A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion

Abstract Objectives Radiation-induced xerostomia is a common sequela in patients who undergo head and neck radiation therapy. This study aims to develop a three-dimensional deep learning model to predict xerostomia by fusing data from the gross tumor volume primary (GTVp) channel and parotid glands...

Full description

Saved in:
Bibliographic Details
Main Authors: Lin Lin, Yuchen Ren, Wanwei Jian, Geng Yang, Bailin Zhang, Lin Zhu, Wenhao Zhao, Haoyu Meng, Xuetao Wang, Qiang He
Format: Article
Language:English
Published: BMC 2025-07-01
Series:BMC Medical Imaging
Subjects:
Online Access:https://doi.org/10.1186/s12880-025-01848-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849738596205985792
author Lin Lin
Yuchen Ren
Wanwei Jian
Geng Yang
Bailin Zhang
Lin Zhu
Wenhao Zhao
Haoyu Meng
Xuetao Wang
Qiang He
author_facet Lin Lin
Yuchen Ren
Wanwei Jian
Geng Yang
Bailin Zhang
Lin Zhu
Wenhao Zhao
Haoyu Meng
Xuetao Wang
Qiang He
author_sort Lin Lin
collection DOAJ
description Abstract Objectives Radiation-induced xerostomia is a common sequela in patients who undergo head and neck radiation therapy. This study aims to develop a three-dimensional deep learning model to predict xerostomia by fusing data from the gross tumor volume primary (GTVp) channel and parotid glands (PGs) channel. Methods Retrospective data were collected from 180 head and neck cancer patients. Xerostomia was defined as xerostomia of grade ≥ 2 occurring in the 6th month of radiation therapy. The dataset was split into 137 cases (58.4% xerostomia, 41.6% non-xerostomia) for training and 43 (55.8% xerostomia, 44.2% non-xerostomia) for testing. XeroNet was composed of GNet, PNet, and a Naive Bayes decision fusion layer. GNet processed data from the GTVp channel (CT, dose distributions corresponding and the GTVp contours). PNet processed data from the PGs channel (CT, dose distributions and the PGs contours). The Naive Bayes decision fusion layer was used to integrate the results from GNet and PNet. Model performance was evaluated using accuracy, F-score, sensitivity, specificity, and area under the receiver operator characteristic curve (AUC). Results The proposed model achieved promising prediction results. The accuracy, AUC, F-score, sensitivity and specificity were 0.779, 0.858, 0.797, 0.777, and 0.782, respectively. Features extracted from the CT and dose distributions in the GTVp and PGs regions were used to construct machine learning models. However, the performance of these models was inferior to our method. Compared with recent studies on xerostomia prediction, our method also showed better performance. Conclusions The proposed model could effectively extract features from the GTVp and PGs channels, achieving good performance in xerostomia prediction.
format Article
id doaj-art-dd66bd39418e4a1b94b5fe01830c32cf
institution DOAJ
issn 1471-2342
language English
publishDate 2025-07-01
publisher BMC
record_format Article
series BMC Medical Imaging
spelling doaj-art-dd66bd39418e4a1b94b5fe01830c32cf2025-08-20T03:06:31ZengBMCBMC Medical Imaging1471-23422025-07-0125111110.1186/s12880-025-01848-1A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusionLin Lin0Yuchen Ren1Wanwei Jian2Geng Yang3Bailin Zhang4Lin Zhu5Wenhao Zhao6Haoyu Meng7Xuetao Wang8Qiang He9Guangdong Pharmaceutical UniversityThe Second Clinical College of Guangzhou University of Chinese MedicineThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineGuangdong Pharmaceutical UniversityThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineGuangdong Pharmaceutical UniversityThe Second Affiliated Hospital of Guangzhou University of Chinese MedicineAbstract Objectives Radiation-induced xerostomia is a common sequela in patients who undergo head and neck radiation therapy. This study aims to develop a three-dimensional deep learning model to predict xerostomia by fusing data from the gross tumor volume primary (GTVp) channel and parotid glands (PGs) channel. Methods Retrospective data were collected from 180 head and neck cancer patients. Xerostomia was defined as xerostomia of grade ≥ 2 occurring in the 6th month of radiation therapy. The dataset was split into 137 cases (58.4% xerostomia, 41.6% non-xerostomia) for training and 43 (55.8% xerostomia, 44.2% non-xerostomia) for testing. XeroNet was composed of GNet, PNet, and a Naive Bayes decision fusion layer. GNet processed data from the GTVp channel (CT, dose distributions corresponding and the GTVp contours). PNet processed data from the PGs channel (CT, dose distributions and the PGs contours). The Naive Bayes decision fusion layer was used to integrate the results from GNet and PNet. Model performance was evaluated using accuracy, F-score, sensitivity, specificity, and area under the receiver operator characteristic curve (AUC). Results The proposed model achieved promising prediction results. The accuracy, AUC, F-score, sensitivity and specificity were 0.779, 0.858, 0.797, 0.777, and 0.782, respectively. Features extracted from the CT and dose distributions in the GTVp and PGs regions were used to construct machine learning models. However, the performance of these models was inferior to our method. Compared with recent studies on xerostomia prediction, our method also showed better performance. Conclusions The proposed model could effectively extract features from the GTVp and PGs channels, achieving good performance in xerostomia prediction.https://doi.org/10.1186/s12880-025-01848-1Deep learningXerostomiaDecision fusion
spellingShingle Lin Lin
Yuchen Ren
Wanwei Jian
Geng Yang
Bailin Zhang
Lin Zhu
Wenhao Zhao
Haoyu Meng
Xuetao Wang
Qiang He
A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
BMC Medical Imaging
Deep learning
Xerostomia
Decision fusion
title A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
title_full A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
title_fullStr A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
title_full_unstemmed A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
title_short A deep learning model for predicting radiation-induced xerostomia in patients with head and neck cancer based on multi-channel fusion
title_sort deep learning model for predicting radiation induced xerostomia in patients with head and neck cancer based on multi channel fusion
topic Deep learning
Xerostomia
Decision fusion
url https://doi.org/10.1186/s12880-025-01848-1
work_keys_str_mv AT linlin adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT yuchenren adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT wanweijian adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT gengyang adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT bailinzhang adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT linzhu adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT wenhaozhao adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT haoyumeng adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT xuetaowang adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT qianghe adeeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT linlin deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT yuchenren deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT wanweijian deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT gengyang deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT bailinzhang deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT linzhu deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT wenhaozhao deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT haoyumeng deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT xuetaowang deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion
AT qianghe deeplearningmodelforpredictingradiationinducedxerostomiainpatientswithheadandneckcancerbasedonmultichannelfusion