Related Experiment Video
Updated: Dec 18, 2025
10:17
Guidelines and Experience Using Imaging Biomarker Explorer IBEX for Radiomics
Published on: January 8, 2018
13.6K
Standardization in Quantitative Imaging: A Multicenter Comparison of Radiomic Features from Different Software
M McNitt-Gray1, S Napel2, A Jaggi2
1David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA.
Tomography (Ann Arbor, Mich.)
|June 18, 2020
Summary
Radiomic feature computation showed good agreement for five of nine features across different software. Standardization is valuable, but some radiomic features require clearer definitions for consistent clinical use.
Area of Science:
- Quantitative Imaging
- Radiomics
- Medical Imaging Analysis
Background:
- Radiomic features are increasingly used in clinical applications.
- Assessing the consistency of radiomic feature computation is crucial for reliable clinical translation.
- Variability in feature extraction can impact diagnostic accuracy and treatment response prediction.
Purpose of the Study:
- To evaluate the agreement of radiomic features computed by different software packages and research groups.
- To identify radiomic features with high inter-software and inter-group consistency.
- To highlight areas where feature definition standardization is needed for robust clinical application.
Main Methods:
- Nine common radiomic features (morphology, intensity, shape, texture) were selected.
- Standardized feature definitions and common image datasets (3D digital reference objects and patient scans) were used.
- The coefficient of variation (CV) was calculated across 13 software/group results for each feature.
Main Results:
- Five out of nine features demonstrated excellent agreement (CV < 1%).
- One feature showed moderate agreement (CV < 10%).
- Three features exhibited significant variation (CV ≥ 10%) despite harmonization efforts.
Conclusions:
- Standardizing radiomic feature definitions is essential for consistent results.
- Further clarification of definitions is needed for specific radiomic features to improve reliability.
- High agreement for some features supports their potential for clinical use with careful implementation.
