{
  "abstract": "Background Synovial inflammation in psoriatic arthritis (PsA) lacks a validated histological classification framework. Current approaches apply rheumatoid arthritis (RA) synovial pathotypes to PsA, which overlooks immunobiological differences between the diseases and limits the predictive power of tissue stratification. We aimed to define data-driven, PsA-specific synovial pathotypes to facilitate tissue-driven precision medicine. Methods In this cross-sectional study, ultrasound-guided synovial biopsies were obtained from 61 patients with PsA (n=22), seropositive RA (n=23), seronegative RA (n=10), and other inflammatory arthritides (n=6). Sections underwent multi-modal staining, including H&E, picrosirius red, and immunohistochemistry for CD3, CD20, CD68, CD138, and CD31. Whole-slide digital image analysis and automated pixel classifiers quantified cellular and stromal metrics. Unsupervised hierarchical clustering identified intrinsic PsA pathotypes, which were correlated with clinical disease activity, radiographic osteoarthritis (OA), and longitudinal treatment responses. Findings While a subset of PsA patients shared an inflammatory signature with seropositive RA, unsupervised clustering of the PsA cohort revealed three distinct, novel pathotypes: Dense Myeloid (macrophage-driven), Fibro-Stromal (fibrosis and pauci-immune), and Angio-Lymphoid (hypervascular with significant CD138+ plasma cell infiltration). Pathotypes were independent of treatment and concurrent radiographic knee OA. The Fibro-Stromal pathotype correlated with significantly higher cutaneous psoriasis severity. Longitudinally, patients with the Fibro-Stromal pathotype exhibited a trend toward lower treatment response (44%) following standard interventions (methotrexate or TNF inhibitors) compared to the Dense Myeloid (50%) and Angio-Lymphoid (80%) pathotypes. Interpretation We present the first computationally derived synovial pathotypes specific to PsA, revealing distinct plasma cell and fibrotic architectures not captured by RA-centric models. To enable clinical translation, we distilled these findings into a pragmatic, four-marker diagnostic decision tree (CD31, CD68, CD138, and picrosirius red) designed for routine histopathology workflows, providing a staging ground for biomarker-stratified clinical trials in PsA. Funding Funding from the NIHR Biomedical Research Centre: Manchester, Arthritis UK, the Wellcome Trust, Brit-PACT, the National Psoriasis Foundation, and the Medical Research Council.",
  "abstract_shown": true,
  "affiliations": [
    "The University of Manchester"
  ],
  "article_type": "PUBLISHAHEADOFPRINT",
  "arxiv": null,
  "author_total": 14,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "R. M. Hum"
    },
    {
      "name": "M. Christofi"
    },
    {
      "name": "S. L. Smith"
    },
    {
      "name": "Y. Tan"
    },
    {
      "name": "R. Shukla"
    },
    {
      "name": "L. M. Marshall"
    },
    {
      "name": "M. Smith"
    },
    {
      "name": "A. Williams"
    },
    {
      "name": "G. Orozco"
    },
    {
      "name": "S. Viatte"
    },
    {
      "name": "M. Buch"
    },
    {
      "name": "A. Barton"
    },
    {
      "name": "P. Ho"
    },
    {
      "name": "P. Martin"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/6qffm995eqcqz.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.11.06.26367442",
  "fulltext_url": "https://www.medrxiv.org/content/10.64898/2026.11.06.26367442v1.full",
  "id": "6qffm995eqcqz",
  "issns": [],
  "landing_url": "https://www.medrxiv.org/content/10.64898/2026.11.06.26367442v1",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "medrxiv",
      "rheumatology"
    ]
  ],
  "openalex": null,
  "pdf_url": "https://www.medrxiv.org/content/10.64898/2026.11.06.26367442v1.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-11-06",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/6qffm995eqcqz.ris",
  "sources": [
    {
      "first_seen": "2026-11-10T00:22:41Z",
      "native_id": "10.64898/2026.11.06.26367442",
      "source": "medrxiv",
      "url": "https://www.medrxiv.org/content/10.64898/2026.11.06.26367442v1"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "medrxiv:rheumatology",
      "subfield": "general"
    }
  ],
  "title": "Redefining Synovial Pathotypes in Psoriatic Arthritis through Multi-Modal Histology and Machine Learning (The PSYNOVIUM Study)",
  "title_text": "Redefining Synovial Pathotypes in Psoriatic Arthritis through Multi-Modal Histology and Machine Learning (The PSYNOVIUM Study)",
  "type": "preprint",
  "updated_at": "2026-11-10T00:22:41Z",
  "url": "https://research.draconitic.com/paper/6qffm995eqcqz",
  "venue": "medRxiv",
  "versions": []
}