新生儿遗传代谢病人工智能疾病风险评估模型的建立及验证  

Establishment of an auxiliary diagnosis system of newborn screening for inherited metabolic diseases based on artificial intelligence technology and a clinical trial

作　　者：杨茹莱[1] 杨艳玲[2] 王挺[3] 徐玮泽[4] 俞刚[5] 杨建滨[1] 孙巧玲 顾茂胜 李海波 赵德华[9] 裴菊英[10] 蒋涛[11] 贺骏[12] 邹卉[13] 毛新梅 耿国兴[15] 强荣 田国力[17] 王艳[18] 韦洪伟 张晓刚[20] 王华[21] 田亚平[22] 邹琳[23] 孔元原[24] 周玉侠[25] 欧明才[26] 药泽蓉[27] 周裕林 朱文斌[29] 黄永兰 王玉红[31] 黄慈丹 镡颖 李龙[34] 尚清 郑宏[36] 吕少磊 王文君 姚艳 乐静 舒强[4] Yang Rulai;Yang Yanling;Wang Ting;Xu Weize;Yu Gang;Yang Jianbin;Sun Qiaoling;Gu Maosheng;Li Haibo;Zhao Dehua;Pei Juying;Jiang Tao;He Jun;Zou Hui;Mao Xinmei;Geng Guoxing;Qiang Rong;Tian Guoli;Wang Yan;Wei Hongwei;Zhang Xiaogang;Wang Hua;Tian Yaping;Zou Lin;Kong Yuanyuan;Zhou Yuxia;Ou Mingcai;Yao Zerong;Zhou Yulin;Zhu Wenbin;Huang Yonglan;Wang Yuhong;Huang Cidan;Tan Ying;Li Long;Shang Qing;Zheng Hong;Lyu Shaolei;Wang Wenjun;Yao Yan;Le Jing;Shu Qiang(Department of Genetics and Metabolism,the Children′s Hospital of Zhejiang University School of Medicine,National Clinical Research Center for Child Health,Hangzhou 310052,China;Department of Pediatrics,Peking University First Hospital,Beijing 100034,China;Center for Reproduction and Genetics,Suzhou Municipal Hospital,Suzhou 215002,China;Department of Cardiac Surgery,the Children′s Hospital of Zhejiang University School of Medicine,National Clinical Research Center for Child Health,Hangzhou 310052,China;Department of Information,the Children′s Hospital of Zhejiang University School of Medicine,National Clinical Research Center for Child Health,Hangzhou 310052,China;Neonatal Screening Center,Anhui Province Maternity&Child Health Hospital,Hefei 230001,China;Genetics Medicine Center,Xuzhou Maternity and Child Health Hospital,Xuzhou 221000,China;The Central Laboratory of Birth Defects Prevention and Control,Ningbo Women&Children′s Hospital,Ningbo 315012,China;Neonatal Screening Center,the Third Affiliated Hospital of Zhengzhou University,Zhengzhou 450052,China;Department of Child Health Care,Hebei Center for Women and Children′s Health,Shijiazhuang 050031,China;Prenatal Diagnosis Center,Nanjing Maternity and Child Health Hospital,Nanjing 210004,China;Department of Maternal Health,Changsha Hospital for Maternal&Child Health Care,Changsha 410001,China;Neonatal Screening Center,Jinan Maternity and Child Care Hospital,Jinan 250000,China;Neonatal Screening Center,Ningxia Maternal and Child Health Care Hospital,Yinchuan 750011,China;Genetic Metabolic Center,Maternity and

