Related Experiment Video
Updated: Aug 5, 2025

06:53
Modeling Primary Bone Tumors and Bone Metastasis with Solid Tumor Graft Implantation into Bone
Published on: September 9, 2020
2.8K
Systematic Review of Tumor Segmentation Strategies for Bone Metastases
Iromi R Paranavithana1,2, David Stirling1, Montserrat Ros1
1Faculty of Engineering and Information Sciences, School of Electrical, Computer and Telecommunications Engineering, University of Wollongong, Wollongong, NSW 2522, Australia.
Cancers
|March 29, 2023
Summary
This review examined bone lesion segmentation methods, finding neural networks and CT imaging dominant. However, challenges in tumor boundary definition and manual correction hinder clinical application of these segmentation techniques.
Area of Science:
- Medical imaging analysis
- Oncology
- Radiology
Background:
- Accurate segmentation of bone lesions is crucial for differentiating benign from malignant conditions.
- Characterizing malignant bone lesions aids in treatment planning and patient management.
Purpose of the Study:
- To investigate segmentation approaches for bone metastases.
- To differentiate benign from malignant bone lesions.
- To characterize malignant bone lesions.
Main Methods:
- A systematic literature search was conducted across major databases (Scopus, PubMed, IEEE, MedLine, Web of Science).
- Included 77 original articles, 24 review articles, and 1 comparison paper published between January 2010 and March 2022.
- Followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
Main Results:
- Neural network-based approaches (58.44%) and CT-based imaging (50.65%) were the most common methods.
- A lack of a gold standard for tumor boundaries and the need for manual correction limit clinical translation.
- Only 24.67% of studies addressed clinical practice feasibility.
Conclusions:
- Combining anatomical and metabolic information in tumor segmentation shows promise.
- An optimal, universally applicable tumor segmentation method is still lacking.
- Data limitations and inherent difficulties require further research for robust clinical application.

