Abstract B31: Depicting spatially-resolved immune landscapes in long-term ovarian cancer survivors by imaging mass cytometry

Sammy Ferri-Borgogno,Javier A Gomez,Ivo Veletic,Jared K Burks,Samuel C Mok

DOI: https://doi.org/10.1158/2326-6074.tumimm22-b31

IF: 10.1

2022-12-03

Cancer Immunology Research

Abstract:High-grade serous ovarian cancer (HGSC) is the most common subtype of ovarian cancer, accounting for 70% of all ovarian cancer deaths. Most patients with HGSC are diagnosed at advanced stages with the tumors spread beyond the pelvis. HGSC is notable sensitive to platinum/taxane-based chemotherapy before or following debulking surgery. Despite the high response rate, most patients will develop recurrent chemoresistant disease, and die from the disease within 2 years (short-term survivors, STS). Despite these dismal statistics, 10-15% of patients with advanced stage HGSC will survive 7 or more years following the diagnosis (long-term survivors, LTS). Immune cells in the tumor microenvironment (TME) have been shown to modulate the malignant phenotypes of HGSC. However, the mechanisms by which these cells interact with other cell types in the TME to modulate patient survival rates remain unclear. We hypothesize that spatially-resolved immune signatures associated with survival in STS and LTS patients can be identified, which may serve as a new generation of prognostic biomarkers for HGSC. To test this, we used 35 metal-tagged antibodies, which detect various cell specific and immune related markers, and imaging mass cytometry (IMC) to generate immune landscapes from a total of 47 advanced stage, and treatment naïve HGSC tumors (obtained from 21 LTS and 26 STS optimally debulked patients). Formalin fixed paraffin embedded (FFPE) tissue sections were stained. Images of each tissue section were acquired by the Fluidigm Helios CyTOF instrument utilizing the laser ablation module of the Hyperion Imaging System, and analyzed with the Visiopharm software. Tumor and stromal areas were first separated based on the presence or absence of Keratin 8/18 and SMA respectively. Cells boundaries and phenotypes were determined by a pretrained Artificial Intelligence algorithm using Visiopharm's unbiased autoclustering module. Cell densities and spatial relationships of identified phenotypes and statistical analysis were then calculated in R. Our results demonstrated significantly higher intratumoral cell densities of NK (CD56+CD25+) and activated T cells (CD8+CD44+Granzyme B+ and CD4+CD25+) in LTS than STS. Moreover, we showed that densities of multiple subpopulations of immunosuppressive macrophages (CD68+CD163+) expressing both TIM3 and CXCR5 markers were higher in STS than LTS, suggesting that the macrophages were actively engulfing CD8+ and CD4+ T cells in the TME. Furthermore, spatial analysis showed that increased densities of TIM3+CXCR5+ macrophages and increased number of neighborhoods between these macrophages and a variety of TME cells in the STS compared to LTS. In conclusions, our findings depict distinct spatially-resolved immune landscapes that are associated with LTS in HGSC patients. Further studies on the crosstalk networks established between specific immune cell types and their neighboring cells in STS and LTS will allow the identification of novel therapeutic targets that can improve patient survival and are warranted. Citation Format: Sammy Ferri-Borgogno, Javier A Gomez, Ivo Veletic, Jared K Burks, Samuel C Mok. Depicting spatially-resolved immune landscapes in long-term ovarian cancer survivors by imaging mass cytometry [abstract]. In: Proceedings of the AACR Special Conference: Tumor Immunology and Immunotherapy; 2022 Oct 21-24; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2022;10(12 Suppl) nr B31.

oncology,immunology