Fig. 1: Profiling of immune cell density and spatial relationships of the urothelial cancer tumor micro-environment by multiplex immunofluorescence.

A Biopsy samples from 24 patients from the NABUCCO trial were profiled using mIF. B Cell type classification by comparing antibody marker positivity. C Tissue segmentation into tumor and stroma regions by comparing the local densities of cancer cell marker positive and negative cells. D Immune cell density in the tumor and stroma compartments was calculated in each tissue compartment (tumor and stroma). E SRs were summarized using the 1-NN statistic studied from a reference cell type to a target cell type. The resulting 1-NN distances vector was studied using 2 approaches: modeling a Weibull distribution to the Probabilistic Density Function (PDF) (top), and using the cumulative distribution function (CDF) using the G-function. F Association of SR parameters with ICI response and comparison of the discriminative power between SR and density TME parameters. G Validation of associations between SR parameters and response identified in UC in an independent cohort of HNSCC tumors. Icons from panel A, B, F and G were adapted from bioIcons (cancerous-cell-1, lymphocytes-4, macrophage, t-lymphocyte, b-lymphocyte, fibroblast-1 licensed under CC-BY 3.0 Unported by Servier), flaticon.com (bladder icon, https://www.flaticon.com/free-icon/bladder_1453578; head neck icon, https://www.flaticon.com/free-icon/injection_4418017). TME: tumor micro-environment; SR: spatial relationship; mIF: multiplex immunofluorescence; ICI: immune checkpoint inhibitors; 1-NN: first nearest neighbor; PDF: probabilistic density function; CDF: cumulative density function; G-AUC-T: G-function evaluated at a threshold T; T: threshold; UC: urothelial cancer; HNSCC: head and neck squamous cell carcinoma. Source data are provided with this paper.