Machine learning helps identifying volume-confounding effects in radiomics. (March 2020)

• Record Type: Journal Article
• Authors: Traverso, Alberto; Kazmierski, Michal; Zhovannik, Ivan; Welch, Mattea; Wee, Leonard; Jaffray, David; Dekker, Andre; Hope, Andrew
• Is Part Of: Physica medica. Volume 71(2020)
• Page Start: 24
• Page End: 30
• Publication Date: 2020-03
• Subjects: Radiomics -- Machine learning -- Predictions -- Lung -- Head and neck; Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2020.02.010
• Languages: English
• ISSNs: 1120-1797

Abstract

Highlights: We investigated the presence of volume-confounding effects in radiomics features. Our results show that many radiomics features are volume-confounded. We offered the radiomics community a workflow for introducing safeguards in signature developments. Abstract: Purpose: Highlighting the risk of biases in radiomics-based models will help improve their quality and increase usage as decision support systems in the clinic. In this study we use machine learning-based methods to identify the presence of volume-confounding effects in radiomics features. Methods 841 radiomics features were extracted from two retrospective publicly available datasets of lung and head neck cancers using open source software. Unsupervised hierarchical clustering and principal component analysis (PCA) identified relations between radiomics and clinical outcomes (overall survival). Bootstrapping techniques with logistic regression verified features' prognostic power and robustness. Results Over 80% of the features had large pairwise correlations. Nearly 30% of the features presented strong correlations with tumor volume. Using volume-independent features for clustering and PCA did not allow risk stratification of patients. Clinical predictors outperformed radiomics features in bootstrapping and logistic regression. Conclusions The adoption of safeguards in radiomics is imperative to improve the quality of radiomics studies. We proposed machine learning (ML) – based methods for robust radiomics signatures development.