Automated Single Cell Phenotyping of Time-of-Flight\nSecondary Ion Mass Spectrometry Tissue Images

Existing analytical\ntechniques are being improved or applied in\nnew ways to profile the tissue microenvironment (TME) to better understand\nthe role of cells in disease research. Fully understanding the complex\ninteractions between cells of many different types and functions is\noften slowed by the intense data analysis required. Multiplexed Ion\nBeam Imaging (MIBI) has been developed to simultaneously characterize\n50+ cell types and their functions within the TME with a subcellular\nspatial resolution, but this results in complex data sets that are\nchallenging to qualitatively analyze. Deep Learning (DL) techniques\nwere used to build the MIBIsight workflow, which can process images\ncontaining thousands of cells into easily digestible reports and plots\nto enable researchers to easily summarize data sets in a study and\nmake informed conclusions. Here we present the three types of DL models\nthat have been trained with annotated MIBI images that have been pathologist\nvalidated as well as the associated workflow for the evolution of\nraw mass spectral data into actionable reports and plots.