Cytoplasm and/or membrane staining

[LisaRainey]

Hi,

I've started using QuPath and my experience in histopathology is very limited. Part of my PhD is to look at particular markers within tumour at the cytoplasm and/or membrane. 

I've uploaded a image of my chosen marker. Staining DAB/Haematoxylin. I'm interested in cytoplasmic staining and where possible apical membrane staining within glandular structures. 

How would I go about measuring either membrane or cytoplasm using QuPath?

Any help is very much appreciated!

Regards,

Lisa

[micros...@gmail.com]

I would start with using Optical Density during Cell detection (or Positive Cell Detection) in order to get an overview of your positive/negative ratio of DAB nuclei.

More on that here: https://petebankhead.github.io/qupath/tips/2018/03/22/setting-positive.html

and if you haven't, there is nice overview of the process here: https://github.com/qupath/qupath/wiki/Getting-started

[LisaRainey]

This is a better image

[LisaRainey]

Thanks for the info, I'll give that a go.

I'm very new to digital analysis, so trying to learning quickly!

[micros...@gmail.com]

In that case it would be best if you can use hematoxylin detection, so refining your color vectors will probably be key.  https://github.com/qupath/qupath/wiki/Preprocessing

Also the cytoplasms are very small, so you will want a Cell expansion that is equally small so that "mean cytoplasmic DAB" values are not diluted with white-space.

[LisaRainey]

Ok, thank you for your help.

I'll get reading and play around with the images I have.

Appreciate all the help,

Lisa

[Pete]

Hi Lisa,

These images look pretty challenging, but a lot depends on what you want to do with them.

To learn the ideas and software, I think it's best to work through some easier examples e.g. with a clear nuclear staining; microscopyra already linked you to the wiki, and it should be possible to work through the earlier sections there fairly quickly.

For your application, the first thing to define exactly what output(s) you want… e.g. area of brown staining, number of positive cells (in both cases, normalized to what?).  You could have a few outputs, some you’d ideally like and some that are more crude but probably easier to calculate.  I assume you want to show a difference, so if you could put screenshots of examples where you can clearly see the difference side-by-side then that could help.

If your number of images is manageable, I’m also very much in favour of doing the analysis manually first. This could mean coming up with some ‘score’ for each image, or possibly drawing around / counting relevant structures. That might be all you need in the end, and at least gives a backup plan if finding a more automated method of analysis that works across all your images proves too troublesome (it could take a while, and I don’t know if your PhD is more focussed on the methods of the analysis or on interpreting the results). It can be quite a lot of work, but not necessarily more work than trying to come up with an automated system. And if you do figure out a good way to apply image analysis, then you’ve got your manual scores for comparison to help validate your method (or your eyes…).

Regards,

Pete