209P Mapping the landscape of immune cells for optimal index calculation using AI-powered image analysis of multiplexed immunohistochemistry in breast cancer patients

Bulk TILs & CD39+CD103+ sorted CD8 TILs were expanded, and underwent FACS analyses and TCRsequencing.Fresh frozen samples from all patients underwent whole exome and RNA sequencing.Spatial transcriptomic profiling and single cell RNA sequencing (10X Genomics) were performed on selected patients.Results: Flow cytometry analyses and T cell receptor (TCR) sequencing of TILs revealed differences in LM and PT.Post CD39+CD103+ sorting, LM TILs had higher proportions of CD39+LAG3+PD1+ triple-positive CD8 cells (p < 0.0001) and CD39+TIGIT+PD1+ cells (p< 0.05) compared to PT TILs, suggesting a potential tumourspecific population with high activation/exhaustion markers.Bulk LM TILs also had lower CD8 and higher proportion of CD4 cells compared to PT TILs.Organoid killing assays from patient 4108 demonstrated that LM TILs recognised and resulted in more killing of both LM and PT organoids than PT derived TILs.Spatial transcriptomic analyses showed immune exclusion of T cells in LM and distinct macrophage markers which may reveal potential immunosuppressive factors.WES/RNAseq revealed a similar neoantigen landscape between PT and LM, suggesting a possibility for a unifying therapeutic target.Single cell RNAseq results are still underway but preliminary bulk TCRseq from CD39+CD103+ CD8 T cells in PT and LM TILs showed overlap in a few TCR clones, suggesting shared TCR lineage.Conclusions: Better understanding of the LM TIL and shared neoantigen landscape of CRLM supports its potential for TIL therapy and TCR discovery.