个体化预测先天性泪道阻塞患儿泪道探通失败的风险Nomgram模型的建立与验证  

Establishment and validation of Nomgram model for predicting the risk of lacrimal passage probing failure in children with congenital lacrimal duct obstruction

作　　者：邓姿峰[1] 肖志刚[1] 汤文权 Deng Zifeng;Xiao Zhigang;Tang Wenquan(Department of Ophthalmology,Hunan Children’s Hospital,Changsha 410007,China)

机构地区：[1]湖南省儿童医院,长沙410007

出　　处：《临床眼科杂志》2022年第1期58-62,共5页Journal of Clinical Ophthalmology

摘　　要：目的分析先天性泪道阻塞患儿泪道探通失败的危险因素,以建立风险Nomgram模型,并对模型进行验证。方法前瞻性临床研究。选取2019年1月至2020年1月350例先天性泪道阻塞患儿,并按照随机数字表分布法划分为建模集(n=240)和验证集(n=110),并统计两组一般临床资料,单因素和多因素Logistic回归分析先天性泪道阻塞患儿泪道探通失败的危险因素,以建立相关预测模型,行Bootstrap法内部验证模型,外部验证通过验证集检验。结果年龄>6个月(OR=4.938,95%CI:2.016~12.094)、复杂性泪道阻塞(OR=2.948,95%CI:1.087~7.994)、合并脓性分泌物(OR=3.823,95%CI:1.372~10.656)、形成假道(OR=2.633,95%CI:0.998~6.946)、无法常规进针(OR=5.328,95%CI:1.591~17.856)及鼻泪管难以进入(OR=4.276,95%CI:1.369~13.353)是先天性泪道阻塞患儿泪道探通失败的独立危险因素,基于此6项危险因素构建预测Nomgram模型,R方程检验结果表明建模集和验证集的C-index指数分别为0.853(95%CI:0.831~0.874)和0.841(95%CI:0.822~0.865);校正曲线和标准曲线结果较为相近,表明该风险Nomgram模型精准度和区分度较好。结论年龄>6个月、复杂性泪道阻塞、合并脓性分泌物、形成假道、无法常规进针及鼻泪管难以进入是先天性泪道阻塞患儿泪道探通失败的独立危险因素,且该Nomgram模型经内部验证及外部验证表明预测精准度较高,临床应用价值较好。Objective To analyze the risk factors of lacrimal passage probing failure in children with congenital lacrimal duct obstruction,so as to establish and verify the risk-prediction Nomgram model.Methods Prospective clinical study.350 children with congenital lacrimal duct obstruction from January 2019 to January 2020 were selected and divided into the training set(n=240)and validation set(n=110)according to the random number table distribution method.The general clinical data of the two groups were statistically analyzed.The risk factors of lacrimal passage probing failure in children with congenital lacrimal duct obstruction were analyzed by univariate and multivariate logistic regression,so as to establish the relevant prediction model and conduct bootstrap method to verify the internal model,the external validation was conducted based on the validation set.Results Age greater than 6 months(OR 4.938,95%C I:2.016~12.094),complex lacrimal duct obstruction(OR=2.948,95%CI:1.087~7.994),complicated purulent secretions(OR=3.823,95%CI:1.372~10.656),false passage formation(OR=2.633,95%CI:0.998~6.946),inability of conventional needle insertion(OR=5.328,95%CI:1.591~17.856)and difficulty in nasolacrimal duct entry(OR=4.276,95%CI:1.369~13.353)were independent risk factors for lacrimal passage probing failure in children with congenital lacrimal duct obstruction.Based on these six risk factors,the Nomgram model was constructed to predict lacrimal passage penetration failure.The results of r-equation test showed that the c-index of the training set and the validation set was 0.853(95%CI:0.831~0.874)and 0.841(95%CI:0.822~0.865),respectively.The calibration curve and the standard curve were similar,indicating good accuracy and discrimination of the risk Nomgram model.Conclusions Age greater than 6 months,complex lacrimal duct obstruction,complicated purulent secretions,false passage formation,inability of routine needle insertion and difficult access of nasolacrimal duct are independent risk factors for lacrimal passage probing fa

关 键 词：先天性泪道阻塞 泪道探通失败 LOGISTIC 危险因素 Nomgram模型 

分 类 号：R777.2[医药卫生—眼科]