Biomarker-based clustering identifies distinct pulmonary function trajectories in early systemic sclerosis

BackgroundSystemic sclerosis (SSc) is a heterogeneous autoimmune disease in which interstitial lung disease (ILD) is a major determinant of mortality. Given that certain chemokines and adhesion molecules may be involved in the inflammation, subsequent vascular injury, and fibrosis observed in SSc, their circulating levels in peripheral blood may reflect the disease processes ranging from inflammation to vascular damage and fibrotic remodeling.

However, the potential of these biomarkers to identify patient subgroups with divergent pulmonary trajectories remains to be elucidated.MethodsWe performed a retrospective analysis of prospectively collected data from patients with early severe SSc (diffuse cutaneous SSc irrespective of ILD status or limited cutaneous SSc with ILD; disease duration <5 years) who were enrolled in a multicenter cohort. Serum levels of five chemokines and four soluble adhesion molecules were quantified at baseline.

Patients were classified based on these biomarker profiles using k-means clustering. Changes in pulmonary function were compared among clusters using relative changes in percent vital capacity (%VC).ResultsPatients (n = 92) were classified into three clusters: Cluster 1 (n = 37) with elevated sICAM-1 and sE-selectin; Cluster 2 (n = 13) with elevated CCL2, CXCL8, and sP-selectin; and Cluster 3 (n = 42) with no distinctive biomarker pattern.

Cluster 3 showed stable %VC and served as the reference group. Cluster 1 showed early decline (one-year difference: −8.41%; 95% CI: −12.62 to −4.20; p < 0.001) that attenuated by year two.

In contrast, Cluster 2 showed progressive decline (two-year difference: −7.77%; 95% CI: −15.25 to −0.29; p = 0.042). These biomarker-defined patterns were consistent with a vasculopathic–fibrotic profile in Cluster 1 and an inflammatory–vascular profile in Cluster 2.ConclusionSerum chemokine and adhesion molecule profiles may help stratify early severe SSc into biologically distinct subgroups with different pulmonary trajectories, supporting their potential utility for early risk stratification in SSc-ILD.