Search

Background:Magnetic resonance imaging (MRI) offers a non-invasive and objective method of early diagnosis and classification, monitoring disease burden and treatment response for patients (pts) with axial spondyloarthritis (axSpA) including ankylosing spondylitis (AS).1Numerous scoring schemes such as the AS Spine MRI Activity (ASspiMRIa) score are available for the quantitative assessment of MRI, but are subject to intra- and inter-rater variability, labor intensive and costly. Nevertheless, quantification of MRI changes has become an important tool to demonstrate treatment success of biologic drugs in axSpA.Objectives:To evaluate the performance of machine learning (ML) based software for automated Berlin grading of spinal MRI bone marrow oedema in pts with AS and compare with expert scoring.Methods:Fully automated ML software (Figure) was developed to detect and label 23 vertebrae, define vertebral units (VU) as per the Berlin modification of the ASspiMRIa score, and score each VU as either 0 (score of 0) or 1 (score of 1, 2 or 3). The ML algorithm was based on the previously developed SpineNet software.2Analysis included 108 pts from the secukinumab MEASURE 1 study3, in which imaging was done using T1 and STIR sagittal MRI at baseline and Weeks 16, 52, 104, 156 and 208. Two expert readers, blinded to treatment and visit, evaluated all images by ASspiMRIa score. The scores from Reader 2 (R2) were binned into two groups: 0 vs 1, 2, or 3. As a result of multiple pt time points and expert reading sessions, the complete dataset comprised of 10,988 VU. Ten-way cross-validation at per-VU was used to train and validate the ML software. The dataset was split into 10 randomly selected subsets, ensuring that each pt appears in only one subset, after which 8 subsets were used for training the ML software, 1 was used to check for correct training and 1 was used for validation. The process was repeated ten times such that all 10 subsets were used for validation. Accuracy weighted for the frequency of each category, sensitivity and specificity were calculated using scores from R2 as reference. Intra-reader accuracy was also calculated.Results:Accuracy of the software in relation to expert reader scores was 67% with a sensitivity of 0.63 and specificity of 0.70. The intra-reader accuracy was 71% and 77% for R1 and R2, respectively. Individual VU scoring of the Software vs. R2 are presented in the Table as a confusion matrix.Conclusion:Automated scoring of MR images in AS pts provided moderate agreement to that of expert reader-based assessments. ML software has potential to provide an automated guided-reading approach to scoring MR images, which may enable further clinical insights.References:[1]Lukas C, et al. J Rheumatol. 2007;34:862-70.[2]Jamaludin A, et al. Eur Spine J. 2017;26:1374-83.[3]Baeten D, et al. N Engl J Med. 2015;373,2534-48.Figure.Processing pipeline of automated Berlin scoring softwareTable.Confusion matrix between the software and R2SoftwareScore = 0SoftwareScore = 1, 2 or 3Total VU scoredR2 Score = 07199 (70%)3068 (30%)10,267R2 Score = 1, 2 or 3251 (35%)475 (65%)7267,4503,54310,993Percentages calculated as a fraction over the total in each row. Overall accuracy is the average of the highlighted percentages.Disclosure of Interests:Amir Jamaludin: None declared, Rhydian Windsor: None declared, Sarim Ather: None declared, Timor Kadir: None declared, Andrew Zisserman: None declared, Juergen Braun Grant/research support from: Abbvie (Abbott), Amgen, BMS, Boehringer, Celgene, Celltrion, Centocor, Chugai, Eli Lilly and Company, Medac, MSD (Schering Plough), Mundipharma, Novartis, Pfizer (Wyeth), Roche, Sanofi- Aventis, and UCB Pharma, Consultant of: Abbvie (Abbott), Amgen, BMS, Boehringer, Celgene, Celltrion, Centocor, Chugai, EBEWE Pharma, Eli Lilly and Company, Medac, MSD (Schering-Plough), Mundipharma, Novartis, Pfizer (Wyeth), Roche, Sanofi-Aventis, and UCB Pharma, Speakers bureau: Abbvie (Abbott), Amgen, BMS, Boehringer, Celgene, Celltrion, Centocor, Chugai, EBEWE Pharma, Eli Lilly and Company, Medac, MSD (Schering-Plough), Mundipharma, Novartis, Pfizer (Wyeth), Roche, Sanofi-Aventis, and UCB Pharma, Lianne S. Gensler Grant/research support from: Pfizer, Novartis, UCB, Consultant of: AbbVie, Eli Lilly, GSK, Novartis, UCB, Pedro Machado Consultant of: Abbvie, Celgene, Janssen, Lilly, MSD, BMS, Novartis, Pfizer, Roche and UCB, Speakers bureau: AbbVie, Centocor, Eli Lilly, Janssen, MSD, Novartis, Pfizer and UCB Pharma, Mikkel Ǿstergaard Grant/research support from: AbbVie, Bristol-Myers Squibb, Celgene, Merck, and Novartis, Consultant of: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Speakers bureau: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Denis Poddubnyy Grant/research support from: AbbVie, MSD, Novartis, and Pfizer, Consultant of: AbbVie, Bristol-Myers Squibb, Eli Lilly, MSD, Novartis, Pfizer, Roche, UCB, Speakers bureau: AbbVie, Bristol-Myers Squibb, Eli Lilly, MSD, Novartis, Pfizer, Roche, UCB, Thibaud Coroller Shareholder of: Novartis, Employee of: Novartis, Brian Porter Shareholder of: Novartis, Employee of: Novartis, Shephard Mpofu Shareholder of: Novartis, Employee of: Novartis, Aimee Readie Shareholder of: Novartis, Employee of: Novartis

