Abstract A036: Spatial multi-omics identifies a tumor microenvironment signature predictive of immunotherapy response in mucosal melanomas

Abstract Background: Mucosal melanoma (MuM) is a rare and aggressive cancer with limited response to immunotherapy. Despite its clinical severity, the mechanisms that drive therapeutic resistance in MuM are still largely unclear. To elucidate the spatial and molecular determinants of therapeutic outcomes, we designed this study to identify the spatial and molecular factors that shape both responsiveness and resistance of MuM to immune checkpoint blockade therapies. Study Design and Methods: We employed a comprehensive spatial multi-omics strategy that integrated COMET-based single-cell spatial proteomics with high-resolution metabolite imaging via MALDI-IMS. Using this platform, we analyzed the cellular composition and mapped spatially organized cellular neighborhoods (CNs) across 97 FFPE tissue cores obtained from 26 MuM patients who received PD-1/PD-L1 or CTLA-4 blockade. We further examined how spatial architecture, cell–cell interactions, proteomic signatures, and metabolomic profiles of CNs varied between treatment responders and non-responders. Results: This integrative strategy enabled spatially resolved characterization of about 700K single cells, which were classified into 25 phenotypic clusters across nine major cellular lineages. Spatial neighborhood analysis identified 15 biologically distinct CNs, each exhibiting unique compositions and spatial arrangements of tumor, immune, and stromal components. Among patients who responded to immunotherapy, three tumor-associated CNs—the central tumor, invasive tumor, and tumor boundary CNs—were markedly enriched, displaying distinctive spatial architecture. Notably, the invasive and boundary CNs showed close spatial proximity among proliferating tumor cells, macrophages, and dendritic cells. These CD163+ macrophages demonstrated decreased IRF4 and Arg1 expression, indicative of reduced immunosuppressive function. In contrast, non-responders showed a predominance of a stromal-rich CN characterized mainly by SMA- stromal cells with limited immune cell infiltration, observed in both baseline and post-treatment samples. Spatial metabolomic analysis further revealed a substantial decrease in tryptophan-derived indole metabolites in responders, which strongly correlated with CD163 and CD11c expression, suggesting a coordinated immunometabolic reprogramming within the tumor microenvironment. Conclusions: Collectively, these insights underscore that immunotherapy response in MuM is orchestrated by an interplay of spatial tumor–immune architecture, stromal insulation, and metabolomic reprogramming collectively underpin differential immunotherapy responses in MuM. The spatially resolved signatures of tryptophan-derived metabolites not only emerge as promising biomarkers but also present tractable therapeutic targets to potentially overcome resistance in this aggressive melanoma subtype. Citation Format: Suhendan Ekmekcioglu, Jun Wang, Priyadharsini Nagarajan, SungNam Cho, Yunhe Liu, Erin H. Seeley, Jared K. Burks, Adi Diab, Linghua Wang. Spatial multi-omics identifies a tumor microenvironment signature predictive of immunotherapy response in mucosal melanomas abstract. In: Proceedings of the AACR Immuno-Oncology Conference (AACR IO): Discovery and Innovation in Cancer Immunology: Revolutionizing Treatment through Immunotherapy; 2026 Feb 18-21; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Immunol Res 2026;14(2 Suppl):Abstract nr A036.