Advances in radiomics for predicting and managing xerostomia following radiotherapy: A systematic review.
Authors
Affiliations (5)
Affiliations (5)
- Department of Medical Physics, School of Medicine, Tabriz University of Medical Sciences, Tabriz, Iran.
- Department of Radio Oncology, School of Medicine, Cancer Prevention Research Center, Sayed Al-Shohada Hospital, Isfahan, Iran.
- Department of Medical Physics, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran.
- Medical Radiation Sciences Research Center, Department of Medical Physics, School of Medicine, Tabriz University of Medical Sciences, Tabriz, Iran. Electronic address: [email protected].
- Medical Radiation Sciences Research Center, Department of Radiology, Faculty of Allied Medical Sciences, Tabriz University of Medical Science, Tabriz, Iran. Electronic address: [email protected].
Abstract
Radiomics has emerged as a promising approach for predicting radiotherapy (RT)- induced xerostomia in head and neck cancer (HNC) patients, potentially enabling more personalized treatment strategies. This systematic review was conducted in accordance with PRISMA guidelines and protocol registered in PROSPERO (CRD420251039081). A comprehensive search of PubMed, Scopus, Web of Science, and the Cochrane Library was done between January 2010 and April 2025. Also, quality assessment of included studies was performed using the radiomics quality score (RQS) tool. This systematic review included 32 eligible studies with 4,167 HNC patients. The RQS ranged from 7 to 33 out of 36, with a mean of 15.13 (42.0 %). A dose-dependent relationship between radiation dose and xerostomia severity was observed. V<sub>30</sub> > 50 % and V<sub>40</sub> > 60 % doses to the parotids were associated with moderate-to-severe xerostomia (Grades 2-3), affecting 50 % of patients in some studies. IMRT still resulted in moderate-to-severe xerostomia when these dose thresholds were exceeded. Additionally, submandibular glands were also critical, especially with V<sub>40</sub> > 60 % doses. Delta-radiomics refers to the analysis of changes in radiomic features over time, typically before and after treatment, to assess tissue response. Delta-radiomics outperformed static radiomics most of studies, particularly with MRI or MVCT. The highest performance was reported with an AUC of 0.97 for a CT + MRI ensemble machine learning model, and an R<sup>2</sup> of 0.98 for a delta-radiomics MRI model. Radiomics-based models, particularly those using delta features and multimodal imaging, show high potential for accurate xerostomia prediction in HNC.