You are here

  • Home
  • Archive
  • Volume 76, Issue 10
  • The yield of a positive MRI of the spine as imaging criterion in the ASAS classification criteria for axial spondyloarthritis: results from the SPACE and DESIR cohorts

Article Text

Article menu

Download PDFPDF

Clinical and epidemiological research
Extended report
The yield of a positive MRI of the spine as imaging criterion in the ASAS classification criteria for axial spondyloarthritis: results from the SPACE and DESIR cohorts
  1. Zineb Ez-Zaitouni1,
  2. Pauline AC Bakker1,
  3. Miranda van Lunteren1,
  4. Manouk de Hooge1,
  5. Rosaline van den Berg1,
  6. Monique Reijnierse2,
  7. Karen Minde Fagerli3,
  8. Robert BM Landewé4,
  9. Roberta Ramonda5,
  10. Lennart TH Jacobsson6,
  11. Alain Saraux7,
  12. Gregory Lenczner8,
  13. Antoine Feydy9,
  14. Jean Baptiste Pialat10,
  15. Fabrice Thévenin9,
  16. Floris A van Gaalen1,
  17. Désirée van der Heijde1
  1. 1Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands
  2. 2Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands
  3. 3Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway
  4. 4Amsterdam Rheumatology & Immunology Center (ARC), Amsterdam, The Netherlands
  5. 5Department of Medicine DIMED, University of Padova, Padova, Italy
  6. 6Department of Rheumatology, University ofGöteborg, Göteborg, Sweden
  7. 7Department of Rheumatology, Laboratoire d’Immunothérapie et Pathologies, lymphocytaires B, Labex ‘Immunotherapy, Graft, Oncology’, Université Bretagneoccidentale, Bretagne, France
  8. 8Department of Radiology, Clinique Hartmann, Neuilly SurSeine, France
  9. 9Department of Radiology, Paris Descartes University, Cochin Hospital, Paris, France
  10. 10Radiology Department, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Claude-Bernard University, Pierre-Bénite, Bernard, France
  1. Correspondence to Zineb Ez-Zaitouni, Department of Rheumatology, LeidenUniversity Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands; z.ez-zaitouni{at}lumc.nl

Abstract

Objectives To assess the prevalence of spinal inflammation on MRI in patients with chronic back pain (CBP) of maximally 3 years duration and to evaluate the yield of adding a positive MRI-spine as imaging criterion to the Assessment of SpondyloArthritis international Society (ASAS) classification criteria for axial spondyloarthritis (axSpA).

Methods Baseline imaging of the sacroiliac joints (X-SI), MRI of the sacroiliac joints (MRI-SI) and MRI-spine were scored by ≥2 experienced central readers per modality in the SPondyloArthritis Caught Early (SPACE) and DEvenir des Spondylarthropathies Indifférenciées Récentes (DESIR) cohorts. Inflammation suggestive of axSpA was assessed in the entire spine. A positive MRI-spine was defined by the presence of ≥5 inflammatory lesions. Alternative less strict definitions were also tested.

Results In this study, 541 and 650 patients with CBP from the SPACE and DESIR cohorts were included. Sacroiliitis on X-SI and MRI-SI was found in 40/541 (7%) and 76/541 (14%) patients in SPACE, and in DESIR in 134/650 (21%) and 231/650 (36%) patients, respectively. In SPACE and DESIR, a positive MRI-spine was seen in 4/541 (1%) and 48/650 (7%) patients. Of the patients without sacroiliitis on imaging, 3/447 (1%) (SPACE) and 8/382 (2%) (DESIR) patients had a positive MRI-spine. Adding positive MRI-spine as imaging criterion led to new classification in only one patient in each cohort, as the other patients already fulfilled the clinical arm. Other definitions of a positive MRI-spine yielded similar results.

Conclusion In two cohorts of patients with CBP with a maximum symptom duration of 3 years, a positive MRI-spine was rare in patients without sacroiliitis on MRI-SI and X-SI. Addition of MRI-spine as imaging criterion to the ASAS axSpA criteria had a low yield of newly classified patients and is therefore not recommended.

