Beginners guide to Ki-67

KI67-tumor spheroid-20x-1.tif (14.3 MB)
Hi, I have a number of images like the one above and would like to automatically count the number of KI-67 stained nuclei and the total number of nuclei in this breast tumor spheroid.
I know that this is a simple analysis and many people have done this before me I would like to know the simplest and most accurate way to do it.

Someone else might be better able to help with FIJI, but this sort of project is almost perfect for QuPath if you ever want to give that a try. It only took a few minutes for me to:
1 Create a thresholder (Classify->Pixel classification->Create thresholder) to identify the tissue (your background is high because the image is dim, so there may be some improvements that could be made on the image acquisition end)

2. Use the thresholder to create an object for the organoid and exclude smaller objects.

3. Set the stain vectors as best I could given the single image (Analyze->Preprocessing->Estimate stain vectors)
4. Analyze->Cell detection->Positive cell detection

More information about each of these steps is available on

And there is a youtube video list from an older version of QuPath (but still very similar) showing Ki67 analysis here:

Once you have completed the steps, you can use the Workflow tab to create a macro to run across a whole project of images, and Measure->Export measurements to generate a spreadsheet. The best part is that you can come back at any time and inspect the results of any given analysis within the image itself, and generate some great visuals.

1 Like