | 98.7 (95.4-99.8) |
| Fat lesion, same location on ≥3 consecutive slices | 8.6 (3.2-17.7) | 100 (97.7-100) |
| Fat lesion ≥5 | 11.4 (5.1-21.3) | 100 (97.7-100) |
| Backfill lesion ≥1 SIJ half | 24.3 (14.8-36) | 99.4 (96.5-100) |
| Backfill lesion ≥2 SIJ halves | 20 (11.4-31.3) | 100 (97.7-100) |
| Ankylosis lesion ≥1 SIJ half | 2 (0.4-5.7) | 100 (94.9-100) |
| Ankylosis lesion ≥2 SIJ halves | 1.3 (0.2-4.7) | 100 (94.9-100) |
| Any structural lesion | 98.6 (92.3-100) | 89.7 (83.9-94) |
| ANY of the following in ≥2 SIJ quadrants: erosion, sclerosis, fat lesion, or ≥2 SIJ halves: backfill, ankylosis | 98.6 (92.3-100) | 93.6 (88.5-96.9) |
| ≥3 SIJ quadrants with erosion or ≥5 with fat lesions, erosion at the same location for ≥2 consecutive slices, fat lesions at the same location for ≥3 consecutive slices, or the presence of a deep fat lesion* | 97.2 (90.2-99.6) | 96.9 (92.8-99.7) |
| Erosion ≥3 quadrants, OR sclerosis or fat lesion ≥2 SIJ quadrants or backfill or ankylosis ≥2 SIJ halves across all slices | 98.6 (92.3-100.0) | 95.5 (91.0-98.2) |
Sacroiliac joint (SIJ) quadrants = upper ilium, lower ilium, upper sacrum, lower sacrum for left and right SIJ. SIJ halves = upper ilium/sacrum, lower ilium/sacrum. Number of quadrants/halves = total # of quadrants/halves with detectable lesion across all slices of study unless specifically indicated as consecutive.
Criteria adopted for quantitative definition of positive MRI in adults.
Deep fat lesion = >1 cm of depth (10). BME = bone marrow edema; SIJ = sacroiliac joint; CI = confidence interval.
Figure 2. Examples of Positive MRI for Structural Lesions in Juvenile SpA.
On coronal oblique T1 sequences: A) Erosions in 3 SIJ quadrants (arrowheads), sclerosis in 4 quadrants (arrows); B) Backfill (arrows) both upper iliacs sides of the SIJs; c) ankylosis of right and left SIJs (arrows); d) Fat lesion involving the right upper and lower sacrum (arrows).
Table 4.
Frequencies of active and structural lesions at thresholds used in positive MRI definition
| Lesion threshold | Designated by ≥2 central raters N (%) |
Designated by ≥4 central raters N (%) |
|---|---|---|
| Subchondral inflammation/bone marrow edema (≥3 quadrants)* | 78 (32.1) | 70 (28.8) |
| Erosion (≥3 quadrants)^ | 73 (30.0) | 61 (25.1) |
| Fat lesion/ fat metaplasia (≥2 quadrants)^ | 23 (9.5) | 9 (3.7) |
| Fat metaplasia in an erosion cavity (Backfill) (≥2 SIJ halves)^ | 16 (6.6) | 10 (4.1) |
| Sclerosis (≥2 quadrants)^ | 51 (21.0) | 31 (12.8) |
| Ankylosis (≥2 SIJ halves)^ | 3 (1.2) | 2 (0.8) |
Total number of studies evaluated for active lesions* was 239 and structural lesions^ was 226.
Performance of the optimal definitions for positive MRI for inflammatory and structural lesions typical of axial JSpA was tested in a validation cohort of 182 children from 11 countries. Using the central imaging majority assessment with high confidence of definite inflammatory lesion as the reference standard, sensitivity and specificity for BME in ≥3 SIJ quadrants were 96.5% (95% CI: 87.9-99.6) and 94.4% (88.8-97.7%), respectively. Using the central imaging majority assessment with high confidence of definite structural lesion as the reference standard, sensitivity and specificity of erosion in ≥3 SIJ quadrants or any one of the other structural lesions in ≥2 SIJ quadrants or halves were 91.2% (95% CI: 80.7-97.1) and 91.9% (85.6-96.0%), respectively.
