Comparative Evaluation of Feature Selection Algorithms for Predicting Radiation-Induced Complications in Head and Neck Cancer Using Multimodal CT and MRI Radiomics
4 affiliations
- 1Department of Medical Physics, School of Medicine, Iran University of Medical Sciences, Tehran 1449614535, Iran
- 2Department of Biomedical Engineering and Medical Physics, School of Medicine, Shahid Beheshti University of Medical Sciences, Tehran 1985717443, Iran
- 3Department of Radiation Oncology, Winship Cancer Institute, Emory University, Atlanta, GA 30322, USA
- 4Department of Radiation Oncology, Northside Hospital, Atlanta, GA 30342, USA
Abstract
Head and neck cancer (HNC) presents a significant challenge due to its anatomic complexity and treatment-related complications such as sticky saliva and xerostomia, which critically impair patients’ quality of life. Recent advancements in imaging technologies, using the concept of radiomics, offer promising approaches for predicting these complications and ultimately reducing their incidence. In this regard, robust feature selection is critical in radiomics-based predictive modeling due to the high dimensionality, redundancy, and instability of extracted imaging biomarkers. This study aims to evaluate nine distinct feature selection algorithms—Boruta, Kruskal, Recursive Maximum Relevance minimum redundancy (R-MRmr), Relief, Recursive Feature Elimination (RFE), Forward Selection (FS), Backward Elimination (BE), Least Absolute Shrinkage and Selection Operator (LASSO), and Principal Component Analysis (PCA)—as a feature-transformation comparator for predicting sticky saliva and xerostomia in HNC patients using clinical, dosimetric, MRI and CT radiomics data. The study involved 250 HNC patients who underwent radiotherapy from three different institutions. From MRI and CT images, a total of 321 radiomic features (107 from each CT, T1, and T2-weighted MRI) along with 20 clinical and dosimetric features were extracted. The predictive robustness of feature selection methods was further evaluated across five machine-learning classifiers, including K-nearest neighbors (KNN), support vector machine (SVM), random forest (RF), logistic regression (LR), and eXtreme Gradient Boosting (XGBoost), were initially benchmarked using a common feature selection procedure, after which SVM was selected as the reference classifier for systematic comparison of the feature-reduction approaches. Model performance was assessed using metrics objectives such as Area Under the Curve (AUC), accuracy, sensitivity, and specificity. Among the evaluated feature selection approaches, R-MRmr and Relief generally yielded better discrimination, although performance varied according to toxicity endpoint and imaging modality. The highest AUCs were 0.85 for early sticky saliva (R-MRmr, T1-weighted MRI), 0.96 for early xerostomia (R-MRmr, T1-weighted MRI), 0.65 for late sticky saliva (Relief, combined CT+T1+T2), and 0.88 for late xerostomia (Relief, T1-weighted MRI). T1-weighted MRI frequently provided better individual-modality performance, whereas combining all three imaging modalities did not consistently improve discrimination. These findings demonstrate that the choice of feature selection strategy and imaging modality can materially influence radiomics-based toxicity prediction.
Related
- Predicting risk of progression of early to late AMD in the aging eye through imaging and multimodal evaluation: PRIME study protocol
- Comparative evaluation of previously reported contralateral testicular length cutoffs for predicting nonviable testes in boys with unilateral nonpalpable testis
- Comparative evaluation of fractal analysis and canny edge-detection algorithms for the quantification and classification of bone trabecular patterns in cone-beam computed tomography
- Nationwide Implementation of Multimodal Prehabilitation and Complications After Colorectal Cancer Surgery
- Predicting complications after hepatectomy for hepatocellular carcinoma: a narrative review focused on liver failure
BibTeX
@article{khajetash2026comparative,
title = {{Comparative Evaluation of Feature Selection Algorithms for Predicting Radiation-Induced Complications in Head and Neck Cancer Using Multimodal CT and MRI Radiomics}},
author = {Benyamin Khajetash and Mohammad Reza Deevband and William A. Stokes and Meysam Tavakoli},
journal = {Cancers},
year = {2026},
month = oct,
doi = {10.3390/cancers18193200},
url = {https://europepmc.org/article/PMC/PMC13664040},
note = {CC BY},
}
RIS
TY - JOUR
TI - Comparative Evaluation of Feature Selection Algorithms for Predicting Radiation-Induced Complications in Head and Neck Cancer Using Multimodal CT and MRI Radiomics
AU - Benyamin Khajetash
AU - Mohammad Reza Deevband
AU - William A. Stokes
AU - Meysam Tavakoli
PY - 2026
DA - 2026/10/01
JO - Cancers
DO - 10.3390/cancers18193200
UR - https://europepmc.org/article/PMC/PMC13664040
AB - Head and neck cancer (HNC) presents a significant challenge due to its anatomic complexity and treatment-related complications such as sticky saliva and xerostomia, which critically impair patients’ quality of life. Recent advancements in imaging technologies, using the concept of radiomics, offer promising approaches for predicting these complications and ultimately reducing their incidence. In this regard, robust feature selection is critical in radiomics-based predictive modeling due to the high dimensionality, redundancy, and instability of extracted imaging biomarkers. This study aims to evaluate nine distinct feature selection algorithms—Boruta, Kruskal, Recursive Maximum Relevance minimum redundancy (R-MRmr), Relief, Recursive Feature Elimination (RFE), Forward Selection (FS), Backward Elimination (BE), Least Absolute Shrinkage and Selection Operator (LASSO), and Principal Component Analysis (PCA)—as a feature-transformation comparator for predicting sticky saliva and xerostomia in HNC patients using clinical, dosimetric, MRI and CT radiomics data. The study involved 250 HNC patients who underwent radiotherapy from three different institutions. From MRI and CT images, a total of 321 radiomic features (107 from each CT, T1, and T2-weighted MRI) along with 20 clinical and dosimetric features were extracted. The predictive robustness of feature selection methods was further evaluated across five machine-learning classifiers, including K-nearest neighbors (KNN), support vector machine (SVM), random forest (RF), logistic regression (LR), and eXtreme Gradient Boosting (XGBoost), were initially benchmarked using a common feature selection procedure, after which SVM was selected as the reference classifier for systematic comparison of the feature-reduction approaches. Model performance was assessed using metrics objectives such as Area Under the Curve (AUC), accuracy, sensitivity, and specificity. Among the evaluated feature selection approaches, R-MRmr and Relief generally yielded better discrimination, although performance varied according to toxicity endpoint and imaging modality. The highest AUCs were 0.85 for early sticky saliva (R-MRmr, T1-weighted MRI), 0.96 for early xerostomia (R-MRmr, T1-weighted MRI), 0.65 for late sticky saliva (Relief, combined CT+T1+T2), and 0.88 for late xerostomia (Relief, T1-weighted MRI). T1-weighted MRI frequently provided better individual-modality performance, whereas combining all three imaging modalities did not consistently improve discrimination. These findings demonstrate that the choice of feature selection strategy and imaging modality can materially influence radiomics-based toxicity prediction.
SN - 2072-6694
ER -