Cell detection in tumor

Hello,

We run in trouble detecting cells in tumor area.
Is this possible with QuPath (see attachment)?

Thanks

In most of those cells you could lower the hematoxylin threshold, but in the top center area, you have lost your nuclear signal so there is essentially no way to segement the nuclei of the cells. For staining that dense your best bet is to use a secondary that is fluorescent.

Some other options including using a non DAB stain, like yellow, which could let you see the overlap with your red cells, and likely still detect the blue nuclei. IF would still be better, but a lighter (cytokeratin/vimentin/etc) could work if that is not an option. If you just want to see the red cells (TILs?) within the tumor area, many red dyes are autofluorescent, and you might be able to image in both brightfield and IF at the same time to get counts that way. Some slide scanners or widefield scopes can do this.

3 Likes

What @Research_Associate said. Or if you still want to use DAB, I would optimize the stain to significantly decrease the intensity of the DAB, so that it does not obscure the nuclei. The first and easiest thing I would try is to titrate the primary antibody. This will usually get you where you want to be. Many IHC stains using DAB are optimized to look like this because the original goal was not image analysis and this was pleasing to the eye of the user. Once you start using the stain for image analysis, you often need to re-optimize. But if you are doing a double stain, that would be another reason to save yourself this trouble and just switch to fluorescence. Brown and red double chromogenic staining, if there is any colocalization, doesn’t work well or at all for image analysis in our experience. If the two signals are in two distinct and not overlapping compartments or cell populations, then it’s usually ok. In this case it seems like it would be challenging.

2 Likes

Many thanks to both of you for this fantastic inputs,
Unfortunately, we don’t own a fluorescence scanner. I will ask in our laboratory what is possible with non-DAB staining and let you know if we were able to do it.

1 Like