{
  "abstract": "Background Despite the significant clinical benefits of immune checkpoint blockade (ICB), a substantial proportion of patients experience primary or acquired resistance, highlighting a critical need for predictive biomarkers that move beyond descriptive immune readouts. Current approaches often lack the mechanistic depth to fully capture the complex tumor microenvironment (TME) dynamics governing ICB response.Methods We developed IMIREG (Immune Microenvironment Regulon Signature), a mechanistically grounded 14-regulon signature, by systematically inferring transcription factor (TF) activity across four discovery immunotherapy cohorts (IMvigor010, I-SPY2, Kim, Gide). IMIREG’s predictive utility was rigorously validated across 40 independent ICB clinical trial datasets spanning 17 cancer types. We further characterized IMIREG’s physiological distribution across healthy tissues and single-cell atlases. Mechanistic dissection involved bulk RNA-seq deconvolution, pathway enrichment analysis, and a novel lineage-resolved classification framework utilizing pan-cancer single-cell RNA-seq compendia to identify immune cell drivers. Spatial immune profiling and longitudinal analyses of a triple-negative breast cancer (TNBC) cohort elucidated IMIREG’s dynamics during anti-PD-1 therapy.Results IMIREG consistently and robustly predicted clinical benefit from ICB (average AUROC ≥ 0.74), outperforming established biomarkers including T cell inflamed signature, IFN-γ signatures, and tumor mutational burden. IMIREG’s utility was shown to be predictive, not prognostic. Mechanistically, IMIREG activity was selectively enriched in functionally competent yet checkpoint-restrained immune compartments, notably encompassing exhausted/effector T cells and pro-inflammatory M1 macrophages. IMIREG-high tumors exhibited an inflamed TME with robust antigen presentation and positive correlations with key checkpoint molecules (PD-1, CTLA-4). Our lineage-resolved framework classified IMIREG-high tumors into T cell-driven, macrophage-driven, and dual-driven immune archetypes, revealing substantial inter-tumoral heterogeneity. Strikingly, the most immunologically active dual-driven phenotype exhibited near-complete erosion during metastatic progression. Furthermore, IMIREG served as a dynamic, real-time pharmacodynamic marker in TNBC patients, with higher baseline and on-treatment levels correlating with improved pathological complete response.Conclusions IMIREG provides a novel, mechanistically grounded biomarker that moves beyond enumerating immune components to measuring the core regulatory circuitry of anti-tumor immunity. This signature significantly enhances patient stratification for ICB and offers a powerful framework to understand the dynamic, multi-lineage nature of immune engagement and its collapse in advanced cancer, enabling more refined immunotherapy strategies.",
  "authors": [
    {
      "affiliations": [
        "University of Pittsburgh, Pittsburgh, PA, USA"
      ],
      "name": "Bashir Lawal"
    },
    {
      "affiliations": [
        "University of Pittsburgh, Pittsburgh, PA, USA"
      ],
      "name": "Renu Sharma"
    },
    {
      "affiliations": [
        "University of Pittsburgh, Pittsburgh, PA, USA"
      ],
      "name": "Akshat Gupta"
    },
    {
      "affiliations": [
        "University of Pittsburgh, Pittsburgh, PA, USA"
      ],
      "name": "Laizhi Zhang"
    },
    {
      "affiliations": [
        "University of Pittsburgh, Pittsburgh, PA, USA"
      ],
      "name": "Xiaosong Wang"
    }
  ],
  "title": "23 Lineage-resolved transcriptional program defines immune engagement archetypes and predicts immunotherapy response across multiple cancer types",
  "uid": "d7614e8d-00b9-5f74-bc38-8797eb6ae1d6"
}