Volker BrunsThis gallery illustrates for a range of different tissue types and IHC markers the cell analysis quality of MIKAIA‘s IHC Cell Detection App. The app first unmixes the stain components, then uses an AI to delineate cells, then computes the stains’ intensity per cell and finally classifies the cells into negative / positive (or low / medium / strong positive) based on the marker intensity.
Cells with a too low intensity, too small or large size, black color, or cells that are located in an “ignore” annotations can optionally be filtered out. Cells can also be grouped by ROI, e.g., the AI Author App can be used to recognize tumor vs non-tumor areas. The IHC Cell Detection App can then either be used to only analyze cells in the previously detected tumor region or it can analyze cells everywhere, but separate cell types based on what ROI (e.g., tumor or non-tumor) they are located in.
The IHC Cell Detection App also includes post-processing functionality, e.g., it can search hotspots, group positive cells into clusters, or create a heatmap. Additional apps offer further functionality, e.g., the Proximity Analysis App can be used to compute distances from cells to the tumor interface or between cell types.
Add comment