spar0010">Three readers blinded recorded backfill seen on T1SE MRI scans from 647 subjects: 297 patients with ankylosing spondylitis (AS), 126 patients with non-radiographic axial SpA (nr-axSpA), 147 patients with NSBP, and 77 healthy controls. The SPARCC SIJ Structural Score (SSS) method was used to assess backfill. The changes of backfill were evaluated by the follow-up MRI scans from 157 patients. We summarized the characteristics of backfill and calculated its sensitivity and specificity for diagnosing axial SpA.
spar0015">Backfill was recorded in 78.8% AS patients, 11.1% nr-axSpA patients, 1.8% patients with NSBP, and no healthy control. Backfill affected more frequently at ilium bone, lower half of sacroiliac joints in axial SpA (both P < 0.05). The SSS score of backfill was much higher in axial SpA than in patients with NSBP (both P < 0.01) and it did not correlate with demographics and BASDAI, BASFI, and CRP (all P > 0.05). The score of backfill only positively correlated with symptom duration in AS (r = 0.251, P < 0.01) and in nr-axSpA (r = 0.743, P < 0.01) patients. Only 8.9% patients had the change of backfill in an average follow-up time of 1.09 years. Backfill had high specificity (0.98) and moderate sensitivity (0.59) for diagnosing axial SpA.
spar0020">We summarized the characteristics of backfill and found that backfill is a specific sign of axial SpA seen on T1SE MRI.