Related Experiment Video
Updated: Jul 27, 2025
07:32
Author Spotlight: Investigating Immune Cell Dynamics in the Tumor Microenvironment — Challenges and Innovations in Cancer Prognosis
Published on: April 12, 2024
1.4K
QuPath Algorithm Accurately Identifies MLH1-Deficient Inflammatory Bowel Disease-Associated Colorectal Cancers in a
Ross J Porter1,2, Shahida Din2, Peter Bankhead1,3
1Edinburgh Pathology, CRUK Scotland Centre, Institute of Genetics and Cancer (IGC), University of Edinburgh, Scotland EH4 2XU, UK.
Diagnostics (Basel, Switzerland)
|June 10, 2023
Summary
Automated image analysis using QuPath accurately identifies MLH1-deficient colorectal cancers in inflammatory bowel disease (IBD-CRC). This method improves efficiency and reduces variability in immunohistochemistry analysis for IBD-CRC diagnosis.
Area of Science:
- Pathology
- Computational Pathology
- Oncology
Background:
- Immunohistochemistry analysis for MLH1 deficiency in colorectal cancer is labor-intensive and prone to inter-observer variability.
- Accurate identification of MLH1-deficient inflammatory bowel disease-associated colorectal cancers (IBD-CRC) is crucial for patient stratification and treatment decisions.
Purpose of the Study:
- To train and validate QuPath, an open-source image analysis software, for automated identification of MLH1-deficient IBD-CRC.
- To assess the accuracy, sensitivity, and specificity of QuPath in classifying MLH1 expression and tissue histology.
Main Methods:
- A tissue microarray (n=162) with normal colon and IBD-CRC tissues was immunostained for MLH1.
- QuPath was trained on a subset (n=14) to differentiate MLH1 expression and tissue types (normal epithelium, tumor, immune infiltrates, stroma).
- The trained algorithm was applied to the entire tissue microarray, with results compared to manual review.
Main Results:
- QuPath correctly identified tissue histology and MLH1 expression in 73.74% of valid cases (73/99).
- The algorithm achieved 100% sensitivity and 98.25% specificity for identifying MLH1-deficient IBD-CRC in classified cores (n=74).
- High agreement (κ = 0.963) was observed between QuPath classification and manual review, indicating robust performance.
Conclusions:
- Automated image analysis with QuPath offers an efficient and accurate method for MLH1 expression analysis in IBD-CRC.
- This approach has the potential to streamline diagnostic workflows in pathology laboratories.
- QuPath-based analysis can reduce inter-observer variability and improve the consistency of IBD-CRC classification.
