Abstract LB076: Spatial transcriptomics reveals differences in the tumor and immune microenvironment of high-grade serous ovarian cancers with differing responses to immune checkpoint inhibitors

Abstract Introduction: High-grade serous ovarian cancer (HGSC) is an aggressive malignancy with poor treatment outcomes. Although transformative for some cancer types, immune checkpoint inhibitors (ICIs) have been only marginally successful at best in HGSC. Comprehensive interrogation of the cellular and molecular features of the tumor and immune microenvironment (TIME) of HGSC cases with varying responses to ICI could provide novel insights into the potential drivers of therapeutic response. Methods: We applied 10x Genomics Visium Spatial Transcriptomics (ST) to archival tissue sections from 10 patients treated with ICI (nivolumab or pembrolizumab) with varying degrees of response. Integrated cluster analysis of 39,147 ST features using Louvain optimization resulted in 20 integrated clusters. Molecular annotation using GSEA, ESTIMATE, and CIBERSORTx was performed on individual clusters to assess spatially resolved cellular transcriptional states within the cohort and by ICI response. Results: Bioinformatics analysis of marker gene data from each group (responder or non-responder) provided insights into cell types for each cluster defined by transcriptional states. Major cell type populations by cluster included tumor, immune, fibroblast, epithelial, and smooth muscle. Cluster 1 cell type of both responders and non-responders is molecularly defined by known HGSC genes including CD24, WFDC2, SLPI, and MUC1. Non-responders also expressed CRIP1 and ELF3, which are associated with metastasis and cancer cell stemness, respectively. Pathway enrichment analysis across the entire cohort identified distinct immune-related pathways, including allograft rejection, IL6-JAK-STAT3 signaling and interferon responses, corroborating the immune cell deconvolution findings from CIBERSORTx. Interestingly, clusters 6 and 8 for both groups seem to represent a specialized population of lipid-associated macrophages (LAMs), marked by expression of LYZ, APOE, APOC1, GPNMB. However, in responders, we observed upregulation of antigen presenting genes, including LYZ, C1QA, C1QC, CD74, and CXCR4. In contrast, non-responders had high expression of immunosuppressive macrophages including MMP9 and IFI30 within the LAM compartment. Non-responders demonstrated higher immune composition scores in several clusters, predominantly for M0 and M2 macrophages, which are associated with immunosuppressive phenotypes. In contrast, responders displayed a more immune-activated profile, with increased infiltration of CD8+ T cells, activated CD4+ memory T cells, and activated NK cells. Notably, M2 macrophages were more abundant in the tumor cluster of non-responders, indicating a more immunosuppressive TME. Conclusions: These results underscore the heterogeneity within the tumor and immune microenvironments in HGSC tumors that may uncover novel markers or targets of response to ICI. Citation Format: Jing Qian, Lynda D. Roman, Rania Bassiouni, Hanim I. Ozkizilkaya, Kathy Kwock, Yuxin Jin, Javier Arias-Stella, Joseph W. Carlson, John D. Carpten. Spatial transcriptomics reveals differences in the tumor and immune microenvironment of high-grade serous ovarian cancers with differing responses to immune checkpoint inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_2):Abstract nr LB076.