弥漫大B细胞淋巴瘤免疫化疗后肺部感染的风险预测模型建立与验证  

Establishment and validation of a predictive model for the risk of pulmonary infection after immunochemotherapy for diffuse large B-cell lymphoma

作　　者：潘思澄 张婧 胡云飞 黄韵红 PAN Sicheng;ZHANG Jing;HU Yunfei;HUANG Yunhong(Department of Oncology,the Affiliated Hospital of Guizhou Medical University,Guiyang 550002,Guizhou,China)

机构地区：[1]贵州医科大学附属医院肿瘤科,贵州贵阳550002

出　　处：《贵州医科大学学报》2025年第2期268-274,共7页Journal of Guizhou Medical University

基　　金：贵州省卫生健康委员会科学技术基金项目(2021XMSB00032675)。

摘　　要：目的建立弥漫大B细胞淋巴瘤(diffuse large B cell lymphoma,DLBCL)免疫化疗后发生肺部感染预测模型,并评价模型预测效能。方法选取2018年1月—2022年5月贵州医科大学附属医院肿瘤科收治的350例DLBCL患者为研究对象,分为肺部感染组(n=148)和无肺部感染组(n=202),收集患者一般临床资料和实验室检查结果,采用lasso回归的方法行变量筛选,基于筛选结果构建多因素logistic模型,再结合模型结果构建带有风险分层的列线图;应用受试者特征曲线(receiver operating characteristic curve,ROC)、校正曲线及Hosmer-Lemeshow检验对模型的预测性能进行验证,采用临床决策曲线(decision curve analysis,DCA)对模型的临床适用性进行评估。结果多因素分析结果显示,年龄≥60岁、KPS评分<80分、人血清白蛋白含量<30 g/L、血红蛋白含量<90 g/L、白细胞计数<1×10^(9)个/L和中性粒细胞减少持续时间≥7 d均是肺部感染的的独立危险因素(P<0.05);进一步构建风险预测模型,训练集和验证集的ROC曲线下面积(area under the curve,AUC)分别为0.866(95%CI=0.822~0.910)和0.829(95%CI=0.736~0.923),提示该模型的区分度较好;Hosmer-Lemeshow拟合优度检验结果(P>0.05);模型的校准曲线略微偏离对角线,提示校准度良好;DCA显示训练集和验证集的曲线均比参考曲线高,提示该模型具有较好的临床意义。结论基于患者年龄、人血清白蛋白含量、血红蛋白含量、白细胞计数、中性粒细胞减少持续时间及KPS评分构建的肺部感染风险预测模型对DLBCL患者免疫化疗后发生肺部感染具有一定预测价值。Objective To establish a predictive model for pulmonary infection after immunochemotherapy in diffuse large B-cell lymphoma(DLBCL)and to evaluate the predictive efficacy of the model.Methods A total of 350 patients with DLBCL admitted to the Department of Oncology in the Affiliated Hospital of Guizhou Medical University from January 2018 to May 2022 were selected as participants of this study.They were divided into the pulmonary infection group(n=148)and the non-pulmonary infection group(n=202).The general clinical data and laboratory examination results of the patients were collected.The Lasso regression was used to screen the variables,and the multi-factor logistic model was constructed based on the screening results.Combined with the model results,a nomogram with risk stratification was constructed.The receiver operating characteristic curve(ROC),calibration curve and Hosmer-Lemeshow test were applied to validate the predictive performance of the model,and the clinical applicability of the model was examined using the clinical decision curve analysis(DCA).Results Multifactor analysis showed that age≥60 years old,KPS score<80 points,human serum albumin content<30 g/L,hemoglobin content<90 g/L,white blood cell count<1×10^(9)/L,and neutropenia duration≥7 days were independent risk factors for pulmonary infection(P<0.05).The risk prediction model was constructed based on the above risk factors,and the area under the curve(AUC)of the ROC curve for the training and validation sets were 0.866(95%CI was 0.822-0.91)and 0.829(95%CI was 0.736-0.923),respectively,suggesting that the model had good differentiation.The results of Hosmer-Lemeshow goodness of fit test were P>0.05,and the calibration curve of the model was slightly deviated from the diagonal,indicating a good calibration.The DCA showed that the curves for both the training and validation sets were higher than the reference curve,suggesting that this model had good clinical significance.Conclusion The risk prediction model of pulmonary infection based on p

关 键 词：弥漫大B细胞淋巴瘤 免疫化疗 肺部感染 风险预测模型 

分 类 号：R733.4[医药卫生—肿瘤]