Li Juejin, He Hong-Gu, Guan Chang, Ding Yuxin, Hu Xiaolin
Department of Nursing, West China Hospital, Sichuan University/West China School of Nursing, Chengdu, PR China; Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore.
Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; National University Health System, Singapore.
Radiother Oncol. 2025 Aug;209:110993. doi: 10.1016/j.radonc.2025.110993. Epub 2025 Jun 20.
Radiation-induced oral mucositis is one of the most common and debilitating side effects in nasopharyngeal carcinoma patients, which can cause decreased quality of life and treatment adherence. But the effective and tailored strategies for preventing radiation-induced oral mucositis are limited due to unclear risk factors and the lack of prediction models for identifying and stratifying high-risk patients.
To develop a dynamic joint prediction model for severe radiation-induced oral mucositis among nasopharyngeal carcinoma patients.
A prospective longitudinal study was conducted. Multidimensional variables were longitudinally evaluated till 3 months after the completion of radiotherapy. The outcome was the occurrence of severe radiation-induced oral mucositis during follow-up. Dynamic joint prediction model was based on the COX regression analysis and linear mixed effects model. Bootstrap resampling method was used to calculate the area under curve (AUC) and plot receiver operator characteristics (ROC) curve for model internal validation. All statistical analyses were performed in RStudio 4.3.1.
A total of 294 participants were included. Risk factors include oral pH > 6.5 (HR = 0.61 [0.37-0.99]), sequential radiotherapy (HR = 1.80 [1.01-3.13]), dose-volume parameter D50 (HR = 1.04 [1.01-1.08]), increased white blood cell count (HR = 4.47 [1.53-14.60]), increased neutrophilic granulocyte percentage (HR = 1.52 [0.97-2.38]), log(anxiety) (HR = 612.78 [7.66-2.71*10^4]), log(depression) (HR = 0.01 [0.0001-0.20]), and log(nutritional status) (HR = 0.12 [0.01-0.82]). The ROC curves and AUC values revealed that the model has acceptable predictive performance.
This study established a dynamic joint prediction model for severe radiation-induced oral mucositis based on multidisciplinary risk factors, which could provide guidance for developing targeted multidisciplinary interventions to improve severe radiation-induced oral mucositis.
放射性口腔黏膜炎是鼻咽癌患者最常见且严重的副作用之一，可导致生活质量下降和治疗依从性降低。但由于风险因素不明确以及缺乏用于识别和分层高危患者的预测模型，预防放射性口腔黏膜炎的有效且针对性策略有限。
建立鼻咽癌患者严重放射性口腔黏膜炎的动态联合预测模型。
进行一项前瞻性纵向研究。对多维变量进行纵向评估，直至放疗结束后3个月。结局为随访期间严重放射性口腔黏膜炎的发生情况。动态联合预测模型基于COX回归分析和线性混合效应模型。采用自助重抽样法计算曲线下面积（AUC）并绘制受试者工作特征（ROC）曲线进行模型内部验证。所有统计分析均在RStudio 4.3.1中进行。
共纳入294名参与者。风险因素包括口腔pH>6.5（HR=0.61[0.37-0.99]）、序贯放疗（HR=1.80[1.01-3.13]）、剂量体积参数D50（HR=1.04[1.01-1.08]）、白细胞计数升高（HR=4.47[1.53-14.60]）、中性粒细胞百分比升高（HR=1.52[0.97-2.38]）、log(焦虑)（HR=612.78[7.66-2.71×10^4]）、log(抑郁)（HR=0.01[0.0001-0.20]）以及log(营养状况)（HR=0.12[0.01-0.82]）。ROC曲线和AUC值显示该模型具有可接受的预测性能。
本研究基于多学科风险因素建立了严重放射性口腔黏膜炎的动态联合预测模型，可为制定针对性的多学科干预措施以改善严重放射性口腔黏膜炎提供指导。