  • axial spondyloarthritis
  • ankylosing spondylitis
  • magnetic resonance imaging
  • classification criteria

https://doi.org/10.1136/annrheumdis-2017-211486

Statistics from Altmetric.com

    Request Permissions

    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.

    Introduction

    MRI has become of great interest as a diagnostic tool in the evaluation of patients suffering from chronic back pain (CBP) suspected of axial spondyloarthritis (axSpA).1–8 For years, clinicians (and researchers) relied on conventional imaging of the sacroiliac joints (X-SI) to detect abnormalities suggestive of axSpA.9 However, X-SI only captures structural damage, which generally takes months to years to develop, leaving aside that not all patients will develop structural bone damage in the axial skeleton. This jeopardises the early recognition of patients with axSpA. MRI, however, can visualise both inflammation and structural damage, and therefore may help in recognising patients with axSpA who do not (yet) have radiographic sacroiliitis.

    The Assessment of SpondyloArthritis international Society (ASAS) classification criteria for axSpA comprise a criteria set combining the information obtained from patient history taking, physical examination, laboratory testing and imaging (X-SI and MRI of the sacroiliac joints).10 A pivotal aspect of these criteria is the ‘two arms’ concept, commonly referred to as the ‘imaging arm’ and the ‘clinical arm’. According to the ASAS classification criteria, patients with CBP with the onset of back pain at <45 years of age are classified via the imaging arm when (1) radiographic sacroiliitis plus ≥1 spondyloarthritis (SpA)-feature is present or (2) sacroiliitis on MRI plus ≥1 additional SpA-feature is present also defined as ‘positive MRI’. Currently, a positive MRI in the ASAS classification criteria is solely based on the presence of inflammatory lesions in the sacroiliac joints and positive findings are defined according to the ASAS definition: one bone marrow oedema (BME) lesion highly suggestive of axSpA present on ≥2 consecutive slices or ≥2 BME lesions highly suggestive of axSpA on a single slice.11 12

    Several studies have shown that besides inflammation on MRI of the sacroiliac joints (MRI-SI), inflammatory lesions in the spine on MRI (MRI-spine) may also occur.8 13 Remarkably, in a phase III, multicentre, randomized, controlled trial of adalimumab versus placebo (ABILITY-1) spinal inflammation—in the absence of sacroiliitis on MRI—was observed in about half of the patients with longer disease duration and clinically active disease.14 A consensus definition for a positive MRI-spine was developed by the ASAS/Outcome Measures in Rheumatology (OMERACT) MRI working group.8 In this consensus definition a positive MRI-spine is described as the presence of ≥3 inflammatory lesions in the vertebrae, whereas each lesion needs to be present on ≥2 consecutive slices. De Hooge et al recently proposed a cut-off value of ≥5 inflammatory lesions that defines a positive MRI-spine with higher specificity of ≥95% (ie, <5% patients without axSpA with a positive MRI-spine).15

    The main objectives of this study were to evaluate the presence of spinal inflammatory lesions on MRI in patients with a maximum CBP duration of 3 years in two different cohorts, the SPondyloArthritis Caught Early (SPACE) cohort and the DEvenir des Spondylarthropathies Indifférenciées Récentes (DESIR) cohort. In addition, we assessed the added value of spinal inflammatory lesions on MRI, represented in various definitions of a positive MRI-spine, as imaging criterion in the ASAS classification criteria for axSpA.

    Methods

    Cohorts and clinical assessments

    For this analysis, baseline data from the SPACE and the DESIR cohorts were used which have been described in detail before.16 17

    In summary, SPACE is an ongoing observational cohort in which patients with a minimum age of 16 years with short-term CBP (≥3 months, ≤2 years and an onset <45 years) are included. Patients are recruited from multiple Rheumatology centres in Europe: the Netherlands, Norway, Italy and Sweden. The clinical database used for the current study was locked on 30 September 2016.

