肺癌患者动态调强放疗后肺部侵袭性真菌感染的风险预测模型构建  

Construction of risk prediction model for invasive pulmonary fungal infection in patients with pulmonary cancer after dynamic intensity modulated radiation therapy

作　　者：吕魏潇[1] 虞荷荷 梁晨艳[1] 应晓珍[1] LYU Wei-xiao;YU He-he;LIANG Chen-yan;YING Xiao-zhen(Lishui Central Hospital,Lishui,Zhejiang 323000,China;不详)

机构地区：[1]丽水市中心医院放疗科,浙江丽水323000 [2]温州市中心医院感染科,浙江温州325000

出　　处：《中华医院感染学杂志》2023年第8期1153-1157,共5页Chinese Journal of Nosocomiology

基　　金：丽水市科技计划基金资助项目(2017GYX09)。

摘　　要：目的探讨肺癌患者动态调强放疗后肺部侵袭性真菌感染的危险因素,建立预测模型,为患者肺部侵袭性真菌感染的预测提供参考。方法选取2018年1月-2020年12月丽水市中心医院301例肺癌放疗治疗患者,其中继发肺部侵袭性真菌感染的患者纳入感染组(n=60),未继发肺部侵袭性真菌感染的患者纳入未感染组(n=241),分析感染组患者放疗后肺部真菌感染病原菌分布特征,采用Logistic回归分析肺部真菌侵袭性感染危险因素,基于危险因素的回归系数构建风险预测模型,通过受试者工作特征(ROC)曲线评估预测模型区分度。结果60例感染组患者共检测出真菌69株,包括假丝酵母菌属65株(94.20%),曲霉菌属4株(5.80%);Logistic回归分析显示年龄≥60岁、合并糖尿病、侵入性操作、同步化疗是肺癌放疗患者肺部真菌侵袭性感染的独立危险因素(P<0.05);肺癌放疗患者肺部真菌侵袭性感染概率值回归方程为:P=1/[1+e^(-(-1.993+1.004×年龄+1.590×糖尿病+1.705×同步化疗+1.131×侵入性操作))],采用Hosmer-Lemeshow检验检测回归方程的拟合优度(P=0.239)。对模型进行内部验证,其ROC曲线下面积(AUC)为0.813[95%CI(0.741~0.885)],灵敏度为73.90%,特异度为86.30%,区分度良好,根据约登指数最大原则选取cut-off点为0.274。结论建立的风险评分模型判别效度良好,可作为加强肺癌放疗患者肺部侵袭性真菌感染预防控制的重要理论依据。OBJECTIVE To explore the risk factors for invasive pulmonary fungal infection in patients with pulmonary cancer after dynamic intensity modulated radiation therapy and to construct prediction model in order to provide reference for predicting invasive pulmonary fungal infection.METHODS A total of 301 patients undergoing pulmonary cancer radiotherapy in Lishui Central Hospital between Jan 2018 and Dec 2020 were enrolled,including 60 cases with invasive pulmonary fungal infection in the infection group and 241 cases without invasive pulmonary fungal infection in the non-infection group.The distribution characteristics of invasive pulmonary fungal infection pathogens were analyzed.The risk factors for invasive pulmonary fungal infection were analyzed by logistic regression analysis.The risk prediction model was constructed based on regression coefficients of risk factors.The discrimination of prediction model was evaluated by receiver operating characteristics(ROC)curves.RESULTS A total of 69 strains of fungi were detected in 60 patients of the infection group,including 65 strains(94.20%)of Candida and 4 strains(5.80%)of Aspergillus.Logistic regression analysis showed that age≥60 years,diabetes mellitus,invasive operation and concurrent chemotherapy were independent risk factors for invasive pulmonary fungal infection(P<0.05).The regression equation to predicting the incidence of invasive pulmonary fungal infection was as follows:P=1[1+e^(-(-2.313+1.004×age+1.590×diabetes mellitus+1.705×concurrent chemotherapy+1.131×invasive operation))].The regression equation′s goodness of fit was detected by Hosmer-Lemeshow test(P=0.239).The internal verification of the model was conducted.The area under the ROC curve(AUC),sensitivity and specificity were 0.813[95%CI:(0.741-0.885)],73.90%and 86.30%,with good discrimination.According to the principle of the maximum Youden index,the cut-off value was 0.274.CONCLUSION The discriminative validity of risk scoring model is good,which can be applied as the important theoretical

关 键 词：肺癌 放疗 肺部侵袭性真菌感染 病原菌 危险因素 预测模型 

分 类 号：R563.1[医药卫生—呼吸系统]