Getting started with QuPath v0.3.0

Overview

QuPath v0.3.0 isn’t quite released yet, but the first (only?) Release Candidate has been made available for this workshop.

Here, I ask you to quickly recap what you’ve (hopefully) already done with QuPath, as a prelude to exploring something new.

You can download QuPath v0.3.0-rc1 here.

The main QuPath ReadTheDocs is still for v0.2, but you can already get the docs for v0.3 at https://qupath.readthedocs.io/en/latest/.

One of the biggest changes (since yesterday) is that there are now video tutorials. Therefore if you had any trouble with the pre-workshop material, these could help:

• Projects
• Separating stains
• Detecting tissue
• Measuring areas
• Cell detection
• Cell classification

But the main purpose of this part of the workshop is to explore an entirely new feature Density maps.

Your challenge

• Create a new project and add the workshop images to it
• Open OS-2.ndpi
• Run Positive cell detection and train an object classifier to distinguish between tumor & non-tumor cells
  • To speed things up, you can focus on the middle part of the image and use a Requested pixel size of 1 µm
• Find the circle with area 0.785 mm² that contains the highest number of positive tumour cells
• Create a screenshot displaying the circle, and count how many positive tumor cells it contains

To do this, you should follow the tutorial at https://qupath.readthedocs.io/en/latest/docs/tutorials/density_maps.html and find the combination of parameters needed to get the answer.