Despite evidence to the contrary, the public, and clinicians, often believe that low back pain (LBP) diagnosis and management are enhanced through the use of imaging such as MRI and CT. In the USA alone, nearly $6 billion is spent on spinal MRI scans annually, stretching limited resources. We have previously discussed1 how imaging rates for LBP may be safely reduced and aligned with guidelines,2 using methods such as clinician decision support to ‘shape the path’.3 Nevertheless, reducing imaging rates for LBP is difficult. There may be additional value in gaining a better understanding of the decision-making processes, and motivations, of referring clinicians. There is considerable variation in imaging referral patterns between clinicians, with variation the greatest for LBP presentations not involving trauma or radiculopathy (the majority of people with LBP). This suggests clinician factors, in addition to system factors, contribute to the variation. Many clinicians are aware of, and broadly agree with, recommended clinical …

Background Spinal MRI lesions suggestive of axial Spondyloarthritis (axSpA) are not included in the ASAS definition of a positive MRI, but do occur in the absence of affected sacroiliac joints (SIJ). It is unknown how often this happens and if it is useful to perform a MRI of the spine in patients (pts) with negative MRI-SIJ Objectives To investigate the prevalence of a positive MRI-spine in pts with short symptom duration and a negative MRI-SIJ Methods Pts aged 18-50 with inflammatory back pain (IBP) (≥3 months, ≤3 years) from 25 participating centers in France were included in the DESIR-cohort (n=708). All available baseline MRIs of the spine were independently scored by 2 well-calibrated central readers who were blind to any other data. MRIs-SI were scored according to the ASAS definition 1 (lesions highly suggestive of sacroiliitis plus ≥1 lesion on ≥2 consecutive slices or >1 lesion on 1 slice). Inflammatory lesions on MRI-spine suggestive of spondylitis were scored when visible on ≥2 consecutive slices and according to the ASAS consensus definition 2 (≥3 lesions). In case of disagreement, an experienced radiologist served as adjudicator. MRI was considered positive if 2/3 readers agreed Results All pts with MRI-spine and MRI-SIJ (n=650) were included in the analyses. There were 231 pts (35,5%) with a positive MRI-SIJ and 102 pts (15.7%) with a positive MRI-spine; 67 pts (10.3%) were positive for both MRI-SIJ and MRI-spine, 384 (59.1%) were negative for both; and 35 pts (5.4%) had a positive MRI-spine but a negative MRI-SI. Thirty of these were 45 years at symptom onset fulfilled the mNY criteria. Overall, only 25 pts (3.8%) had a pos MRI-spine without sacroiliitis on MRI or radiographs Conclusions In 3.8% of IBP pts aged 18-50 ≥3 spinal inflammatory lesions suggestive of axSpA are found in absence of sacroiliitis on MRI or radiograph. Therefore the yield of including MRI-spine as additional imaging criterion in the ASAS axSpA classification criteria is considered unacceptably low References Rudwaleit, ARD 2009;68:1520-7. Hermann, ARD 2012;71:1278-88. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.4054

