Background/Purpose:

In ankylosing spondylitis (AS), it is hypothesized that stepwise pathologic changes in the spine begin with active inflammation, then fatty degeneration (Fat), leading to syndesmophyte formation.  Using data from the golimumab (GLM) AS study (GO-RAISE), we assessed  how fatty and inflammatory lesions detected by MRI related to subsequent syndesmophytes and bridging on x-ray.

Methods:

In GO-RAISE, 356 patients with definite AS, BASDAI >4, and back pain score >4 were randomized to SC GLM 50mg, 100mg, or PBO q4wks. All patients were to have lateral view radiographs of the cervical and lumbar spine performed at baseline (BL), wks104 and 208. At MRI substudy sites, serial spine MRI scans in the sagittal plane were acquired using 1.5T scanners with T1 and short tau inversion recovery (STIR) sequences at BL, wk14, and wk104 for 98 patients. Two imaging readers blinded to treatment and image order independently evaluated anterior and posterior vertebral corners on x-rays for syndesmophytes/bridging.  Blinded MRI reads were done using ASspiMRI-a scoring of active inflammation (Infl) in vertebral units (VU), with notation of lesional location by VU quadrant (upper, lower x anterior, posterior); and the modified ASspiMRI-c scoring method to assess structural changes including presence of Fat in each of the 23 VU. This analysis focused on percentage of vertebral corners with new syndesmophytes/bridging visible on x-ray at wk 104 and 208, and how this percentage varied by presence or absence of prior Infl or Fat lesions in the corresponding VU quadrants at BL and wk14.

Results:

There were 91 patients in the substudy representing over 1200 VU (cervical, lumbar) with evaluable Fat data by MRI at BL and wk14 and corresponding vertebral corner x-ray data at wk104 and 208.  Of VU with Fat at wk14, more frequently there was also Fat at BL and Fat persisting to wk104, and Infl lesions were more frequently absent at BL than present.  Overall, as assessed by both readers individually and concordantly, the percentage of corners with new syndesmophytes at wk 104 or 208 was greater in VU where Fat was persistently present at both BL and wk14, compared to those where Fat was only present at wk14.  New syndesmophytes were significantly more likely to develop in VU with Fat at both BL and wk14 than those that did not have persistent fat (odds ratios for wk208 syndesmophyte/bridging, Reader 1, 3.27; Reader 2, 2.42).  The combined presence at BL of Fat and Infl lesions in a VU quadrant appeared to further increase the chance of new syndesmophytes developing at the corresponding vertebral corner ([Reader #1] 7.9% and 2.5% of corners with new syndesmophytes at wk104 for VU with BL Fat+Infl/ wk14 Fat  vs. BL Fat-no Infl/ wk14 Fat,  respectively, and [Reader #2] 4.4% and 2.7%, respectively).   Conversely, for VU without Fat or Infl at BL but wk14 Fat present, only 0-1.5% of corners developed new syndesmophytes later.

Conclusion:

This detailed analysis of spine MRI and radiographs at the vertebral level supports the hypothesis that in AS, fatty degeneration and active inflammation within vertebral unit favors progression toward subsequent syndesmophyte growth and ankylosis. 

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/persistent-fatty-lesions-in-the-vertebrae-in-ankylosing-spondylitis-favor-subsequent-new-syndesmophytes-imaging-results-of-a-phase-iii-randomized-placebo-controlled-study/