Currently, our annotation software supports using the Notes & Stories Panel feature in the visualizer to manually annotate which areas of each tissue are tumor and which are stroma. Once you have collected all of the masks into a single story, you will need to follow the instructions in Creating Cell Annotations from Story/ROIs to export the notes for downstream analysis.
When creating this type of annotation, you can draw annotations indicating only one class or the other (e.g. only “Tumor”) and then use the Tissue Foreground Detection tool to annotate the entire tissue area—tissue foreground areas that are not annotated can be marked as the other class (e.g. “Stroma”) by default.
We have also developed a tool for internal use that can generate tissue masks based on the areas uniformly expressing large amounts of one or several biomarkers in the image (e.g. PanCK & EpCAM for tumors of epithelial origin). We are working on developing a version of this tool that will be available for use by subscribers to our Workbench platform. In the meantime, when you purchase Tumor/Stroma mask annotation services from our team, we will be making use of this tool.
You can also provide binary mask files generated in any analysis software you prefer to use for tissue annotation. All that is required in terms of formatting is:
1 for pixels annotated as “Tumor” and 0 for pixels annotated as “Stroma”.