Background/Purpose: Fat metaplasia is detected as bright signal on T1W MRI and has been shown to develop after resolution of inflammation in spine and sacroiliac joints (SIJ). Tissue with bright signal on T1W MRI, termed backfill (BF), may also fill areas of excavated sacral and iliac bone along the joint space and is thought to reflect repair of erosion. Imaging studies in the spine have indicated that fat metaplasia predicts development of new bone. We hypothesized that ankylosis in the SIJ develops following repair of erosion and that fat metaplasia is a key intermediary step in this pathway.

Methods: We used the SPARCC SIJ Structural Score (SSS) method to assess fat metaplasia (FAT), erosion (ER), BF, and ankylosis (ANK). This score relies on the T1W sequence and assesses 5 consecutive coronal slices anteriorly through the cartilaginous portion of the joint from the transitional slice. Lesions are scored dichotomously (present/absent) in SIJ quadrants (fat, erosion) or halves (backfill, ankylosis). Scoring ranges are: FAT (0-40), ER (0-40), BF (0-20), ANK (0-20). Four readers assessed 45 pairs of MRI scans blinded to time point (baseline, 2 years) from 45 cases in a prospective cohort receiving either standard (n=22) or anti-TNF (n=23) therapies. In a second study, two readers assessed 147 pairs of scans blinded to time point (baseline, 2 years) from cases either on standard (n=69) or anti-TNF (n=78) therapies. Univariate analyses and multivariate linear regression focused on identifying significant MRI predictors of change in BF and ANK scores, adjusted for age, sex, symptom duration, treatment, CRP (baseline and 2-year change), SPARCC SIJ inflammation score (baseline and 2-year change), and baseline SSS scores for FAT, ER, BF, and ANK. 

Results: Using mean SSS scores for 4 readers in the 45 cases, resolution of ER was significantly associated with the development of BF (p = 0.0082) and new ANK (p=0.045) at 2 years. Using mean scores of two readers in the 147 cases, resolution of ER was significantly associated with the development of BF (p<0.0001), FAT (p<0.0001) and new ANK (p=0.0001) at 2 years. New ANK was also significantly associated with development of FAT (p=0.0005). Associations were also significant in both treatment groups. A decrease in ER score was a significant predictor for development of new BF in the multivariate regression model (adjusted R² =0.44, F ratio 14.6, p<0.0001) (change in SSS erosion: β= -0.74, t= -4.1, p=0.0001). 31 (21.1%) of patients developed new ANK and these had significantly more resolution of ER than patients without new ANK (p=0.014, Mann-Whitney). Significant independent predictors of new ANK in the multivariate model (adjusted R²= 0.24, F ratio = 10.0, p <0.0001) were baseline BF score, decreased ER score and development of new FAT (Table).

	β coefficient	SE	t	p value
Age	-0.022	0.010	-2.25	0.026
BL SSS Backfill score	0.073	0.027	2.75	0.0068
Change in SSS Erosion score	-0.14	0.045	-3.11	0.0023
Change in SSS Fat score	0.20	0.058	3.36	0.001

Conclusion: Ankylosis in the SIJ develops following repair of erosion and fat metaplasia is a key intermediary step in this pathway.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/evidence-that-fat-metaplasia-is-a-key-intermediary-in-the-development-of-sacroiliac-joint-ankylosis-following-repair-of-erosions-in-patients-with-spondyloarthritis/