    DESIR is a longitudinal cohort for which the inclusion period was from December 2007 until April 2010 in 25 centres in France. Patients aged 18–50 years with inflammatory back pain (IBP) according to the Calin18 or Berlin19 criteria, persisting ≥3 months but <3 years, suggestive of axSpA according to the treating rheumatologist, were included. The clinical database used for the current study was locked on 30 April 2015.

    As part of both study protocols and to determine fulfilment of the ASAS-criteria, presence of all SpA-features was assessed. These include human leucocyte antigen B27 (HLA-B27), positive family history of SpA, IBP, psoriasis, peripheral arthritis, dactylitis, heel enthesitis, acute anterior uveitis, inflammatory bowel disease (IBD), good response to non-steroidal anti-inflammatory drugs (NSAIDs), elevated C reactive protein (CRP) and erythrocyte sedimentation rate (ESR) and sacroiliitis on X-SI and MRI-SI, all according to the published definitions.

    In both cohorts, informed consent forms from all study participants as well as approval from all local medical ethical committees were obtained beforehand.

    Imaging

    Detailed descriptions of the applied scoring methods in both cohorts have been published previously. In brief, all available baseline imaging modalities were scored by experienced central readers. In the SPACE cohort, each imaging modality was scored by three central readers. In the DESIR cohort, each imaging modality was scored by two central readers and an adjudicator in case of disagreement. The readers were blinded for all clinical and laboratory data as well as the other imaging modalities, and the various modalities were scored separately.

    Imaging assessments

    Sacroiliac joints were evaluated on pelvic radiographs using the modified New York criteria in which sacroiliitis is defined as grade ≥2 bilaterally or grade 3–4 unilaterally.9 MRI-SI and MRI-spine were performed on a 1T–1.5T scanner. For both modalities, Short Tau Inversion Recovery (TR2500-4000/TE600) and T1-weighted Turbo Spin-Echo (TR500-700/TE10-55) sequences were acquired. MRI-SI were performed in coronal oblique plane and MRI-spine in sagittal plane with a slice thickness of 4 mm. Readers provided an overall judgement on a positive MRI-SI (yes/no) according to the ASAS definition.11 X-SI and MRI-SI were considered positive if ≥2 readers agreed.

    MRI-spine

    In the SPACE cohort, a positive MRI-spine was defined by the presence of ≥5 corner BME lesions highly suggestive of axSpA each visible on ≥2 consecutive slices and if ≥2 readers agreed.15 A positive MRI-spine was also defined by the ASAS consensus definition (≥3 corner BME lesions on ≥2 consecutive slices) and if ≥2 readers agreed.8

    In addition, spinal inflammation suggestive of axSpA was scored according to the Spondyloarthritis Research Consortium of Canada (SPARCC) method. Presence of BME was marked on three consecutive sagittal slices per vertebral unit (VU).20 In case of BME lesions on more than three slices, the three most affected consecutive slices were selected. The 23 VUs each are divided into four quadrants and assessed for the presence (score of 1) or absence (score of 0) of inflammatory lesions (maximum possible score 276). In addition, for each VU, a score of 1 could be assigned to the presence of an ‘intense’ signal (maximum possible score of 69), defined as a signal as intense as a blood vessel and a ‘deep’ lesion (maximum possible score of 69), defined as a homogeneous, unequivocal increase in STIR signal extending at least 1 cm from the vertebral end plate. In total, the maximum possible total SPARCC-score is 414. We tested two additional definitions of a positive MRI-spine using the SPARCC-score (including intensity and depth of a lesion): first, in which a positive MRI-spine was defined by a SPARCC-score ≥5 by≥2 readers and second, in which a positive MRI-spine was defined by a mean SPARCC-score ≥5 of all 3 readers. All four of the described definitions of a positive MRI-spine (ie, ≥5 BME lesions on ≥2 consecutive slices; ≥3 BME lesions on ≥2 consecutive slices (ASAS definition); SPARCC score ≥5 by ≥2 readers and mean SPARCC score ≥5 of 3 readers) were applied to assess the added value of a positive MRI-spine in the ASAS classification criteria.

