基于临床—影像组学特征经皮肾镜取石术后全身炎症反应综合征预测模型的建立  

Establishment of prediction model of systemic inflammatory response syndrome after percutaneous nephrolithotomy based on clinical-imaging characteristics

作　　者：王飞[1] 曹海明 董壮 吴宇[1] 谷明利[1] WANG Fei;CAO Haiming;DONG Zhuang;WU Yu;GU Mingli(Department of Urology,The Second Affiliated Hospital of Bengbu Medical University,Bengbu 233000,China)

机构地区：[1]蚌埠医学院第二附属医院泌尿外科,安徽蚌埠233000

出　　处：《山东医药》2022年第4期36-41,共6页Shandong Medical Journal

基　　金：蚌埠医学院自然科学类项目(BYKY2019164ZD)。

摘　　要：目的建立一种新的临床—影像组学列线图模型,用于预测经皮肾镜取石术(PCNL)治疗上尿路结石后出现全身炎症反应综合征(SIRS)的发生风险。方法208例PCNL治疗成功的上尿路结石患者,按照术后是否出现SIRS分为SIRS组和对照组。对比两组患者的临床资料及基于CT的影像组学资料,采用多因素Logistic回归分析建立临床—影像组学列线图模型,通过计算受试者工作特征曲线下面积(AUC)、GiViTI校准曲线带和临床决策曲线进行验证评估模型性能。结果208例PCNL受试者中,共有34例(16%)术后发生了SIRS。多因素Logistic分析结果提示:性别(OR=3.389,95%CI:1.187~10.231,P=0.025)、术前中段尿培养(OR=4.463,95%CI:1.605~12.411,P=0.004)、术前发热(OR=8.146,95%CI:1.195~41.814,P=0.007)、手术时间(OR=1.030,95%CI:1.011~1.052,P=0.003)、Guy′s分级(OR=2.565,95%CI:1.465~4.842,P=0.002)和影像组学评分(OR=2.601,95%CI:1.568~4.443,P<0.001)与PCNL术后SIRS的发生密切相关,将这些因素选入列线图。PCNL术后发生SIRS的临床—影像组学预测模型的AUC为0.915(95%CI:0.869~0.961),GiViTI校准曲线带展示出模型校准度良好,决策曲线分析表明模型具有较高的临床收益。结论成功建立临床—影像组学预测模型,其对PCNL术后SIRS的发生其有较好的预测能力,可以为临床是否需要使用药物干预提供一定参考。Objective To establish a new clinical-radiomics nomogram model to predict the risk of systemic inflammatory response syndrome(SIRS)after percutaneous nephrolithotomy(PCNL)in the treatment of upper urinary tract calculi.Methods Totally 208 patients with upper urinary tract calculi successfully treated by PCNL were divided into the SIRS group and control group according to the occurrence of SIRS after operation.The clinical data and CT-based radiomics data of the two groups were compared,and the clinical-radiomics nomogram model was established by Multivariate Logistic regression analysis.The performance of the model was verified by calculating area under the receiver operating characteristic curve(AUC),GiViTI calibration curve and clinical decision curve.Results Among 208 PCNL subjects,34 cases(16%)had SIRS after operation.Multivariate Logistic analysis showed that gender(OR=3.389,95%CI:1.187-10.231,P=0.025),preoperative urine culture(OR=4.463,95%CI:1.605-12.411,P=0.004),preoperative fever(OR=8.146,95%CI:1.195-41.814,P=0.007),operation time(OR=1.030,95%CI:1.011-1.052,P=0.003),Guy′s stone score(OR=2.565,95%CI:1.465-4.842,P=0.002)and radiomics score(OR=2.601,95%CI:1.568-4.443,P<0.001)were closely related to the occurrence of SIRS after PCNL,and these factors were selected into the nomogram.The AUC of the clinical-radiomics nomogram model in predicting SIRS after PCNL was 0.915(95%CI:0.869-0.961),the GiViTI calibration curve band showed good calibration degree of the model,and decision curve analysis showed that the model had high clinical benefits.Conclusion The successful establishment of clinical-radiomics nomogram model has a good ability to predict the risk of SIRS after PCNL,and can provide some reference for clinical intervention.

关 键 词：经皮肾镜取石术 全身炎症反应综合征 影像组学 列线图 

分 类 号：R692.4[医药卫生—泌尿科学]