DIMPLE: An R package to quantify, visualize, and model spatial cellular interactions from multiplex imaging with distance matrices

A Spatial Omnibus Test (SPOT) for Spatial Proteomic Data

“…The advantage of the SPOT framework is that it is adaptable to the application and hypotheses of practitioners. For example, one could consider any measure of spatial cellular configuration, such as the mark connection function used by Vu et al (2022) or estimate the spatial intensity of cells and consider the Jensen−Shannon distance between density estimates as done in Masotti et al (2023) . One could consider spatial summary measures that accommodate inhomogeneity and could implement any outcome model depending on the clinical outcome-of-interest.…”

“…Our data analyses suggest that this range is reasonable, but one could expand beyond this. Other approaches for choosing the radii may incorporate biological knowledge, such as the size of the cell and prior knowledge about the “radii-of-influence” between cell types ( Masotti et al 2023 ). Consideration of additional radii can boost power if useful radii are included, but consideration of poor choices can lead to reduced power.…”

“…A challenge with these spatial measures is the choice of radius, t , for characterizing proximal relationships. This could be guided by clinical knowledge ( Barua et al 2018 , Masotti et al 2023 ), but there is no consensus or guideline across applications and hypotheses. Fixing the radius at one value may neglect clinically relevant spatial patterns observed at smaller or larger values of t .…”