    In the DESIR cohort, spinal inflammation suggestive of axSpA was scored according to the Berlin method.21 In total, the maximum possible total Berlin-score is 69. Additionally, the 23 VUs were each divided into four quadrants and assessed for the presence (score of 1) or absence (score of 0) of inflammatory lesions. Inflammatory lesions were scored when present on ≥2 consecutive slices. When central readers disagreed on the presence of inflammation, an adjudicator provided scores for all 23 VUs. A positive MRI-spine was defined by the presence of ≥5 corner inflammatory lesions and if ≥2 readers agreed. A positive MRI-spine was also defined by the ASAS consensus definition (ie, ≥3 inflammatory lesions) and when ≥2 readers agreed. Only these two definitions were used to assess the additional value of a positive MRI-spine in the ASAS classification criteria, as the SPARCC method was not applied.

    Data analysis

    Baseline demographic and clinical characteristics are presented using descriptive statistics for both the SPACE and DESIR cohort (reported in this order). Results are presented as mean and SD or numbers and percentages (%). Agreement (Cohen’s kappa, ƙ) between readers regarding a positive MRI-SI and MRI-spine was calculated.

    Stata 14 (StataCorp) was used for data analysis.

    Results

    For 541/639 and 650/708 patients from the SPACE and DESIR cohorts, complete scores of MRI-SI, MRI-spine, X-SI and clinical data were available for analyses. Of these, 34% and 46% were male, mean symptom duration (SD) was 13.0 (7.1) and 18.2 (10.5) months and mean age was 29.1 (8.4) and 33.7 (8.7) years (table 1), respectively.

    View this table:
    Table 1

    Baseline clinical and demographic characteristics of patients with CBP suspected of axSpA in the SPACE and DESIR cohorts

    Inter-reader reliability

    The agreement between the readers in the SPACE and DESIR cohorts regarding a positive MRI-spine according to the ASAS definition was ƙ=0.66 and ƙ=0.58, respectively. Inter-reader agreement for a positive MRI-spine according to the cut-off of ≥5 BME lesions in SPACE was ƙ=0.60 and in DESIR ƙ=0.49. For MRI-SI, the reliability between the readers in the SPACE cohort was ƙ=0.76 and in the DESIR cohort ƙ=0.73.

    In both cohorts, the majority of patients did not show abnormalities on MRI-SI, MRI-spine and X-SI (444/541 (82%) and 374/650 (58%), SPACE and DESIR, respectively) (table 2). In total, 40/541 (7%) and 134/650 (21%) patients in the SPACE and DESIR cohorts had radiographic sacroiliitis (figure 1). Of the remaining patients without radiographic sacroiliitis, 447/501 (89%) and 382/516 (74%) in SPACE and DESIR had no evidence of inflammation on MRI-SI. In both cohorts, an isolated positive MRI-spine (without sacroiliitis on either MRI-SI or X-SI), applying the cut-off of ≥5 BME lesions, was found in 3/447 (1%) and 8/382 (2%) patients (table 2).

    View this table:
    Table 2

    Cross-tabulations of baseline sacroiliac imaging (MRI-SI and X-SI) and MRI-spine of patients with CBP suspected of axSpA in the SPACE (n=541) and DESIR cohorts (n=650)

    View this table:
    Figure 1
    Figure 1

    ASAS classification of patients with CBP with negative MRI-SI and positive MRI-spine defined by ≥5 inflammatory lesions and the effect of adding positive MRI-spine as an imaging criterion to the ASAS axSpA criteria on classification of patients. (A) Additional classification of patients with CBP in the SPACE cohort. (B) Additional classification of patients with CBP in the DESIR cohort. ASAS, Assessment of SpondyloArthritis international Society; axSpA, axial spondyloarthritis; CBP, chronic back pain; SPACE, SPondyloArthritis Caught Early.

