Related Experiment Video
Updated: Mar 16, 2026
08:40
Quantitation of Protein Expression and Co-localization Using Multiplexed Immuno-histochemical Staining and Multispectral Imaging
Published on: April 8, 2016
13.5K
A High-Resolution Tile-Based Approach for Classifying Biological Regions in Whole-Slide Histopathological Images
R A Hoffman1, S Kothari2, J H Phan1
1Department of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA, USA.
IFMBE Proceedings
|August 18, 2016
Summary
This study presents a computational method for identifying biological regions in whole slide images (WSIs) for improved cancer diagnosis. The automated system accurately classifies tissue types, aiding in cancer prognosis.
Area of Science:
- Digital pathology
- Computational oncology
- Image analysis
Background:
- Whole slide images (WSIs) offer potential for cancer diagnosis and prognosis.
- Automated identification of biological regions (tumor, stroma, necrotic tissue) in WSIs remains a challenge.
Purpose of the Study:
- To develop and validate a computational method for classifying WSI tiles into distinct biological regions.
- To assess the accuracy of automated region identification against whole-slide level ground truth.
Main Methods:
- Extraction of 461 image features from 512x512 pixel WSI tiles.
- Optimization of tile-level prediction models using nested cross-validation on a small annotated set.
- Validation against a large dataset (1.7x10^6 tiles) with whole-slide level ground truth.
Main Results:
- Significant correlation (p < 0.001) between predicted and ground truth biological region prevalences for 8 out of 9 cases.
- Demonstrated accuracy of the automated system in quantifying tissue region distribution.
Conclusions:
- The developed computational method effectively classifies biological regions within WSIs.
- This approach holds promise for enhancing automated analysis in digital pathology and cancer research.
