Changes in specific immune cell lineages, such as T and B cells, play a central role in the pathogenesis of rheumatoid arthritis (RA). However, a comprehensive evaluation of systemic immune cell changes in RA remains limited. Immune cell proportions of 104 subsets across granulocyte, T-cell, B-cell, and innate lineages were profiled by flow cytometry in 21 new-onset RA patients and 21 healthy controls. Non-parametric tests compared groups, followed by training a logistic regression-based AI model with cross-validation to characterize RA immune profiles and assess each subset's contribution. Among 104 immune cell subsets analyzed, 16 were indicative of RA. Increased proportions of marginal zone B cells, IgMhi B cells, CD11b+lineage- cells, monocytes, and MHC II+ monocytes, along with decreased eosinophils, reflected activation of innate and humoral immune responses in RA patients. Elevated levels of FoxP3+CD4+ regulatory T cells (FoxP3+ CD4 Treg) and CTLA4+ CD4 Treg cells, as well as increased MHC II+CD4+ and CD8+ T cells, PD-L1+ NK cells, and PD-L1+CD8+ NKT cells, suggested a compensatory immune response. The AI model distinguished immune profiles between RA patients and healthy controls with 100% sensitivity and specificity in this dataset, identifying RA by lower MHC II+ monocytes, higher CTLA4+ CD4 Treg cells, and elevated monocytes. These findings demonstrate the potential of using ICP hallmarks to develop novel diagnostic tools and therapeutic strategies for RA.
Keywords: artificial intelligence; flow cytometry; immune cell profile; precision diagnosis; rheumatoid arthritis.
© The Author(s) 2026. Published by Oxford University Press on behalf of Society for Leukocyte Biology.