Fig. 3: Spatial distribution of immune cell subsets using CODEX.

a Multiplex IF of immune cell aggregate for a representative partial response (PR) and progressive disease (PD) case showing staining for CD4, CD8, CD20, CD68, Ki-67, Pan CK and CD31 (PR, n = 4; PD, n = 4) (b) Heatmap of the average expression of 25 markers in the different cell clusters is shown. (c) The UMAP plots for PR and PD cases shows a total of 15 cell clusters that were validated by manual inspection of multiplexed immunostains from a 25 markers panel staining using CODEX. Color- code correspond to those for each cluster in the fig. (b). d Pie graph shows percentages of most abundant cell clusters in PR and PD cases. (e, f) Neighborhood (CN) are defined based on the presence of the 15 validated clusters. A total of six CN are identified. The stacked bar graph shows distribution of each cell neighborhood between PR (n = 4) and PD (n = 4) cases. The statistical test used is two-sided pairwise t-tests on the transformed proportions, comparing PR and PD cohorts. The resulting p-values were corrected for multiple testing by the Bonferroni method. g The bar plots show average cell distance to nearest B cells from different cell subsets when comparing 25 regions of interest from a total of eight cases (PR, n = 4; PD, n = 4), each dots represent each region of interest from all PR and PD cases. Box plots represent the median, interquartile range and the whiskers represent 1.5 x the upper and lower interquartile range values. The statistical test used is two-sided Welch Two Sample t-test on the average minimum distance metric between our two patient cohorts.