Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma
Abstract Background Xerostomia is one of the most common side effects in nasopharyngeal carcinoma (NPC) patients after chemoradiotherapy. To establish a Delta radiomics model for predicting xerostomia secondary to chemoradiotherapy for NPC based on magnetic resonance T1-weighted imaging (T1WI) seque...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2024-02-01
|
Series: | Radiation Oncology |
Subjects: | |
Online Access: | https://doi.org/10.1186/s13014-024-02417-6 |
_version_ | 1797273746478727168 |
---|---|
author | Mengze Wang Yuzhen Xi Luoyu Wang Haonan Chen Feng Jiang Zhongxiang Ding |
author_facet | Mengze Wang Yuzhen Xi Luoyu Wang Haonan Chen Feng Jiang Zhongxiang Ding |
author_sort | Mengze Wang |
collection | DOAJ |
description | Abstract Background Xerostomia is one of the most common side effects in nasopharyngeal carcinoma (NPC) patients after chemoradiotherapy. To establish a Delta radiomics model for predicting xerostomia secondary to chemoradiotherapy for NPC based on magnetic resonance T1-weighted imaging (T1WI) sequence and evaluate its diagnostic efficacy. Methods Clinical data and Magnetic resonance imaging (MRI) data before treatment and after induction chemotherapy (IC) of 255 NPC patients with stage III-IV were collected retrospectively. Within one week after CCRT, the patients were divided into mild (92 cases) and severe (163 cases) according to the grade of xerostomia. Parotid glands in T1WI sequence images before and after IC were delineated as regions of interest for radiomics feature extraction, and Delta radiomics feature values were calculated. Univariate logistic analysis, correlation, and Gradient Boosting Decision Tree (GBDT) methods were applied to reduce the dimension, select the best radiomics features, and establish pretreatment, post-IC, and Delta radiomics xerostomia grading predictive models. The receiver operating characteristic (ROC) curve and decision curve were drawn to evaluate the predictive efficacy of different models. Results Finally, 15, 10, and 12 optimal features were selected from pretreatment, post-IC, and Delta radiomics features, respectively, and a xerostomia prediction model was constructed with AUC values of 0.738, 0.751, and 0.843 in the training set, respectively. Only age was statistically significant in the clinical data of both groups (P < 0.05). Conclusion Delta radiomics can predict the degree of xerostomia after chemoradiotherapy for NPC patients and it has certain guiding significance for clinical early intervention measures. |
first_indexed | 2024-03-07T14:48:36Z |
format | Article |
id | doaj.art-6862ae493eae4ecea8983c40c6e7923d |
institution | Directory Open Access Journal |
issn | 1748-717X |
language | English |
last_indexed | 2024-03-07T14:48:36Z |
publishDate | 2024-02-01 |
publisher | BMC |
record_format | Article |
series | Radiation Oncology |
spelling | doaj.art-6862ae493eae4ecea8983c40c6e7923d2024-03-05T19:51:28ZengBMCRadiation Oncology1748-717X2024-02-011911910.1186/s13014-024-02417-6Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinomaMengze Wang0Yuzhen Xi1Luoyu Wang2Haonan Chen3Feng Jiang4Zhongxiang Ding5Department of Radiology, Zhejiang Cancer HospitalDepartment of Radiology, 903 RD Hospital of PLADepartment of Radiology, Hangzhou First People’s HospitalDepartment of Radiology, Zhejiang HospitalDepartment of Head and Neck Radiotherapy, Zhejiang Province Key Laboratory of Radiation Oncology, Zhejiang Cancer HospitalDepartment of Radiology, Hangzhou First People’s HospitalAbstract Background Xerostomia is one of the most common side effects in nasopharyngeal carcinoma (NPC) patients after chemoradiotherapy. To establish a Delta radiomics model for predicting xerostomia secondary to chemoradiotherapy for NPC based on magnetic resonance T1-weighted imaging (T1WI) sequence and evaluate its diagnostic efficacy. Methods Clinical data and Magnetic resonance imaging (MRI) data before treatment and after induction chemotherapy (IC) of 255 NPC patients with stage III-IV were collected retrospectively. Within one week after CCRT, the patients were divided into mild (92 cases) and severe (163 cases) according to the grade of xerostomia. Parotid glands in T1WI sequence images before and after IC were delineated as regions of interest for radiomics feature extraction, and Delta radiomics feature values were calculated. Univariate logistic analysis, correlation, and Gradient Boosting Decision Tree (GBDT) methods were applied to reduce the dimension, select the best radiomics features, and establish pretreatment, post-IC, and Delta radiomics xerostomia grading predictive models. The receiver operating characteristic (ROC) curve and decision curve were drawn to evaluate the predictive efficacy of different models. Results Finally, 15, 10, and 12 optimal features were selected from pretreatment, post-IC, and Delta radiomics features, respectively, and a xerostomia prediction model was constructed with AUC values of 0.738, 0.751, and 0.843 in the training set, respectively. Only age was statistically significant in the clinical data of both groups (P < 0.05). Conclusion Delta radiomics can predict the degree of xerostomia after chemoradiotherapy for NPC patients and it has certain guiding significance for clinical early intervention measures.https://doi.org/10.1186/s13014-024-02417-6Nasopharyngeal carcinomaMagnetic resonance imagingDelta RadiomicsParotid gland injuryXerostomia |
spellingShingle | Mengze Wang Yuzhen Xi Luoyu Wang Haonan Chen Feng Jiang Zhongxiang Ding Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma Radiation Oncology Nasopharyngeal carcinoma Magnetic resonance imaging Delta Radiomics Parotid gland injury Xerostomia |
title | Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma |
title_full | Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma |
title_fullStr | Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma |
title_full_unstemmed | Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma |
title_short | Predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage III-IV nasopharyngeal carcinoma |
title_sort | predictive value of delta radiomics in xerostomia after chemoradiotherapy in patients with stage iii iv nasopharyngeal carcinoma |
topic | Nasopharyngeal carcinoma Magnetic resonance imaging Delta Radiomics Parotid gland injury Xerostomia |
url | https://doi.org/10.1186/s13014-024-02417-6 |
work_keys_str_mv | AT mengzewang predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma AT yuzhenxi predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma AT luoyuwang predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma AT haonanchen predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma AT fengjiang predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma AT zhongxiangding predictivevalueofdeltaradiomicsinxerostomiaafterchemoradiotherapyinpatientswithstageiiiivnasopharyngealcarcinoma |