    Two of the three patients in the SPACE cohort with a positive MRI-spine (definition of ≥5 BME lesions) but without sacroiliitis on MRI-SI and X-SI, already fulfilled the clinical arm of the ASAS-criteria, were male and their mean (SD) number of SpA-features was 3.5 (0.7) (table 3). When hypothetically adding a positive MRI-spine as imaging criterion to the ASAS-criteria for axSpA, the remaining patient could be additionally classified via the imaging arm. Therefore, 447 MRIs of the spine have to be performed to additionally classify this single patient. This patient was a male with one SpA-feature (ie, good response to NSAIDs).

    View this table:
    Table 3

    Baseline clinical characteristics and disease activity of patients with CBP additionally classified according to the ASAS axSpA criteria by baseline presence of sacroiliitis on MRI and spinal inflammation defined by ≥5 inflammatory lesions

    View this table:

    In the DESIR cohort, eight patients had a positive MRI-spine (definition of ≥5 BME lesions) without having any signs of inflammation or structural damage suggestive of axSpA on MRI-SI and X-SI (figure 1). As patients within the DESIR cohort are included until the age of 50 and the ASAS-criteria are, in principle, to be applied to patients≤45 years of age with CBP, we report this separately. In total, seven out of eight patients already fulfilled the clinical arm of the ASAS-criteria: 6/8 patients≤45 years and 1/8>45 years (table 3). Of these seven patients, one was male and the mean (SD) number of SpA-features was 3.7 (2.0). When hypothetically adding a positive MRI-spine as imaging criterion to the ASAS-criteria for axSpA, the remaining patient could be additionally classified via the imaging arm. Therefore, 382 MRI of the spine have to be performed to additionally classify this single patient. This was a HLA-B27 negative male patient with the following six SpA-features: positive family history of SpA, IBP, good response to NSAIDs, heel enthesitis, IBD and peripheral arthritis.

    Alternative definitions of a positive MRI-spine

    SPARCC-score

    Two other definitions of a positive MRI-spine were tested in the SPACE cohort using the SPARCC-score (table 4). When a positive MRI-spine was defined as a SPARCC-score of ≥5 by ≥2 central readers, a total of 21/447 (5%) patients without sacroiliitis on MRI-SI and X-SI had a positive MRI-spine (see online supplementary figure S1). Eight patients (38%) already fulfilled the clinical arm and when adding MRI-spine to the ASAS-criteria, 13/21 (62%) patients could be additionally classified via the imaging arm. Of these 13 patients almost half had ≥2 SpA-features. Most frequently reported SpA-features were good response to NSAIDs, positive family history of SpA and IBP (data not shown). When a positive MRI-spine was defined as a mean SPARCC-score ≥5 of all readers, a total of 13/447 (3%) patients had a positive MRI-spine (see online supplementary figure S2). Of these, 7/13 (54%) already fulfilled the clinical arm and the remaining six (46%) patients could be additionally classified via the imaging arm when adding positive MRI-spine to the ASAS-criteria. Half of these patients had ≥2 SpA-features and positive family history of SpA and IBP were among the frequent features (data not shown).

    Supplementary Material

    Supplementary Figure 1

    Supplementary Material

    Supplementary Figure 2
    View this table:
    Table 4

    Patients additionally classified through positive MRI-spine according to four different definitions and diagnosis of axSpA in the SPACE and DESIR cohorts

    View this table:

    ASAS consensus definition

    According to the ASAS consensus definition (≥3 spinal inflammatory lesions on ≥2 consecutive slices), 5/447 (1%) patients from the SPACE cohort and 25/382 (7%) patients from the DESIR cohort had a positive MRI-spine without sacroiliitis on MRI-SI and X-SI (table 4). Of these, 3/5 (60%) patients and 18/25 (72%) patients (16 patients≤45 years and two patients>45 years) already fulfilled the clinical arm (see online supplementary figure S3). The addition of a positive MRI-spine as a criterion in the ASAS-criteria would have resulted in the additional classification of 2/5 (40%) and 7/25 (28%) patients (six patients≤45 years and one patient >45 years) for SPACE and DESIR, respectively.

