Automatic Tumor Infiltration Detection Algorithm
Description
These are the pieces necessary to replicate our automatic tumor detection algorithm as described in my Bachelor's Thesis, High-Throughput Automated Cell Classification for Measuring the Connection Between Tumor Infiltration and Radiation Response (June 2025).
Files
Steps to reproduce
To carry out our immune infiltration detection protocol in QuPath, it is necessary to 1) Create an annotation encompassing the entire tissue slice you want to analyze; 2) Exclude any recognizable epithelium from the highlighted area 3) Select the annotation and run 'Positive Cell Detection' with the parameters shown in the file named 'PSD OD Sum Parameters.PNG' 4) Copy the 'classifiers' file into the project folder 5) Select the annotation and choose 'Load Object Classifier' 6) Select the object classifier 'TumorMucosaV3.json' 7) Run the object classifier You can now view tumor infiltration counts, percentages, and exports the measurements for this slice :)
Institutions
- Erasmus Universiteit Rotterdam Faculteit der Sociale Wetenschappen