Background In clinical practice radiologists and rheumatologists assess whether lesions compatible with axial SpondyloArthritis (axSpA) are present on spinal MRI Objectives To compare the results of local readings (LocR) to centralized reading (CentR) as external standard of BME and structural lesions on MRI-spine, in patients (pts) with inflammatory back pain (IBP) Methods Pts aged 18-50 with recent IBP (≥3 months, ≤3 years) from 25 participating centers in France were included in the DESIR-cohort (n=708). All available baseline MRIs-spine were scored on BME and structural lesions as present, absent or doubtful by the local radiologist/rheumatologist who might have access to clinical data. In addition, 2 well-calibrated centralized readers independently scored the same MRIs for BME lesions and structural lesions (fatty lesions, erosions and (bridging) syndesmophytes). In case the centralized readers disagreed, an experienced radiologist served as adjudicator. Agreement between CentR and LocR was calculated excluding the cases assessed as doubtful by LocR (kappa κ) Results BME/structural lesions were in 492/492 pts scored by a radiologist, 206/205 by a rheumatologist and in 32/32 pts by both. The k agreement between LocR and CentR was 0.27 for BME lesions and 0.13 for structural lesions. For radiologists, κ=0.36 for BME, and κ=0.15 for structural lesions. For rheumatologists κ=0.006 for BME and κ=0.12 for structural lesions. Overall, local specialists are highly overrating positive findings: 42.3% and 85.7% of the positive MRIs for BME are scored negative by the central read (radiologists and rheumatologist respectively). Similarly findings for structural lesions: 48.4% and 70% of MRIs positive for structural lesions are scored normal by central reading Conclusions Both local radiologists, but especially rheumatologists overrate the presence of BME lesions and structural lesions on MRI of the spine compared to trained central readers. These results do not even take doubtful cases into account Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.5456

Background A recent consensus statement based on a systematic literature review by the Assessment of SpondyloArthritis International Society suggested the presence of ≥3 corner inflammatory lesions (CIL) or of several corner fat lesions (CFL) as candidate definitions for a positive MRI of the spine in axial spondyloarthritis (SpA) [1]. Objectives To determine data-driven cut-off values for spinal CIL and CFL yielding a specificity ≥90% and to evaluate their diagnostic utility in non-radiographic axial SpA (nr-axSpA) and ankylosing spondylitis (AS). Methods The study sample comprised 2 independent cohorts A/B of 130 consecutive patients with back pain ≤50 years newly referred to 2 university clinics, and 20 healthy controls, in whom MRI scans of the entire spine were available. Patients were classified according to clinical examination and pelvic radiography as having nr-axSpA (n=50), AS (n=33), or non-specific back pain (NSBP; n=47). Spinal MRI were assessed by 4 blinded readers according to the standardized CanDen modules. Readers recorded bone marrow edema and fat infiltration in the central and lateral/posterior compartment of all 23 discovertebral units. We calculated cut-off values for CIL and CFL to obtain ≥90% specificity and the corresponding area under the curve (AUC) with confidence interval (CI). Finally, we tested the diagnostic utility (mean sensitivity/specificity of 4 readers) of cut-off values for spinal MRI as proposed in the literature (≥3 CIL [1] and ≥5 CFL [2]) for nr-axSpA and AS patients in both cohorts. Results For cohorts A/B, the rounded lesion cut-offs to obtain ≥90% specificity were 3/2 CIL and 7/10 CFL, respectively. The corresponding AUC for CIL were 0.69 (CI 0.49-0.84) and 0.69 (CI 0.47-0.85) in the 2 cohorts, and for CFL 0.60 (CI 0.43-0.75) and 0.71 (CI 0.56-0.82), respectively. The diagnostic utility of the spinal thresholds of ≥3 CIL and of ≥5 CFL was low in both cohorts when comparing nr-ax SpA versus NSBP. Diagnostic utility of 2 candidate definitions of a positive MRI of the spine in cohorts A/B Conclusions In this controlled study, the definitions of a positive spinal MRI proposed in a recent consensus statement showed low diagnostic utility in nr-axSpA. While a cut-off of ≥2/≥3 CIL for a positive MRI was optimal, the threshold for CFL was as high as 10. References Hermann KG et al. ARD 2012;71:1278. [2] Bennett A et al. ARD 2010;69:891. Disclosure of Interest None Declared

