Related Experiment Video
Updated: May 5, 2026
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
1.5K
Computer-based image analysis system designed to differentiate between low-grade and high-grade laryngeal cancer
Konstantinos Ninos1, Spiros Kostopoulos, Konstantinos Sidiropoulos
1Department of Physics, School of Natural Sciences, University of Patras, Rio, Patras, Greece.
Analytical and Quantitative Cytopathology and Histopathology
|November 29, 2013
Summary
A new pattern recognition (PR) system accurately distinguishes low- and high-grade laryngeal cancer from histopathology images. This AI tool, utilizing p63 expression and GPU technology, achieved 90.9% overall accuracy, aiding physician diagnosis.
Area of Science:
- Oncology
- Biomedical Engineering
- Computer Science
Background:
- Laryngeal cancer grading is crucial for treatment and prognosis.
- Accurate differentiation between low- and high-grade tumors is essential for effective patient management.
- Histopathology images provide key diagnostic information, but manual analysis can be subjective.
Purpose of the Study:
- To develop an automated pattern recognition (PR) system for classifying laryngeal cancer as low- or high-grade.
- To utilize immunohistochemically stained histopathology images for p63 expression analysis.
- To enhance diagnostic accuracy and potentially improve patient survival rates.
Main Methods:
- A PR system was designed using 55 laryngeal cancer cases (21 low-grade, 34 high-grade).
- p63-expressed nuclei were automatically segmented from histopathology images.
- Fifty-two features (texture, shape, topology) were extracted and analyzed using a Probabilistic Neural Network classifier on a GPU with CUDA parallel programming.
Main Results:
- The PR system achieved 85.7% accuracy for low-grade and 94.1% for high-grade laryngeal cancer.
- Overall system accuracy was 90.9% using 7 selected features.
- Estimated accuracy for unseen cases was 80%.
Conclusions:
- Parallel processing and GPU technology enabled optimal PR system design.
- The system is designed for clinical use as a research tool and is adaptable for future case additions.
- The developed PR system shows promise for objective and efficient laryngeal cancer grading.