DISCUSSION
A robust quantitative definition for a positive MRI of the sacroiliac joints typical of active and structural axial disease in the juvenile SpA population is critical to incorporate into classification criteria that are being developed for this population and potentially for selection of patients for clinical trials. Prior work demonstrates that what constitutes a positive MRI varies widely across pediatric academic centers in the US (4). Using majority imaging expert rater decision as the reference criterion and an international cross-sectional cohort of patients, we determined quantitative SIJ imaging lesion thresholds to define a positive MRI for inflammatory and structural lesions typical of axial JSpA. JAMRIS-defined BME in 3 or more quadrants on an MRI scan across the study was optimal for defining a definite active inflammatory lesion typical of axial disease in juvenile SpA. ASAS-defined erosion in ≥3 quadrants OR JAMRIS-defined sclerosis in ≥2 quadrants OR JAMRIS-defined fat lesion in ≥2 quadrants OR ASAS-defined backfill in ≥2 SIJ halves OR JAMRIS-defined ankylosis in ≥2 SIJ halves was optimal for defining a definite structural lesion typical of axial disease in juvenile SpA. Both definitions had excellent interrater agreement and surpassed the interrater agreement for the ASAS definition of positive MRI. Both definitions were validated in an independent cohort and had similarly high sensitivity and specificity.
Our data-driven definition for active lesions typical of axial SpA are similar to but not identical to the adult definition. The definition chosen for adults was ASAS-defined BME in ≥4 SIJ quadrants at any location or at the same location on ≥3 consecutive slices. Importantly, our central imaging team chose to apply the JAMRIS definition of BME as it accounts for the metaphyseal-equivalent signal that is frequently observed (and misinterpreted) in the juvenile population (3, 13). Further the high threshold of affected quadrants in adults is necessary due to the relatively higher prevalence of BME that meets ASAS criteria in the healthy population. However, rates of false positive BME in children are much lower (1, 4). While the accepted adult threshold of ≥4 quadrants had a high specificity, the sensitivity was lower than for ≥3 quadrants.
The data-driven definition for structural lesions typical of axial SpA in the juvenile population are likely different than those for adults for several reasons. The prevalence of the various lesions and conditions that mimic these lesions differ in the two populations. The ASAS erosion definition includes full-thickness cortical loss and loss of underlying bright T1 marrow signal. In the juvenile population, the bony cortex may not be fully ossified and may not appear as a dark line, while underlying marrow may still be ‘red’ marrow with relatively low T1 signal, leading to confusion between physiologic appearances and erosion. Fat lesions, backfill, and ankylosis are more prevalent in adult cases with longstanding disease than pediatric cases with relatively short disease duration. The relative rarity of these lesions in children translates into a definition of positive MRI that is inclusive of more than 2 involved quadrants of any of these lesions that performs with low sensitivity. The juvenile definition of positive MRI is inclusive of a threshold for sclerosis while the adult definition omits this lesion all together. The exclusion of sclerosis from adult scoring and the adult definition of positive MRI is primarily based on lack of specificity. However, these issues have not been observed in the juvenile population and the juvenile SI joints are also not subject to the confounding factor of degenerative sclerosis. In fact, sclerosis is included in the juvenile adaptation of the SPARCC SIJ structural damage score and the JAMRIS tool (18).
There are few limitations of this study that should be acknowledged. First, there was imaging protocol variability as scans were obtained as part of routine clinical care and not according to a specified protocol. However, included images had the minimum sequences necessary to perform SIJ quadrant-based scoring for active inflammatory (N=17) or structural lesions (N=4) or both (N=222). The lack of an imaging protocol improves the generalizability of the findings because the definitions proposed in this work will be applicable to a broader collection of imaging protocols already in use around the world. Second, this is purely an imaging study, using an expert consensus imaging gold standard, and no direct correlation to clinical findings or outcomes is considered in this study. Deriving a definition of a positive MRI is aimed toward incorporation into classification criteria and not diagnosis. Consequently, the focus is more on specificity than sensitivity. We are not likely to see meaningful correlations with clinical outcomes because the SIJ have very limited mobility. We also already know that there is limited overlap in imaging and clinical parameters of disease activity but that is not an objective of this study. Lastly, since no control subjects were included in the project, the readers were not blinded to underlying clinical diagnosis of SpA but they were blinded to all other clinical details including those pertaining to axial disease. However, there were a sufficient number of normal SIJ MRI scans in the cohort that the lack of control subjects probably did not substantially bias reader interpretation.
