Phenoptrreports

Whole slide scans were imaged on the Vectra Polaris Automated Quantitative Pathology Imaging System (Akoya) using the 20x objective. The images were analyzed with inForm software (v2.4.8, Akoya) to unmix adjacent fluorochromes, subtract autofluorescence, segment the tissue into tumor, vascular and necrotic regions, segment the cells into nuclear compartments, and to phenotype the cells according to morphology and cell marker expression. For the myeloid panel independent projects were created to phenotype each cellular marker, then merged, consolidated, and analyzed in R Studio using Phenoptr Reports (Akoya Biosciences).

To assess the topological arrangement of immune cells in bladder cancer microenvironment and cell-to-cell interactions, spatial metrics between cells were calculated using phenoptrReports (add-ins for R Studio from Akoya Biosciences). In particular, the nearest neighbor analysis calculates the average distance between each feature’s centroid and its nearest neighbors’ centroid location, and it was used to analyse the mean distance between different cell subtypes. Moreover, the count within analysis was employed to calculate for each pair of phenotypes, the number of cells with a distinct phenotype having a cell of another phenotype within a specified radius. Since a distance radius of 20-25 μm between two cell subtypes is considered indicative of an enhanced probability for cell-to-cell contact [13 (link)], we calculated the number of reference cells that are present within a 20-25 μm radius from a cell with a different phenotype, and normalized for the total number of reference cells expressed as the percentage among the total number of reference cells. This methodology was used in order to not only consider the absolute number of cell-to-cell contacts, which might be viewed as an epiphenomenon of cellular density, but also to correct for the different number of immune cells present in the bladder TME.

Methods for multispectral imaging have previously been published and the methods herein are described in brief^{33 (link),39 (link)}. Following the manufacturer’s protocol, FFPE slides were sequentially stained using Opal IHC Multiplex Assay (PerkinElmer, Waltham, MA) by the Human Immune Monitoring Shared Resource at the University of Colorado Anschutz Medical Campus. The Vectra Polaris Imaging System (PerkinElmer) scanned the whole slide. The panel used for the Vectra Polaris were as follows in sequential order: CD11b, CD64, EpCAM. CD11c, B220, CD8, F4/80. Approximately 5 regions of interest (ROIs) were evaluated per tumor. Images were analyzed using inForm Tissue Analysis Software (v2.4.8, Akoya, Menlo Park, CA) to un-mix fluorochromes, remove autofluorescence, segment tissue and cells, and phenotype cells. Data analysis was performed as previously described by our group^{21 (link),39 (link)}. Briefly, data acquired from inForm was analyzed using the R package Akoya Biosciences phenoptrReports. The count_phenotypes function was used to aggregate phenotype counts for each slide. Data are presented as cell counts.