47P Correlation of clinical, genetic and transcriptomic traits with PD-L1 positivity in TNBC patients

Background: Compared with other breast cancer subtypes, TNBC is characterised by higher tumour mutational burden, elevated levels of PD-L1 expression and increased levels of immune cell infiltration.Predictors of immunotherapy response are not yet fully elucidated with several responders harbouring both PD-L1-and TMB-low tumours.We investigated the clinical and molecular traits associated with PD-L1 expression in TNBC patients with the aim to uncover distinct signatures that may impact on PD-L1 status and could potentially explain divergent responses to ICIs.Methods: We performed WES, RNAseq and IHC using FFPE blocks from primary tumour resections derived from 196 treatment-nave TNBC patients.Slides were stained with the anti-PD-L1 clone 22C3 and scored using a cut-off of CPS>10 for PD-L1-high samples.DGE analysis was conducted between PD-L1-high and -low cases.The read count matrix was normalized using the median of ratios via DESeq2.A generalized linear model identified DEGs, followed by pathway analysis using overrepresentation analysis with KEGG pathways.Additionally, GSEA was employed to evaluate specific gene sets within the expression data.Results: Almost one fourth of TNBC patients were PD-L1-high.The most frequent mutations in both groups involved TP53, PIK3CA, BRCA1, RB1 and PTEN genes with PIK3CA mutations almost exclusively encountered in PD-L1-high tumours.In PD-L1high tumours, most DEGs (n1708) were downregulated, including cell adhesion and collagen-related genes, while the up-regulated genes (n481) included HLA isotypes, IFN-gamma, JAK2 and STAT1.Most DEGs were involved in the neuroactive ligand-receptor interactions, calcium, cAMP and cell adhesion signalling pathways.Immune, chemokine and JAK-STAT signalling pathways were positively enriched in PD-L1-high tumours.Conclusions: This integrative analysis of clinical, genetic, transcriptomic and immunohistochemical data revealed differences in mutational signatures and gene expression patterns between PD-L1-high and -low TNBC tumours.