    Supplementary Material

    Supplementary Figure 3

    Discussion

    In patients with CBP of a maximum duration of 3 years and suspected of axSpA from two different cohorts (SPACE and DESIR), the overall prevalence of spinal inflammation is low. In addition, spinal inflammation in the absence of sacroiliitis on MRI-SI and X-SI occurred in a very small proportion of patients. Consequently, adding a positive MRI-spine (represented in various definitions) as an imaging criterion to the ASAS-criteria resulted in a very low percentage of newly classified patients. Considering the number of MRI-spine needed to additionally classify a few patients, the longer scanning time for the patient and higher costs, we conclude that the yield of adding MRI-spine to the ASAS-criteria is unacceptably low in relation to the number of MRI-spine needed to be performed in patients with early disease.

    Multiple studies assessed the presence of spinal inflammation in different patient groups (eg, non-radiographic axSpA); however, to the best of our knowledge, this is the first study to report on the value of a positive MRI-spine in classifying patients according to the ASAS-criteria using different definitions in patients with recent onset CBP. Our findings are in line with a recent study by Weber et al in which they state that the combination of MRI-spine and MRI-SI has little incremental value compared with MRI-SI alone.22 Using global assessment in patients with non-radiographic axSpA and normal MRI-SI, approximately 20% of MRI-spine were judged as being consistent with axSpA. However, reading of MRI-spine in non-specific back pain patients and healthy controls yielded similar percentages.

    In the SPACE and DESIR cohorts, the majority of patients with only spinal inflammation (≥5 BME lesions and normal MRI-SI and X-SI) already fulfilled the clinical arm of the ASAS-criteria (2/3 and 6/7, respectively). The remaining patients, hypothetically classified via the imaging arm solely because of a positive MRI-spine, can either be misclassified or truly ‘missed’ cases.

    We did not formally investigate the role of MRI-spine in diagnosis making in routine clinical practice and this may still be considered as an imaging tool in the differential diagnosis of patients with CBP or in the confirmation of the diagnosis in specific patients.

    This study has several strengths. Two independent early axSpA cohorts were examined in which patients with CBP have been assessed according to a similar standardised protocol. Both cohorts were rather complete in terms of images being present and scored by multiple readers and the findings in both cohorts were similar providing credit to the robustness of our data. Furthermore, the observation that in both cohorts several definitions of a positive MRI-spine in the ASAS-criteria for axSpA yielded comparable misclassifications strengthens the validity of the findings.

    It can be seen as a limitation that we only used patients with short symptom duration. Consequently, the results can only be extrapolated to this group of patients. However, we feel that this is the most applicable group to test the usefulness of MRI-spine as part of the ASAS classification criteria as abnormalities on imaging of the sacroiliac joints are more likely to become positive with longer symptom duration. Moreover, our data are very much in line with the data from Weber et al in a study with more established disease.

    In summary, in both the SPACE and the DESIR cohorts, a positive MRI-spine in patients with CBP suspected of early axSpA was infrequent. Furthermore, spinal inflammation in the absence of sacroiliitis was observed in very few patients. In this early disease stage, the addition of a positive MRI-spine as imaging criterion to the ASAS classification criteria for axSpA yielded a very low number of newly classified patients in both cohorts. Therefore, the use of MRI-spine is not recommended in the classification of patients with CBP suspected of axSpA.

