摘要
目的:用神经网络模型预测妊娠合并SLE患者的妊娠结局。方法:回顾分析2011年9月至2018年6月上海交通大学医学院附属仁济医院妇产科收治的469例妊娠合并SLE患者的临床资料,训练并建立预测妊娠结局的神经网络模型。结果:多层感知器(MLP)和径向基函数(RBF)两种神经网络模型可用于妊娠合并SLE患者妊娠结局的预测:前者的训练准确率为92.6%,测试准确率为92.9%,ROC曲线的AUC为0.911;后者训练准确率为92.0%,测试准确率为87.5%,ROC曲线的曲线下面积(AUC)为0.880。MLP神经网络模型在预测妊娠合并SLE患者妊娠结局方面更具优势。结论:神经网络模型可用于妊娠合并SLE患者妊娠结局的预测并且具有较高的准确性。
Objective: To predict pregnancy outcomes of women with systemic lupus erythematous(SLE) by artificial neural network. Methods: A retrospective cohort study was performed in Renji Hospital between Sep.2011 and Jun.2018.469 pregnant women with SLE were included in the study and their clinical data were collected.Two artificial neural network models were trained and established to predict pregnancy outcomes. Results: Artificial neural network model of multilayer perceptron(MLP) and radial basis function(RBF) were trained and established to predict pregnancy outcomes of pregnant SLE patients:the former got a training accuracy of 92.6% and a testing accuracy of 92.9%,with an area under curve(AUC)ROC of 0.911 .The later got a training accuracy of 92.0% and a testing accuracy of 87.5%,with an area under curve(AUC)ROC of 0.880.MLP network was more accurate in prediction of pregnancy outcomes. Conclusions: Artificial neural network can be used to predict pregnancy outcomes of pregnant SLE patients with a high accuracy.
作者
马京杭
张羽
吴珈悦
冯振
狄文
Ma Jinghang;Zhang Yu;Wu Jiayue(Department of Obstetrics and Gynecology,RenJi Hospital,School of Medicine,Shanghai Jiaotong University,Shanghai 200127;Shanghai Key Laboratory of Gynecologic Oncology,Shanghai 200127)
出处
《现代妇产科进展》
CSCD
北大核心
2019年第2期120-123,127,共5页
Progress in Obstetrics and Gynecology
基金
上海卫生和计划委员会课题(No:15GWZK0701)
关键词
系统性红斑狼疮
妊娠结局
神经网络
Systemic lupus erythematosus
Pregnancy outcomes
Artificial neural network
