Cell segmentation

Image Analysis
#1

Hi all,

Would like to know if it’s possible to perform a cell segmentation (nucleus and membrane) using the QuPath? I’d like to do it for a double staining…

Thanks!
Bianca.

#2

You could use the multiplex guide in the readthedocs page to attempt to classify cells as single or double positive (building a composite classifier, works similarly for brightfield and fluorescence).

  1. If the stains overlap, they will not be as separable in brightfield as in fluorescent experiments. The stains cover each other and interfere.
  2. If you are using either StarDist or the built in Cell Detection, the cytoplasmic expansion is blind, so mean/median values might not do what you want as far as detecting positive cells.
    More on what can go wrong here: QuPath-Accurate cytoplasmic stain measurements