Background/Purpose: This study aims to characterize the metabolomic fingerprint of several systemic autoimmune diseases (SADs) and apply machine learning (ML) techniques to identify disease-specific biomarkers and improve patient diagnosis and stratification.

Methods: A total of 716 individuals from the PRECISESADS study, were analyzed, including 272 with rheumatoid arthritis (RA), 183 with systemic lupus erythematosus (SLE), 148 with primary antiphospholipid syndrome (PAPS), 70 with systemic sclerosis (SSc), and 43 healthy donors (HDs). Serum metabolomic profiles were assessed using nuclear magnetic resonance (Nightingale), alongside extensive clinical and analytical data. A combination of ML approaches, including unsupervised patient stratification, and a supervised Logistic Regression using 5-Fold Cross Validation was used to identify potential diagnostic signatures. Correlation analyses explored associations between metabolic alterations and clinical features.

Results: Several metabolites (metabs) were differentially expressed in each disease compared to HDs, with the most alterations in SSc (99 metabs) and APS (68 metabs), followed by SLE (30 metabs) and RA (17 metabs). Notably, some cardiovascular-related metabolites (e.g., histidine, albumin, L-LDL-C, L-LDL-CE) were simultaneously altered across all diseases, while each also exhibited unique metabolic change. ML generated disease-specific diagnostic signatures with AUCs above 0.8 in all cases.Unsupervised clustering analysis of the entire cohort identified three patient clusters (C1, C2, C3), with each disease represented across all clusters in varying proportions. C1 showed significant differences compared to C3 in both clinical features and metabolomic profiles, while C2 represented an intermediate group. Patients in C1 had increased cardiovascular risk markers compared with C3, while C2 represented an intermediate group.In APS, C1 was linked to higher thrombotic risk (aGAPSS), arterial events, and metabolic comorbidities, while C2 showed more venous thrombosis and pregnancy complications. In SLE, C1 had more cardiovascular risk factors, lupus nephritis, and anti-dsDNA positivity than C3. In RA, C1 showed higher rates of obesity, diabetes, and atherosclerosis, whereas C3 had greater disease activity and autoantibody positivity. In SSc, C1 was associated with dyslipidemia, lung fibrosis, and anti-U1-RNP, while C3 had more skin involvement and anti-centromere positivity.

Conclusion: Distinct metabolomic profiles were identified with shared and disease-specific alterations across systemic autoimmune diseases. ML revealed accurate diagnostic signatures and uncovered patient subgroups with unique clinical and metabolomic features, highlighting shared pathogenic mechanisms and disease-specific features and supporting the potential for personalized therapies targeting metabolic pathways.Supported by CPS: RYC2021-033828-I; PID2022-141500OA-I00; DIN2022-012766 Minister of Science, Innovation and Universities co-financed by the European Union; and CLP: (PI24/00959, CD21/00187 and RICOR-24/0007/0019), co-financed by European Union. EU/EFPIA-IMI-PRECISESADS (n° 115565)

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/data-driven-metabolomics-identifies-diagnostic-signatures-and-patient-subgroups-in-systemic-autoimmune-disorders/