Abstract Background: Single-cell spatial profiling enables detailed insights into the tumor microenvironment while maintaining the native tissue context. Germline pathogenic variants (PVs) are associated with decreased lymphocyte and increased tumor-associated macrophages from bulk transcriptomic data; however, this relationship has yet to be deconvoluted. To examine how spatial features vary by germline cancer predisposition, we performed single-molecule RNA imaging in a racially and ethnically diverse cohort with detailed genetic and clinical annotation. Methods: We performed single molecule RNA imaging on the Bruker CosMx platform of a cohort in the population-based Northern California Breast Cancer Family Registry (NC-BCFR) consisting of 222 samples from 144 women diagnosed 1995-2005 with 20 years of follow-up. We deployed an in-house machine learning pipeline to perform imaging-based cell segmentation and quantification of cell-level transcriptomic data for 6,519 genes across 1,083,345 cells. Normalization for each of 9 tissue microarrays and batch integration was followed by cell typing using well-established canonical markers and the reference-based SingleR algorithm. Spatial transcriptomics data were used to predict tumor HER2 and hormone receptor status (HR). Cell neighborhood was determined by scikit NearestNeighbors package with neighborhood number determined by finding the inflection point in an ElbowPlot analysis. Co-localization was defined with the colocation quotient analysis. We performed univariate and multivariate analyses to examine the correlation of germline PV status, race/ethnicity, age, tumor stage, and tumor subtype with proportion of each identified cell neighborhood. P-values were corrected for multiple hypothesis testing using Benjamini-Hochberg correction. Results: Of 144 NC-BCFR participants, 28 (19.4%) had at least one germline PV, with 11 BRCA1, 8 BRCA2, 3 PALB2, 2 ATM, and 1 FANCM PV represented. Tumor samples with germline PVs globally contained increased regulatory T cells and CD8+ T cell (Wilcox FDR=0.0059, 0.0070) and decreased fibroblast (Wilcoxon FDR=0.0085) populations compared to non-PV tumors. Cell neighborhood analysis defined 12 distinct microenvironmental compartments including including cancer core (4 distinct clusters), epithelial/fibrotic, epithelial-innate immune, stromal, adaptive immune, innate immune, fibrotic-immune depleted, and mixed (2 clusters). Tumor samples with germline PVs were enriched for neighborhood 0 (Wilcoxon FDR.001), defined by adaptive B (19%) and T lymphocytes (36%), of which 30% were CD4+, 26% are CD8+, 34% regulatory T cells, and 9.6% NK cells. In a multivariate model that included clinical subtype, age, race/ethnicity, and stage as covariates, the association between germline pathogenic PV and proportion neighborhood 0 remained significant (multivariate p=0.0035). We observed an increase in co-localization of cancer epithelial cells to CD8+ T cells (Wilcoxon FDR= 0.0306) and macrophages (Wilcoxon FDR=0.0349) within tumors from germline PV carriers vs non-carriers. Conclusions: Dynamic interplay among tumor cells and the host immune system is captured by single molecule spatial transcriptomic profiling. In the diverse, population-based, genetically characterized NC-BCFR, we observed substantial differences in the adaptive tumor immune environment of germline PV carriers vs. non-carriers. Our results suggest that inherited genetic risk shapes the tumor-immune landscape, which may inform strategies to optimize germline-targeted therapies. Citation Format: Immune microenvironment features of germline mutation-associated breast tumors revealed by single-molecule imaging in the Breast Cancer Family Registry. Immune microenvironment features of germline mutation-associated breast tumors revealed by single-molecule imaging in the Breast Cancer Family Registry abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS2-07-28.
McClure et al. (Tue,) studied this question.