{
  "abstract": "Background The lung histopathology of idiopathic pulmonary fibrosis (IPF) patients undergo extensive tissue remodeling during its asynchronous disease evolution giving rise to the spatial heterogeneity observed in the fibrotic lung. Co-occurring pathologic programs within distinct spatial regions are hypothesized to have underlying cellular and molecular interactions that mediate disease pathogenesis. Having a multomic understanding of the heterogeneous nature of PF will be critical for improving patient diagnosis and treatment in the future.Methods Using the multiomic SpaceIQ TM platform, we analyzed a publicly available data set presented in Vannan et al. Sci. Adv. 2025 consisting of H&E images and spatial transcriptomics of Xenium from 41 patient lung tissue samples (GEO GSE250346: 8 healthy controls and 33 peripheral PF patients). 34 nuclei morphology and color stain features were derived from H&E and used to perform unbiased cell typing. PF-specific microdomains were identified based on differential spatial arrangements of cells between PF and healthy patients. We computed tile-based gene expression on the Xenium spots and registered tiles to the H&E image. Feature correlation analysis was performed between the tile gene expressions and the spatial microdomains. Biological pathway analysis was inferred on the top correlated genes with discovered microdomains.Results Unbiased cell typing on the H&E features yielded 19 distinct cell types, from which 4 cell types were identified to be involved in 3 significant PF-specific microdomains. Feature correlation analysis yielded ATG7, CD44, and LPAR1 genes to be among the top correlated genes from the H&E cell type microdomains. Biological pathway inference on the correlated genes identified epithelial mesenchymal transition (EMT), TGF-beta regulation of ECM, alveolar epithelial cell autophagy, regular glucocorticoid receptor, uPA/uPAR-mediated signaling, AHR signaling in Treg/Dcs as significant pathways (p<0.001) implicated in the PF-specific microdomains ( figure 1).Conclusions Our multiomic analysis with SpaceIQ platform reveals microdomains showing consistent known biology associated with PF. Specifically, we identified ATG7 central to autophagy and LPAR1 involved in fibroblast recruitment as potential therapeutic targets. We also identified a top correlated gene, CD44, which has been shown to be a reliable diagnostic marker for PF. This study demonstrates the benefit of using a multiomic spatial approach to generate hypotheses that are both confirmatory and potentially novel to improve our understanding of a complex and heterogeneous disease like PF.Abstract 105 Figure 1Mapping spatial heterogeneity in IPF with a multiomic feature network. Spatial multiomic analysis reveals correlated molecular programs underlying heterogeneity in PF pathogenesis through H&E unbiased cell types",
  "authors": [
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA"
      ],
      "name": "Raymond Yan"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA"
      ],
      "name": "Brian Falkenstein"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA"
      ],
      "name": "A Burak Tosun"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA"
      ],
      "name": "Filippo Pullara"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA",
        "University of PIttsburgh, Pittsburgh, PA, USA"
      ],
      "name": "S Chakra Chennubhotla"
    }
  ],
  "title": "105 Spatial multiomic analysis captures immuno-modulatory programs of ATG7, CD44, and LPAR1 genes in mediating disease heterogeneity of pulmonary fibrosis (PF)",
  "uid": "f8e6417d-dc6e-54b5-9d92-5e3069960491"
}