This study demonstrates an automated spatial biology workflow combining GeoMx® Digital Spatial Profiling (DSP), AI-driven image analysis (Visiopharm Oncotopix® Discovery), and spatial whole-transcriptome profiling to investigate the relationship between hypoxia and immune infiltration in head and neck cancer (HNC).
The approach uses deep-learning models to automate segmention ofbiologically relevant regions of interest (ROIs) and areas of illumination (AOIs), reducing manual effort and selection bias while enabling scalable analysis of larger patient cohorts.
The workflow successfully generated hypoxia-based spatial gradients and tumor/stroma compartment-specific transcriptomic data, revealing associations between hypoxic regions, gene-expression programs, and immune cell distribution within the tumor microenvironment.
