Factors associated with rapid spinal radiographic progression in patients with axial spondyloarthritis: A...

To identify associated and protective factors of rapid spinal radiographic progression in axial spondyloarthritis (axSpA) using artificial intelligence (AI).

We conducted a hospital-based retrospective cohort study involving 242 axSpA patients taken ≥2 lateral spine radiographs between 2002 and 2024. Spinal damage was assessed with modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) by a deep learning model. Each pair of consecutive radiographs defined an observational interval (total 379 intervals); annual mSASSS progression rate was calculated for each interval. Demographics, clinical features, baseline mSASSS, activity indices, cumulative dosage of prescriptions, and laboratory recordings were collected. Time-dependent generalized estimating equations (GEE) were applied to identify associated or protective factors of rapid spinal radiographic progression (ΔmSASSS/year >1), accounting for within-patient correlation.

For recorded intervals, mean mSASSS progression was 0.5/year; 26.7% of intervals showed progression >1/year. For enrolled patients, mean mSASSS progression was 0.6/year; 27.3% of intervals showed progression >1/year. Conditional multivariable GEE analysis revealed age at baseline mSASSS, especially ≥40 years, was independently associated with rapid mSASSS progression [adjusted odds ratio (aOR), 1.03; 95% confidence interval (CI), 1.003-1.06]. Higher cumulative dosage of non-steroidal anti-inflammatory drugs (NSAIDs) during the intervals was negatively associated with rapid mSASSS progression (aOR, 0.38; 95% CI, 0.19-0.75). Cumulative dosage of tumor necrosis factor inhibitors and secukinumab during the intervals was independent of rapid mSASSS progression.

Using AI-assisted mSASSS scoring, this retrospective cohort study identified older age at assessment as an associated factor and full-dose NSAIDs use as protective factor for rapid spinal radiographic progression.