Related Experiment Video
Updated: Nov 11, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Decoding and Systematization of Medical Imaging Features of Multiple Human Malignancies
Lu Wang1, Zhaoyu Liu1, Jiayi Xie1
1School of Medical Informatics, China Medical University, Shenyang, Liaoning, China (L.W., M.Y., J.S.); Department of Radiology, Shenjing Hospital of China Medical University, Shenyang, Liaoning, China (Z.L.); Department of Radiology, China Medical University, Shenyang, Liaoning, China (J.X., Y.C., X.Z., Z.Y.); Department of Electric and Computer Engineering, University of Texas-El Paso, El Paso, Tex (W.Q.); CAS Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing, China (J.T.); and Department of Radiology, Stanford University School of Medicine, 1201 Welch Rd Lucas Center PS055, Palo Alto, CA 94305 (K.Y., J.S.).
This study summarizes medical imaging features for human malignancies, establishing a scientific basis for selecting features in future cancer research. It identifies patterns in texture features and their correlation with gene mutations, aiding in more credible oncology studies.
Area of Science:
- Medical Imaging
- Oncology
- Bioinformatics
Background:
- Medical imaging features are crucial for cancer diagnosis and research.
- Previous studies have reported numerous imaging features for human malignancies.
- A need exists to consolidate and analyze these features for more robust future studies.
Purpose of the Study:
- To systematically summarize reported medical imaging features across human malignancies.
- To establish a scientific foundation for credible imaging feature selection in oncology.
- To identify patterns and rules for applying imaging features in malignancy analysis.
Main Methods:
- A comprehensive literature search was conducted in PubMed (up to March 2018).
- Included studies were quality-assessed using the Newcastle-Ottawa scale.
- Unsupervised hierarchical clustering and meta-feature constructs were employed for analysis.
- CT images of 1000 non-small cell lung cancer patients were analyzed for texture feature distribution.
Main Results:
- 5026 imaging features from 930 articles across 20 human body parts were collated.
- A correlation atlas was developed to outline general rules for applying imaging features.
- Patterns in value distributions of common texture features were identified across malignancies.
- The run length imaging feature showed consistency with gene mutational signature 1B expression in human cancer.
Conclusions:
- This study provides a basis for more credible imaging feature selection in oncology.
- The findings can help reduce bias and redundancy in future medical imaging research.
- The identified patterns and correlations offer insights into malignancy characteristics.
More Related Videos
12:23Multi-modal Imaging of Angiogenesis in a Nude Rat Model of Breast Cancer Bone Metastasis Using Magnetic Resonance Imaging, Volumetric Computed Tomography and Ultrasound
Published on: August 14, 2012
10:37A Multimodal Imaging Framework to Advance Phenotyping of Living Label-free Breast Cancer Cells
Published on: August 22, 2025
