Spatial insights into immunotherapy response in non-small cell lung cancer (NSCLC) by multiplexed tissue imaging

Abstract The spatial localisation of immune cells within tumours are key to understand the intercellular communications that can dictate clinical outcomes. Here, we demonstrate an analysis pipeline for highly multiplexed CODEX data to phenotype and profile spatial features and interactions in NSCLC patients that subsequently received PD1 axis immunotherapy. We found that regulatory T cells (Tregs) are enriched in non-responding patients and this was consistent with their localization within stromal and peripheral tumour-margins. Proximity-based interactions between Tregs and both monocytes (p = 0.009) and CD8 + T cells (p = 0.009) were more frequently found in non-responding patients, while macrophages were more frequently located in proximity to HLADR + tumour cells (p = 0.01) within responding patients. Cellular neighbourhoods analysis indicated that both macrophages (p = 0.003) and effector CD4 + T cells (p = 0.01) in mixed tumour neighbourhoods, as well as CD8 + T cells (p = 0.03) in HLADR + tumour neighbourhoods were associated with favorable clinical response. Evaluation of the inferred regulatory functions between immune cells relative to the tumour suggested that macrophages exhibit an immunosuppressive phenotype against both CD4 + and CD8 + T cells, and that this association scores more highly in ICI refractory patients. These spatial patterns are associated with overall survival in addition to ICI response and may thus indicate features for the functional understanding of the tumour microenvironment.