Background The definition of a positive MRI as major criterion in the Assessment of SpondyloArthritis classification criteria for axial spondyloarthritis (SpA) is based on MRI of the sacroiliac joints (SIJ) alone. It is not known whether additional MRI of the spine may enhance diagnostic certainty over and above SIJ MRI alone. Objectives To assess the incremental diagnostic value of spinal MRI evaluated both separately from and combined with SIJ MRI in non-radiographic SpA (nr-axSpA) compared to SIJ MRI alone. Methods The study sample comprised 2 independent cohorts A/B of 130 consecutive patients with back pain ≤50 years newly referred to 2 university clinics, and 20 healthy controls (HC), in whom both SIJ and spinal MRI were available. Patients were classified according to clinical examination and pelvic radiography as having nr-axSpA (n=50), ankylosing spondylitis (AS; n=33), or mechanical back pain (MBP; n=47). SIJ and spinal MRI were assessed by 3 blinded readers according to standardized modules. Readers recorded presence/absence of SpA and their level of confidence in this conclusion by global evaluation of the MRI scans (T1SE and STIR sequences) on a 0-10 scale (0 = definitely not SpA; 10 = definite SpA). SIJ alone and spinal MRI alone were read independently 6 months apart, with another interval of 1-12 months to the combined assessment of both SIJ and spinal MRI. We analysed differences between SIJ alone versus spinal MRI alone, and SIJ alone versus combined read of SIJ and spinal MRI. This was done descriptively by the percentage (number) of subjects according to the mean of 3 readers for SIJ MRI and spinal MRI separately, and for combined MRI. Results For cohorts A/B, 15.8%/28.0% of nr-axSpA patients considered having a negative SIJ MRI were re-classified as being positive for SpA by global evaluation of combined SIJ and spinal scans. However, 28.6%/11.1% of MBP controls and 18.3% of healthy volunteers, who showed a negative SIJ MRI, were falsely re-classified as having SpA by combined SIJ and spinal MRI. Low confidence (5-7) in a diagnosis of SpA by global evaluation of SIJ MRI increased to high confidence (8-10) by combined MRI in 3.5% and 8.6% of nr-axSpA patients in the 2 cohorts. Percentage of subjects (mean of 3 readers) Conclusions Combined spinal and SIJ MRI added little incremental value compared to SIJ MRI alone for diagnosing nr-axSpA patients and enhancing confidence in this diagnosis. Disclosure of Interest None Declared

Longitudinally extensive transverse myelitis refers to inflammation of the spinal cord causing T2 hyperintensity on spinal MRI that extends over three or more vertebral segments. Although commonly associated with neuromyelitis optica, there are many other causes that should be considered. These include other inflammatory monophasic myelitises, infection, and mimics such as malignancy and spinal cord infarction. We review the diverse causes of longitudinally extensive transverse myelitis, with illustrative cases from the John Radcliffe Hospital, Oxford. We also summarise laboratory and radiological features that can help to differentiate these causes.