In summary, we propose optimal cut-offs to define active inflammatory and structural MRI lesions typical of axial disease in juvenile SpA that have high specificity and sensitivity using central imaging global assessment as the reference standard and were validated in an independent cohort. Since these cut-offs are intended for use as imaging domains in classification criteria and not for diagnosis, we prioritized specificity over sensitivity, choosing relatively high thresholds for active and structural disease that require a substantial burden of imaging abnormalities to meet the definition. In clinical practice, cases of possible SpA where there is high clinical suspicion but only relatively subtle disease at MRI are likely to frequently fall below these thresholds. How these cases are managed is beyond the scope of this paper. These are the first quantitative imaging cut-offs developed for use in juvenile SpA and leverage a large international cohort of children with SpA evaluated for axial disease. These definitions have applicability for inclusion in classification criteria and selection of patients for clinical trials.
Supplementary Material
Significance and Innovations.
It is well established that the maturing sacroiliac joint (SIJ) looks different from the adult SIJ but a similar definition for a positive MRI in the sacroiliac joint in juvenile spondyloarthritis (JSpA) does not exist.
These are the first quantitative imaging cut-offs developed for use in juvenile SpA and leverage a large international cohort of children with SpA evaluated for axial disease.
These definitions have similarities to the adult definitions with important distinctions adapted for the juvenile population.
These definitions have direct applicability as imaging domains in classification criteria and selection of patients for clinical trials.
ACKNOWLEDGEMENTS AND AFFILIATIONS
The authors thank Joel Paschke (CARE Arthritis) for preparing the imaging review modules and the following collaborators for their contributions of patient clinical data and MRI scans: Matthew L. Stoll (University of Alabama-Birmingham), Gerd Horneff (Asklepios Clinic Sankt Augustin), Giulia Armaroli (Asklepios Clinic Sankt Augustin), Ariane Klein (Asklepios Clinic Sankt Augustin), Rebekka Heidebrecht (Asklepios Clinic Sankt Augustin), Hemalatha Srinivasalu (NIH/NIAMS and Children’s National Hospital), Manuk Manukyan (NIH/NIAMS), Judith A. Smith (University of Wisconsin-Madison), Thomas P. Callaci (University of Wisconsin-Madison), and John W. Garrett (University of Wisconsin-Madison).
This project was funded by the National Institutes of Health 1R01AR074098, awarded to Dr. Weiss.
Grant(s)/Financial supporter(s):
Support for the present manuscript was provided by NIH NIAMS 1R01AR074098 and K24AR078950
Competing interests:
PFW - Support for the present manuscript: NIH NIAMS 1R01AR074098 and K24AR078950 (payment to institution); Grants: Patient-Centered Outcomes Research Institute, NIH, Spondylitis Association of America (payment to institution); Royalties/licenses: Up-to-date (<$10K to author); Consulting fees: Site investigator for Pfizer and Abbvie Clinical Trials (Payment to institution), Advisory Board member: Lily, Biogen, Novartis (all <$10K to author), and Consulting fees: Pfizer, Cerecor (payment to institution); Speaking payment or honoraria: 2022 Rheum Now Speaker (<$5K to author) and Spondyloarthritis Research and Treatment Network – honoraria for educational materials (<$5k to author).
RGL - Consulting fees: Calyx, CARE Arthritis, Image Analysis Group; Payment or honoraria: Novartis.
OK - Speaking payments or honoraria: Novartis, Abbvie, Pfizer, Roche, Sanofi, Amgen; Leadership or fiduciary role: Chairman of Turkish Pediatric Association.
WPM - Leadership or fiduciary role: Board of Directors of SPARTAN; Other financial or non-financial interests: Chief Medical Officer, CARE Arthritis Limited.
Footnotes
Contributor Information
Pamela F. Weiss, Department of Pediatrics, Division of Rheumatology and Center for Pediatric Clinical Effectiveness, Children’s Hospital of Philadelphia and Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine, University of Pennsylvania, Philadelphia PA, USA.
