Comprehensive Analysis of Radiomic Datasets by RadAR

Area of Science:

• Biomedical imaging analysis
• Computational oncology
• Radiomics

Background:

• Quantitative analysis of biomedical images (radiomics) aids clinical decisions and patient stratification.
• Current radiomic workflows face challenges in processing high-dimensional datasets.
• A need exists for specialized tools to manage and analyze radiomic data.

Purpose of the Study:

• Introduce RadAR (Radiomics Analysis with R), a novel software for comprehensive radiomic feature analysis.
• Provide a tool for end-to-end processing of radiomic datasets, from import to visualization.
• Implement diverse statistical methods for robust radiomic data analysis.

Main Methods:

• Development of RadAR software using the R programming language.
• Integration of data import, feature processing, and visualization capabilities.
• Application of multiple statistical methods for high-dimensional data analysis.

Main Results:

• RadAR successfully processed entire radiomic datasets, including feature extraction and statistical analysis.
• Analysis of radiomic profiles from over 850 cancer patients using RadAR recapitulated expected findings.
• Demonstrated reliability and value of RadAR through validation on independent datasets.

Conclusions:

• RadAR is a reliable and valuable computational tool for the radiomics community.
• The software facilitates comprehensive analysis of high-dimensional radiomic datasets.
• RadAR supports improved clinical decision-making and patient stratification in oncology.