Cartilage Proteoglycan Loss Analysis using QuPath

Hi, I was wondering if anyone has analysed cartilage proteoglycan loss on SafO stained sections using QuPath?

The type of images that I am trying to analyse are similar to this uploaded example. I want to find the proportion of red staining in a given ROI. Can I do thresholding or calculate stain intensity using QuPath? Thanks.

2 Likes

You might want to consider using the Simple thresholder Pete just mentioned in a previous post:


You will definitely want to set the stain vectors prior to doing this. For example, I set the stain vectors to the reddish and teal colors, though that could cause some problems with the purple objects. You may also want to consider the Pixel classifier if the Simple thresholder will not work for you.

In this case I used small ROIs and assigned new colors to the Hematoxylin and Eosin labels.
image
More on color vectors here

2 Likes

Oh, and as far as “stain intensity,” note that in general stain intensities in brightfield are not really quantitative/linear. So while it might be useful to set a threshold, double the threshold does not mean double the stain or source of the stain.

Also, ideally, stain vectors should be set from single stains.

Thankyou. Vector setting followed by simple thresholding worked out really well.

1 Like