Timothy G. Brandon, Department of Pediatrics, Division of Rheumatology and Center for Pediatric Clinical Effectiveness, Children’s Hospital of Philadelphia, Philadelphia, PA, USA.
Robert G. Lambert, Department of Radiology and Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada.
David M. Biko, Department of Radiology, Children’s Hospital of Philadelphia and Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA.
Nancy A. Chauvin, Department of Radiology, Penn State Health Milton S. Hershey Children’s Hospital, Hershey, PA, USA.
Michael L. Francavilla, Department of Radiology, Whiddon College of Medicine, University of South Alabama, Mobile, AL, USA.
Jacob L. Jaremko, Department of Radiology and Diagnostic Imaging, University of Alberta, Edmonton, AB, Canada.
Nele Herregods, Department of Radiology and Nuclear Medicine, Ghent University Hospital, Ghent, Belgium.
Ozgur Kasapcopur, Department of Pediatric Rheumatology, Istanbul University-Cerrahpasa Cerrahpasa Medical School, Istanbul, Turkey.
Mehmet Yildiz, Department of Pediatric Rheumatology, Istanbul University-Cerrahpasa Cerrahpasa Medical School, Istanbul, Turkey.
Alison M. Hendry, General Medicine and Rheumatology, Division of Medicine, Emergency and Integrated Care, Counties Manukau District Health Board, Auckland, New Zealand
Walter P. Maksymowych, Department of Medicine, University of Alberta and CARE Arthritis, Edmonton, AB, Canada.
REFERENCES
- 1.Chauvin NA, Xiao R, Brandon TG, Biko DM, Francavilla M, Khrichenko D, et al. MRI of the Sacroiliac Joint in Healthy Children. AJR Am J Roentgenol. 2019:1–7. [DOI] [PubMed] [Google Scholar]
- 2.Herregods N, Lambert RGW, Schiettecatte E, Dehoorne J, Renson T, Laloo F, et al. Blurring and irregularity of the subchondral cortex in pediatric sacroiliac joints on T1 images: incidence of normal findings that can mimic erosions. Arthritis Care Res (Hoboken). 2021. [DOI] [PubMed] [Google Scholar]
- 3.Herregods N, Jans LBO, Chen M, Paschke J, De Buyser SL, Renson T, et al. Normal subchondral high T2 signal on MRI mimicking sacroiliitis in children: frequency, age distribution, and relationship to skeletal maturity. Eur Radiol. 2021;31(5):3498–507. [DOI] [PubMed] [Google Scholar]
- 4.Weiss PF, Brandon TG, Bohnsack J, Heshin-Bekenstein M, Francavilla ML, Jaremko JL, et al. Variability in Interpretation of Magnetic Resonance Imaging of the Pediatric Sacroiliac Joint. Arthritis Care Res (Hoboken). 2021;73(6):841–8. [DOI] [PubMed] [Google Scholar]
- 5.Rudwaleit M, Jurik AG, Hermann KG, Landewe R, van der Heijde D, Baraliakos X, et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group. Annals of the rheumatic diseases. 2009;68(10):1520–7. [DOI] [PubMed] [Google Scholar]
- 6.Weber U, Lambert RG, Ostergaard M, Hodler J, Pedersen SJ, Maksymowych WP. The diagnostic utility of magnetic resonance imaging in spondylarthritis: an international multicenter evaluation of one hundred eighty-seven subjects. Arthritis Rheum. 2010;62(10):3048–58. [DOI] [PubMed] [Google Scholar]
- 7.de Winter J, de Hooge M, van de Sande M, de Jong H, van Hoeven L, de Koning A, et al. Magnetic Resonance Imaging of the Sacroiliac Joints Indicating Sacroiliitis According to the Assessment of SpondyloArthritis international Society Definition in Healthy Individuals, Runners, and Women With Postpartum Back Pain. Arthritis Rheumatol. 2018;70(7):1042–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Weber U, Jurik AG, Zejden A, Larsen E, Jorgensen SH, Rufibach K, et al. Frequency and Anatomic Distribution of Magnetic Resonance Imaging Features in the Sacroiliac Joints of Young Athletes: Exploring “Background Noise” Toward a Data-Driven Definition of Sacroiliitis in Early Spondyloarthritis. Arthritis Rheumatol. 2018;70(5):736–45. [DOI] [PubMed] [Google Scholar]
