Radiogenomic biomarkers

Medical images contain valuable information beyond the interpretation of human eyes. Radiomics is an emerging discipline to extract large numbers of quantitative features from various medical imaging modalities using computational or statistical approaches to develop supportive tool for clinical decision. Radiogenomics is the extension of radiomics through the correlation/comparison or combination of genetic and radiomic data. Radiogenomics may play an important role in providing imaging surrogates which are correlated with genetic expression, thereby serving as a substitute for genetic testing.

Before radiogenomics can become a useful clinical tool, many challenges and hurdles have to be overcome. Many studies have shown that imaging parameters, different scanners/platforms may affect the results tremendously. Moreover, imaging processing, imaging segmentation, imaging features extraction methods, and statistical approach all contribute to variation of study results.

For this special collection of "Radiogenomic biomarkers," we cordially invite manuscripts from different disciplines. High quality manuscripts related to disease diagnosis, prediction of treatment response, and assessment of disease risk and outcome are encouraged. Manuscripts outside of these topics which are within the scope of translational research are also welcome.

Authors are encouraged to include any or all of the following important elements:

1. Detailed methodologies, including: acquisition parameters of the raw data; how the regions of interest are selected and segmented; methods/software to extract radiomics features; statistical learning methods; and a framework proposal for robust radiomics analysis.

2. Robust methods for reliable and reproducible automatic segmentations of various regions of interest.

3. Justification of why the specific features extraction method and statistical model are used.

4. Use of the Imaging Biomarkers Standardization Initiative (IBSI), which standardizes radiomics feature extraction using different toolboxes and facilitates radiomic features interchangeability across platforms.

5. Evidence that quantitative features extracted from imagery are stable and reproducible.

6. Reasonably large sample size and internal and external validation datasets, which show acceptable results.

7. Correlation of radiomic results with genomic data or interpreted on biological/molecular basis.


  • Dr. Jeon-Hor Chen

    Dr. Jeon-Hor Chen, affiliated with the University of California, Irvine, is a research radiologist and an Associate Editor for Biomarker Research. Dr. Chen has been devoted to various Radiogenomics and deep learning studies for years.


Articles will be displayed here once they are published.