{
  "abstract": "Background Lung cancer remains the tumour type with the highest mortality rate globally. Whilst therapeutic options, both systemic and targeted are increasing, selecting patients for these therapies become a critical factor. The tumour microenvironment (TME) has been recognized as an increasingly important component to profile, both phenotypically and functionally, to understand the cellular composition and how cells communicate in the TME. Moreover, these profiles can be used in the development of predictive signatures and companion diagnostic assays for patient triage.Methods Using our SpaceIQ TM platform, we developed a simple prognostic test on key biomarkers predictive of the likelihood of patients responding to adjuvant immunotherapies. Briefly, we applied cell segmentation and unbiased cell typing on the spatial proteomics data (28 patient cores in 44-plex mIF panel with 10 responders and 18 non-responders) yielding 23 distinct cell types. Differential spatial arrangements of cells based on pointwise mutual information between response groups resulted in response-specific microdomains. Biomarker profiles from each of the cell types involved in the microdomains served as the basis for a de-plexing algorithm in obtaining a low-dimensional proxy representation of the corresponding cell types. Finally, a spatial score between the de-plexed representation of cell types and its estimated mean biomarker expression were derived as predictive features for response outcome in a simplified prognostic model.Results Two predictive spatial interactions for response outcome emerged involving metabolic activity in immune cells, in particular macrophages. Moreover, a resistant signature defined by spatial interactions between an immune cell subpopulation (unbiased cell type 12111) having increased LDHA and decreased CPT1A activity with a macrophage subpopulation (unbiased cell type 22121) was predictive for IO response (in-sample/cross-validated (CV) AUC: 0.95/0.75). A response signature promoting spatial interaction between epithelial/tumor cell population with decreased Ki67 expression in tumor-associated macrophages (unbiased cell type 22200) was predictive for IO response (in-sample/CV AUC: 0.83/0.70) ( figure 1).Conclusions Through the SpaceIQ platform, we were able to de-plex a 44-plex IF panel to a few key predictive biomarkers involving metabolic and macrophage signatures that yielded good predictive performance on immunotherapy response in NSCLC patients. We hypothesize that metabolic reprogramming of glycolysis in macrophages through lowered CPT1A stabilization of LDHA may lead to less effective pro-inflammatory response of M1 macrophages contributing to the resistance of IO therapy. We also observed evidence that interaction between epithelial/tumor cells with mitigated proliferation of GLUT1-active tumor-associated macrophages led to improved response to IO therapy.Abstract 95 Figure 1(A) unbiased cell typing extracts numerically stable, spatially distinct and biologically interpretable recursive cell types; (B) microdomain families emerge from spatial domain PMI network; (C) microdomains shaping cancer outcomes",
  "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": [
        "Frazer Institute, Faculty of Health, Medicine and Behavioural Sciences, The University of Queensland, Woolloongabba, QLD, Australia"
      ],
      "name": "James Monkman"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA"
      ],
      "name": "Filippo Pullara"
    },
    {
      "affiliations": [
        "Frazer Institute, Faculty of Health, Medicine and Behavioural Sciences, The University of Queensland, Woolloongabba, QLD, Australia"
      ],
      "name": "Arutha Kulasinghe"
    },
    {
      "affiliations": [
        "PredxBio, Inc., Pittsburgh, PA, USA",
        "University of PIttsburgh, Pittsburgh, PA, USA"
      ],
      "name": "S Chakra Chennubhotla"
    }
  ],
  "title": "95 Metabolic reprogramming of glycolysis, fatty acid oxidation, and anti-oxidative stress in macrophages is predictive of immunotherapy response in patients with non-small cell lung cancer (NSCLC)",
  "uid": "9251280d-e86b-55b9-a451-64c09b2490d7"
}