Article Text

FRI0153 Validation of the New Concept of Backfill on Mri: A Distinct Reparative Tissue That Follows Resolution of Inflammation at Sites of Sacroiliac Joint Erosion in Axial Spondyloarthritis
  1. W.P. Maksymowych1,
  2. S.J. Pedersen2,
  3. S. Wichuk1,
  4. P. Chiowchanwisawakit3,
  5. M. Østergaard2,
  6. U. Weber1,
  7. R.G. Lambert4
  1. 1Medicine, University of Alberta, Edmonton, Canada
  2. 2Copenhagen Center for Arthritis Research, University of Copenhagen, Copenhagen, Denmark
  3. 3Medicine, Mahidol University, Bangkok, Thailand
  4. 4Radiology, University of Alberta, Edmonton, Canada


Background We have hypothesized that resolution of inflammatory lesions in erosions of sacroiliac joints in patients with SpA is followed by development of a new tissue, which on T1W MRI has high signal intensity resembling the fat metaplasia in bone marrow seen in SpA. We have called this type of fat lesion “Backfill' due to its appearance in the excavated area caused by erosion.

Objectives 1. To demonstrate that Backfill can be reliably detected on T1W MRI. 2. To demonstrate that resolution of inflammation and reduction of erosion are both independently associated with development of Backfill using longitudinal data.

Methods We adopted standardized definitions for structural lesions: Backfill is defined as complete loss of iliac or sacral cortical bone at its anticipated location and increased signal on T1WSE that is demarcated from adjacent normal marrow by irregular dark signal reflecting sclerosis. Backfill, erosion, and fat metaplasia were scored dichotomously (present/absent) on 5 consecutive coronal slices anteriorly through the cartilaginous portion of the joint. Four readers assessed baseline and 2 year scans from 20 patients (exercise 1) and then 45 patients after calibration (exercise 2). Inter-observer reliability was assessed by intra-class correlation coefficient (ICC3,1). Two readers independently scored 147 pairs of scans (baseline, 2 years) from a prospective cohort of patients with SpA on either NSAID (n=69) or anti-TNF (n=78) therapies. SIJ inflammation was scored using the SPARCC MRI SIJ method. Correlations between MRI features were assessed by Pearson chi-square. Predictors of new Backfill were analyzed by univariate and multivariate regression adjusted for patient demographics, treatment, baseline and 2-year change in inflammation and damage scores.

Results ICC for detection of Backfill in exercises 1 and 2 were 0.86/0.66 for status scores and 0.55/0.56 for 2-year change scores. Development of new Backfill correlated significantly with reduction in SPARCC SIJ score for inflammation (r=-0.50, p<0.0001) reduction of SSS erosion score (r=-0.47, p<0.0001), and development of new fat metaplasia (r=0.29, p=0.003), and was observed for both categories of treatment. In univariate analysis the following variables were associated with new Backfill: SPARCC SIJ inflammation baseline (β=0.08, p=0.01) and change (β=-0.15, p=0.0002) scores, erosion score at baseline (β=0.21, p=0.0001) and change in erosion (β=-0.39, p<0.0001), Backfill score at baseline (β-0.24, p<0.0001), change in fat metaplasia (β=0.28, p=0.009). Multivariate regression revealed the following as independent predictors of new Backfill: change in SPARCC SIJ inflammation (β=-0.10, p=0.003), 2-year change in SSS erosion (β=-0.39, p<0.0001), baseline backfill SSS score (β=-0.32, p<0.0001).

Conclusions Backfill is an MRI feature of SpA that can be reliably detected and its development is associated with the resolution of inflammation and reduction of erosion. Reparative tissue at the site of erosion has high signal intensity on T1WSE resembling fat metaplasia in the bone marrow.

Disclosure of Interest None declared

DOI 10.1136/annrheumdis-2014-eular.3828

Statistics from

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.