    Acknowledgments

    The authors thank the different regional participating centres: Pr M Dougados (Paris—Cochin B), Pr A Kahan (Paris—Cochin A), Pr O Meyer, Pr P Dieudé (Paris—Bichat), Pr P Bourgeois, Pr L.Gossec (Paris—La Pitié Salpetrière), Pr F Berenbaum (Paris—Saint Antoine), Pr P Claudepierre (Créteil), Pr M Breban (Boulogne Billancourt), Dr B Saint-Marcoux (Aulnay-sous-Bois), Pr P Goupille (Tours), Pr J-F Maillefert (Dijon), Dr X Puéchal, Dr E Dernis (Le Mans), Pr D Wendling (Besançon), Pr B Combe (Montpellier), Pr L Euller-Ziegler (Nice), Pr P Orcel, Dr P Richette (Paris - Lariboisière), Pr P Lafforgue (Marseille), Dr P Boumier (Amiens), Pr M Soubrier (Clermont-Ferrand), Dr N Mehsen (Bordeaux), Pr D Loeuille (Nancy), Pr R-M Flipo (Lille), Pr A Saraux (Brest), Dr S Pavy (Le Kremlin Bicêtre), Pr A Cantagrel (Toulouse), Pr O Vittecoq (Rouen). The authors also thank URC-CIC Paris Centre for the coordination and monitoring of the study.

    References

    1. ↵
      1. Bollow M,
      2. Enzweiler C,
      3. Taupitz M, et al
      . Use of contrast enhanced magnetic resonance imaging to detect spinal inflammation in patients with spondyloarthritides. Clin Exp Rheumatol 2002;20:S167–74.
      OpenUrlPubMedWeb of Science
    2. ↵
      1. Braun J,
      2. Baraliakos X,
      3. Golder W, et al
      . Analysing chronic spinal changes in ankylosing spondylitis: a systematic comparison of conventional x rays with magnetic resonance imaging using established and new scoring systems. Ann Rheum Dis 2004;63:1046–55.doi:10.1136/ard.2003.019968
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Rudwaleit M,
      2. van der Heijde D,
      3. Khan MA, et al
      . How to diagnose axial spondyloarthritis early. Ann Rheum Dis 2004;63:535–43.doi:10.1136/ard.2003.011247
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Hermann KG,
      2. Althoff CE,
      3. Schneider U, et al
      . Spinal changes in patients with spondyloarthritis: comparison of MR imaging and radiographic appearances. Radiographics 2005;25:559–69.doi:10.1148/rg.253045117
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Landewé RB,
      2. Hermann KG,
      3. van der Heijde DM, et al
      . Scoring sacroiliac joints by magnetic resonance imaging. A multiple-reader reliability experiment. J Rheumatol 2005;32:2050–5.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Bennett AN,
      2. Rehman A,
      3. Hensor EM, et al
      . Evaluation of the diagnostic utility of spinal magnetic resonance imaging in axial spondylarthritis. Arthritis Rheum 2009;60:1331–41.doi:10.1002/art.24493
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Weber U,
      2. Lambert RG,
      3. Østergaard M, et al
      . The diagnostic utility of magnetic resonance imaging in spondylarthritis: an international multicenter evaluation of one hundred eighty-seven subjects. Arthritis Rheum 2010;62:3048–58.doi:10.1002/art.27571
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Hermann KG,
      2. Baraliakos X,
      3. van der Heijde DM, et al
      . Assessment in SpondyloArthritis international Society (ASAS). Descriptions of spinal MRI lesions and definition of a positive MRI of the spine in axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI study group. Ann Rheum Dis 2012;71:1278–88.doi:10.1136/ard.2011.150680
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. van der Linden S,
      2. Valkenburg HA,
      3. Cats A
      . Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum 1984;27:361–8.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Rudwaleit M,
      2. Landewé R,
      3. van der Heijde D, et al
      . The development of Assessment of SpondyloArthritis international Society classification criteria for axial spondyloarthritis (part I): classification of paper patients by expert opinion including uncertainty appraisal. Ann Rheum Dis 2009;68:770–6.doi:10.1136/ard.2009.108217
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Lambert RG,
      2. Bakker PA,
      3. van der Heijde D, et al
      . Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: update by the ASAS MRI working group. Ann Rheum Dis 2016;75:1958–63.doi:10.1136/annrheumdis-2015-208642
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Rudwaleit M,
      2. Jurik AG,
      3. Hermann KG, 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;68:1520–7.doi:10.1136/ard.2009.110767
