Related Experiment Video
Updated: Mar 11, 2026
06:45
Automated Joint Space Detection Improves Bone Segmentation Accuracy
Published on: November 28, 2025
234
COMPUTER AIDED IMAGE SEGMENTATION AND CLASSIFICATION FOR VIABLE AND NON-VIABLE TUMOR IDENTIFICATION IN OSTEOSARCOMA.
Harish Babu Arunachalam1, Rashika Mishra, Bogdan Armaselu
1Department of Computer Science, University of Texas at Dallas, Richardson, TX, USA, harishb@utdallas.edu.
Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing
|November 30, 2016
Summary
This study introduces an automated digital image analysis method to accurately assess osteosarcoma treatment response by quantifying tumor necrosis in histopathology slides. The new technique significantly improves accuracy and reduces pathologist assessment time.
Area of Science:
- Digital pathology
- Computational oncology
- Medical image analysis
Background:
- Osteosarcoma treatment response is manually assessed via H&E slides, which is time-consuming and prone to bias.
- Accurate assessment of tumor necrosis is crucial for evaluating treatment efficacy in osteosarcoma patients.
Purpose of the Study:
- To develop and validate an automated digital image analysis technique for segmenting and classifying tumor regions in osteosarcoma histopathology Whole Slide Images (WSIs).
- To improve the accuracy and efficiency of assessing tumor necrosis for treatment response evaluation.
Main Methods:
- Whole Slide Images (WSIs) were processed using a combination of pixel-based and object-based image segmentation techniques.
- K-Means clustering and multi-threshold Otsu segmentation were employed for tumor isolation and classification of viable/non-viable regions.
- Flood-fill algorithm was used to cluster pixels into cellular objects for further analysis.
Main Results:
- The automated method achieved 100% accuracy in identifying viable tumor and coagulative necrosis.
- An accuracy of approximately 90% was obtained for fibrosis and acellular/hypocellular tumor osteoid identification.
- The approach demonstrated high accuracy and consistency across sampled osteosarcoma datasets.
Conclusions:
- The developed digital image analysis technique offers a comprehensive and accurate solution for classifying viable and non-viable tumor regions in osteosarcoma.
- This automated approach is expected to significantly reduce inter-observer variability and assessment time for pathologists.
- The software has the potential to enhance the evaluation of cancer treatment response in osteosarcoma patients.
