Abstract
Background: Systemic sclerosis (SSc) is a heterogeneous autoimmune disease with high mortality driven by multiorgan involvement and limited therapeutic options. Traditional classifications based on skin involvement or serology are insufficient to capture disease complexity or predict outcomes accurately. Methods: K means clustering was applied to clinical and serological data from 402 SSc patients in the PRECISESADS cohort. The resulting clusters were validated in an independent local cohort (n = 213). To explore molecular differences, a random subset of 154 PRECISESADS patients underwent serum proteomic profiling using a panel of 92 organ damage-related proteins. Functional relevance was further investigated by exposing dermal fibroblasts to patient serum and assessing gene expression. Results: Two distinct clusters were identified and validated, differing in organ involvement and autoantibody profiles. Cluster 2 was associated with more severe disease, including higher prevalence of ILD, PAH, and musculoskeletal manifestations, and enriched in anti-Scl-70 antibodies. Proteomic analysis revealed upregulation of 26 proteins in Cluster 2, related to fibrosis, inflammation, and endothelial dysfunction. Serum from these patients induced the in vitro expression of pro-fibrotic and inflammatory genes in fibroblasts. Altered levels of several proteins also correlated with relevant clinical features, suggesting potential biomarker utility. Conclusion: Unsupervised clustering and proteomic profiling reveal biologically distinct subgroups within SSc, beyond traditional clinical or serological classifications. Our findings support the integration of molecular tools into patient stratification strategies, paving the way toward personalized medicine in SSc.