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Baraliakos X,
      2. Landewé R,
      3. Hermann KG, et al
      . Inflammation in ankylosing spondylitis: a systematic description of the extent and frequency of acute spinal changes using magnetic resonance imaging. Ann Rheum Dis 2005;64:730–4.doi:10.1136/ard.2004.029298
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. van der Heijde D,
      2. Sieper J,
      3. Maksymowych WP, et al
      . Spinal inflammation in the absence of sacroiliac joint inflammation on magnetic resonance imaging in patients with active nonradiographic axial spondyloarthritis. Arthritis Rheumatol 2014;66:667–73.doi:10.1002/art.38283
      OpenUrl
    15. ↵
      1. de Hooge M,
      2. van den Berg R,
      3. Navarro-Compán V, et al
      . Patients with chronic back pain of short duration from the SPACE cohort: which MRI structural lesions in the sacroiliac joints and inflammatory and structural lesions in the spine are most specific for axial spondyloarthritis? Ann Rheum Dis 2016;75:1308–14.doi:10.1136/annrheumdis-2015-207823
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Dougados M,
      2. d’Agostino MA,
      3. Benessiano J, et al
      . The DESIR cohort: a 10-year follow-up of early inflammatory back pain in France: study design and baseline characteristics of the 708 recruited patients. Joint Bone Spine 2011;78:598–603.doi:10.1016/j.jbspin.2011.01.013
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. van den Berg R,
      2. de Hooge M,
      3. van Gaalen F, et al
      . Percentage of patients with spondyloarthritis in patients referred because of chronic back pain and performance of classification criteria: experience from the Spondyloarthritis Caught Early (SPACE) cohort. Rheumatology 2013;52:1492–9.doi:10.1093/rheumatology/ket164
      OpenUrlCrossRefPubMed
    18. ↵
      1. Calin A,
      2. Porta J,
      3. Fries JF, et al
      . Clinical history as a screening test for ankylosing spondylitis. JAMA 1977;237:2613–4.doi:10.1001/jama.1977.03270510035017
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Rudwaleit M,
      2. Metter A,
      3. Listing J, et al
      . Inflammatory back pain in ankylosing spondylitis: a reassessment of the clinical history for application as classification and diagnostic criteria. Arthritis Rheum 2006;54:569–78.doi:10.1002/art.21619
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Maksymowych WP,
      2. Inman RD,
      3. Salonen D, et al
      . Spondyloarthritis Research Consortium of Canada magnetic resonance imaging index for assessment of spinal inflammation in ankylosing spondylitis. Arthritis Rheum 2005;53:502–9.doi:10.1002/art.21337
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Braun J,
      2. Baraliakos X,
      3. Golder W, et al
      . Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and after successful therapy with infliximab: evaluation of a new scoring system. Arthritis Rheum 2003;48:1126–36.doi:10.1002/art.10883
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Weber U,
      2. Zubler V,
      3. Zhao Z, et al
      . Does spinal MRI add incremental diagnostic value to MRI of the sacroiliac joints alone in patients with non-radiographic axial spondyloarthritis? Ann Rheum Dis 2015;74:985–92.doi:10.1136/annrheumdis-2013-203887
      OpenUrlAbstract/FREE Full Text

    Footnotes

    • Contributors ZEZ designed the study, performed the statistical analyses, interpreted the data, drafted and revised the manuscript. DvdH and FvG designed the study, interpreted the data and revised the manuscript. All authors contributed to the acquisition of data, revised the manuscript and approved the final manuscript.

    • Funding The DESIR cohort was sponsored by the Département de la Recherche Clinique et du Développement de l’Assistance Publique–Hôpitaux de Paris. This study is conducted under the umbrella of the French Society of Rheumatology and INSERM (Institut National de la Santé et de la Recherche Médicale). The database management is performed within the department of epidemiology and biostatistics (Professor Paul Landais, D.I.M., Nîmes, France). An unrestricted grant from Pfizer was allocated for the 10 years of the follow-up of the recruited patients.

    • Competing interests None declared.

    • Patient consent Obtained.

    • Ethics approval All local medical ethical committees.

    • Provenance and peer review Not commissioned; externally peer reviewed.