- 9.Lambert RG, Bakker PA, van der Heijde D, Weber U, Rudwaleit M, Hermann KG, et al. Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: update by the ASAS MRI working group. Ann Rheum Dis. 2016;75(11):1958–63. [DOI] [PubMed] [Google Scholar]
- 10.Maksymowych WP, Lambert RG, Baraliakos X, Weber U, Machado PM, Pedersen SJ, et al. Data-driven definitions for active and structural MRI lesions in the sacroiliac joint in spondyloarthritis and their predictive utility. Rheumatology (Oxford). 2021;60(10):4778–89. [DOI] [PubMed] [Google Scholar]
- 11.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. Journal of biomedical informatics. 2009;42(2):377–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Elm Ev, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335(7624):806–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Herregods N, Maksymowych WP, Jans L, Otobo TM, Sudol-Szopinska I, Meyers AB, et al. Atlas of MRI findings of sacroiliitis in pediatric sacroiliac joints to accompany the updated preliminary OMERACT pediatric JAMRIS (Juvenile Idiopathic Arthritis MRI Score) scoring system: Part I: Active lesions. Semin Arthritis Rheum. 2021;51(5):1089–98. [DOI] [PubMed] [Google Scholar]
- 14.Herregods N, Maksymowych WP, Jans L, Otobo TM, Sudol-Szopinska I, Meyers AB, et al. Atlas of MRI findings of sacroiliitis in pediatric sacroiliac joints to accompany the updated preliminary OMERACT pediatric JAMRIS (Juvenile Idiopathic Arthritis MRI Score) scoring system: Part II: Structural damage lesions. Semin Arthritis Rheum. 2021;51(5):1099–107. [DOI] [PubMed] [Google Scholar]
- 15.Rudwaleit M, Jurik AG, Hermann KGA, Landewé R, van der Heijde D, Baraliakos X, et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group. Annals of the Rheumatic Diseases. 2009;68(10):1520. [DOI] [PubMed] [Google Scholar]
- 16.Maksymowych WP, Inman RD, Salonen D, Dhillon SS, Williams M, Stone M, et al. Spondyloarthritis research Consortium of Canada magnetic resonance imaging index for assessment of sacroiliac joint inflammation in ankylosing spondylitis. Arthritis Rheum. 2005;53(5):703–9. [DOI] [PubMed] [Google Scholar]
- 17.Maksymowych WP, Wichuk S, Chiowchanwisawakit P, Lambert RG, Pedersen SJ. Development and preliminary validation of the spondyloarthritis research consortium of Canada magnetic resonance imaging sacroiliac joint structural score. J Rheumatol. 2015;42(1):79–86. [DOI] [PubMed] [Google Scholar]
- 18.Weiss PF, Maksymowych WP, Lambert RG, Jaremko JL, Biko DM, Paschke J, et al. Feasibility and Reliability of the Spondyloarthritis Research Consortium of Canada Sacroiliac Joint Structural Score in Children. J Rheumatol. 2018;45(10):1411–7. [DOI] [PubMed] [Google Scholar]
- 19.Weiss P, Brandon T, Aggarwal A, Burgos-Vargas R, Colbert R, Horneff G, et al. Data-driven MRI Definitions for Active and Structural Sacroiliac Joint Lesions in Juvenile Spondyloarthritis Typical of Axial Disease [abstract]. Arthritis Rheumatol. 2021;73(suppl 10). [Google Scholar]
- 20.Weiss PF, Brandon TG, Aggarwal A, Burgos-Vargas R, Colbert RA, Horneff G, et al. Data-Driven Mri Definitions for Active and Structural Sacroiliac Joint Lesions in Juvenile Spondyloarthritis Typical of Axial Disease. Annals of the Rheumatic Diseases. 2022;81(supplement 1). [Google Scholar]
- 21.Rudwaleit M, Jurik AG, Hermann KG, Landewe R, van der Heijde D, Baraliakos X, et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group. Ann Rheum Dis. 2009;68(10):1520–7. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.