NSCLC同步放化疗后中性粒细胞减少症风险预测  

Predicting the risk of neutropenia after synchronous radiotherapy and chemotherapy for NSCLC

作　　者：饶井芬 于晓磊[1] 肇爽[1] 张聪[1] 冉欣欣 朱彤[1] RAO Jingfen;YU Xiaolei;ZHAO Shuang;ZHANG Cong;RAN Xinxin;ZHU Tong(Department of Oncology,Affiliated Hospital of Chengde Medical College,Chengde 067000,China;Department of Respiratory and Critical Care,Affiliated Hospital of Chengde Medical College,Chengde 067000,China)

机构地区：[1]承德医学院附属医院肿瘤科,河北承德067000 [2]承德医学院附属医院呼吸与危重症科,河北承德067000

出　　处：《东南大学学报（医学版）》2024年第2期209-215,共7页Journal of Southeast University(Medical Science Edition)

基　　金：2022年承德市科技计划自筹经费立项项目(202204A035)。

摘　　要：目的:探讨随机森林模型预测非小细胞肺癌(NSCLC)同步放化疗后中性粒细胞减少症的价值,以期为临床早期预测中性粒细胞减少症发生风险,制定相应干预措施提供参考。方法:选取2020年1月至2022年8月我院295例NSCLC患者,均接受同步放化疗,根据化疗后是否发生中性粒细胞减少症分为中性粒细胞减少症组(n=91)、非中性粒细胞减少症组(n=204),采用随机数字表法按照2∶1比例建立训练集和测试集,构建随机森林预测模型,受试者工作特征(ROC)曲线分析随机森林模型对中性粒细胞减少症的预测效能。结果:两组年龄、体重、化疗延迟、化疗方案剂量、同步放化疗周期、肠内营养支持、高血压、糖尿病、自身免疫性疾病、预防性应用长效粒细胞集落刺激因子(G-CSF)比较差异有统计学意义(P<0.05);化疗延迟、同步放化疗周期、肠内营养支持、年龄、化疗方案剂量、体重均为中性粒细胞减少症发生的影响因素(P<0.05);基于化疗延迟、同步放化疗周期、肠内营养支持、年龄、化疗方案剂量、体重的随机森林模型预测中性粒细胞减少症发生的ROC曲线下面积(AUC)为0.857,敏感度为80.00%,特异度为95.00%。结论:基于化疗延迟、同步放化疗周期、肠内营养支持、年龄、化疗方案剂量、体重的随机森林模型对于NSCLC患者同步放化疗后中性粒细胞减少症具有较高评估预测价值,临床可通过上述因素针对性制定干预方案,降低中性粒细胞减少症发生风险。Objective:To explore the value of random forest model in predicting neutropenia in non-small cell lung cancer(NSCLC)after synchronous radiotherapy and chemotherapy,in order to provide reference for early clinical prediction of the risk of neutropenia and the development of corresponding intervention measures.Methods:295 NSCLC patients were selected from January 2020 to August 2022 in our hospital,all of whom received synchronous radiotherapy and chemotherapy,they were divided into neutropenia group(n=91)and non-neutropenia group(n=204)according to whether neutropenia occurred after chemotherapy,and the random number table method was used to establish the training set and the test set according to the ratio of 2∶1,and the random forest model was constructed.The predictive efficacy of random forest model for neutropenia was analyzed by receiver operating characteristic(ROC)curve.Results:The differences in age,weight,chemotherapy delay,synchronous regimen dose,concurrent radiotherapy and chemotherapy cycle,enteral nutrition support,hypertension,diabetes mellitus,autoimmune disease,and prophylactic granulocyte colony stimulating factor(G-CSF)application between the two groups were statistically significant(P<0.05).Chemotherapy delay,synchronous radiotherapy and chemotherapy cycle,enteral nutrition support,age,chemotherapy regimen dose,and weight were all factors associated with the development of neutropenia(P<0.05).The random forest model based on chemotherapy delay,synchronous radiotherapy and chemotherapy cycle,enteral nutrition support,age,chemotherapy regimen dose,and body weight predicted neutropenia with an AUC of 0.857,sensitivity 80.00%,and specificity 95.00%.Conclusion:A random forest model based on chemotherapy delay,synchronous radiotherapy and chemotherapy cycle,enteral nutrition support,age,chemotherapy regimen dose,and body weight has a high predictive value for assessing neutropenia after synchronous radiotherapy and chemotherapy in patients with NSCLC,and these factors can be used to develop a targ

关 键 词：非小细胞肺癌 同步放化疗 中性粒细胞减少症 随机森林模型 预测价值 影响因素 

分 类 号：R734.2[医药卫生—肿瘤]