Abstract C051: Genomic and spatial immune biomarkers predictive of chemoimmunotherapy response in triple-negative breast cancer

Abstract Triple-negative breast cancer (TNBC) is an aggressive and heterogeneous subtype with a 5-year overall survival (OS) below 70%. Although the addition of immune checkpoint inhibitors to neoadjuvant chemotherapy (NAC-IO) has improved outcomes, only 50–65% of patients achieve pathological complete response (pCR), while others experience residual disease (RD) and poorer prognosis. Current biomarkers remain inadequate for accurate patient stratification, highlighting the need for more reliable predictors of NAC-IO response. Emerging evidence suggests that genetic alterations, including MYC amplification, PTEN loss, MAPK pathway activation, and the spatial distribution of immune cells, influence NAC-IO response, though underlying mechanisms remain poorly defined. Our preliminary analysis of TCGA and METABRIC datasets demonstrates that high MYC and low PTEN expression correlate with reduced IFN-gamma response and decreased CD8+ T-cell infiltration signatures in TNBC patients, providing a rationale for integrated genomic and spatial profiling. These findings lead to our hypothesis that somatic genetic alterations modulate tumor immune microenvironment and thereby influence treatment response to NAC-IO. This study integrates genomic and spatial multi-omics to identify predictive biomarkers of NAC-IO response in TNBC. Treatment-naïve formalin-fixed paraffin-embedded (FFPE) tumor biopsies from 39 TNBC patients treated with NAC-IO were collected from the University of Alabama at Birmingham, University of Florida, and Baylor Scott and White Health, along with clinicopathological data, including treatment response (pCR vs RD) and OS. Whole-exome sequencing was performed to identify somatic mutations and copy number alterations, including pathway-level assessment, mutational signatures, and clonal structure. Spatial profiling was performed on adjacent FFPE sections using spatial transcriptomics (Visium CytAssist) and multiplex spatial proteomics (Lunaphore COMET™) with a 36-plex antibody panel to characterize tumor cells, stromal cells, major immune cell subsets (CD8+/CD4+ T cells, regulatory T cells, B cells, M1/M2 macrophages, dendritic cells, NK cells), and functional cell states including antigen presentation, exhaustion, and proliferation. Computational integration will use machine-learning frameworks to relate genomic alterations, immune-cell densities, and spatial interactions to treatment response. We anticipate identifying genomic alterations and spatial immune cell signatures that predict NAC-IO response in TNBC. Spatial immune profiling may reveal differential patterns such as variation in immune infiltration, presence of immunosuppressive myeloid niches, or exclusion of cytotoxic T cells in tumors of RD vs pCR patients. Integration of these data may uncover associations between specific oncogenic alterations and spatial immune architectures that predict NAC-IO response, ultimately supporting development of a multimodal biomarker to guide NAC-IO optimization. Citation Format: Mahak Bhargava, Jaya Ruth. Asirvatham, Brian D. Stewart, Bohan Ning, Hua Guo, Emiel A. M. Janssen, Geetanjali Saini, Nusrat Jahan, Surojit Sarkar, Vandana Kalia, Julienne L. Carstens, Selvarangan Ponnazhagan, Ritu Aneja. Genomic and spatial immune biomarkers predictive of chemoimmunotherapy response in triple-negative breast cancer [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 C051.