Abstract 715: Robust segmentation-free stain quality concordance metrics in the SpaceIQ™ multi-omic analysis platform.

Key Points

• This research aims to develop a quantitative method for assessing concordance between IHC and mIF stains in a cancer dataset.
• Utilized a pan-cancer dataset with a 4-plex mIF panel and matching IHC sections.
• Co-registered images using global rigid and non-rigid transformations.
• Isolated IHC stain channels with stain-matrix-based deconvolution.
• Generated tissue masks using Otsu thresholding and morphology operations.
• Assessed stain concordance within divided image tiles using statistical tests.
• Concordance between mIF and IHC varies by marker, with CD8 showing highest agreement and FoxP3 showing lowest.
• Spatial analysis revealed regions of both high and low concordance in staining.
• Quantitative findings correlated with expert visual review.