机构地区：[1]浙江大学医学院附属儿童医院遗传与代谢科国家儿童健康与疾病临床医学研究中心,杭州310052 [2]北京大学第一医院儿科,100034 [3]苏州市立医院生殖与遗传中心,215002 [4]浙江大学医学院附属儿童医院心脏外科国家儿童健康与疾病临床医学研究中心,杭州310052 [5]浙江大学医学院附属儿童医院信息科国家儿童健康与疾病临床医学研究中心,杭州310052 [6]安徽省妇幼保健中心新生儿疾病筛查中心,合肥230001 [7]徐州市妇幼保健院遗传医学中心,221000 [8]宁波市妇女儿童医院出生缺陷综合防治重点实验室,315012 [9]郑州大学第三附属医院新生儿筛查中心,450052 [10]河北省妇幼保健中心儿童保健科,石家庄050031 [11]南京市妇幼保健院产前诊断中心,210004 [12]长沙市妇幼保健院孕产保健部,410001 [13]济南市妇幼保健院新生儿疾病筛查中心,250000 [14]宁夏回族自治区妇幼保健院新生儿疾病筛查中心,银川750011 [15]广西壮族自治区妇幼保健院遗传代谢中心,南宁538001 [16]西北妇女儿童医院医学遗传中心,西安710003 [17]上海市儿童医院新生儿筛查中心,200040 [18]解放军总医院第七医学中心出生缺陷干预工程实验室,北京100000 [19]临沂市妇幼保健院新生儿疾病筛查中心,276016 [20]山西省儿童医院新生儿疾病筛查中心,太原030013 [21]湖南省妇幼保健院遗传中心,长沙410008 [22]解放军总医院第一医学中心出生缺陷防控中心实验室,北京100000 [23]重庆医科大学附属儿童医院临床分子医学中心,400014 [24]首都医科大学附属北京妇产医院儿童保健科,100010 [25]山东大学附属山东省妇幼保健院新生儿疾病筛查中心,济南250000 [26]四川省妇幼保健院新生儿疾病筛查科,成都610000 [27]山西省长治市妇幼保健院产科,046011 [28]厦门市妇幼保健院新生儿疾病筛查中心,361003 [29]福建省妇幼保健院新生儿疾病筛查中心,福州350005 [3

出　　处：《中华儿科杂志》2021年第4期286-293,共8页Chinese Journal of Pediatrics

基　　金：国家重点研发计划(2017YFC1001700,2017YFC1001703,2018YFC1002700,2018YFC1002703)。

摘　　要：目的应用人工智能技术建立新生儿遗传代谢病疾病风险评估模型并验证其用于辅助新生儿串联筛查结果的判断。方法回顾性研究。收集2010年2月至2019年5月来自全国31家医院新生儿遗传代谢病筛查(串联质谱法)5907547例数据和34家医院临床确诊的3028例数据进行回顾性分析,建立新生儿遗传代谢病人工智能疾病预测模型;以2018年1至9月浙江大学医学院附属儿童医院360814例新生儿筛查数据进行遗传代谢病人工智能疾病预测模型的单盲试验验证,通过比较临床确诊病例的检出率、串联初筛阳性率和阳性预测值在人工判读和遗传代谢病人工智能预测模型中的结果,验证人工智能疾病风险预测模型的有效性。结果经数据筛选,共有3665697例新生儿串联初筛数据符合数据库建模的标准,选取所有临床确诊患儿数据3019例共构建了16种人工智能预测模型可涵盖32种遗传代谢病;在单盲试验验证入组的360814例新生儿中,临床确诊病例共45例,人工判读和遗传代谢病人工智能预测模型结果一致,所有临床确诊病例均为阳性或高风险。串联初筛阳性人工判读为2684例,遗传代谢病人工智能疾病风险预测模型判读为串联初筛高风险1694例,串联初筛阳性率分别为0.74%(2684/360814)、0.46%(1694/360814);与人工判读相比,遗传代谢病人工智能疾病风险预测模型判读阳性人数总体减少了36.89%(990/2684);人工判读和遗传代谢病人工智能疾病风险预测模型的阳性预测值分别为1.68%(45/2684)、2.66%(45/1694)。结论所建立的新生儿遗传代谢病人工智能疾病风险预测模型具有准确、快速、假阳性率低的优点,具有重要临床应用价值。Objective To establish a disease risk prediction model for the newborn screening system of inherited metabolic diseases by artificial intelligence technology.Methods This was a retrospectively study.Newborn screening data(n=5907547)from February 2010 to May 2019 from 31 hospitals in China and verified data(n=3028)from 34 hospitals of the same period were collected to establish the artificial intelligence model for the prediction of inherited metabolic diseases in neonates.The validity of the artificial intelligence disease risk prediction model was verified by 360814 newborns'screening data from January 2018 to September 2018 through a single-blind experiment.The effectiveness of the artificial intelligence disease risk prediction model was verified by comparing the detection rate of clinically confirmed cases,the positive rate of initial screening and the positive predictive value between the clinicians and the artificial intelligence prediction model of inherited metabolic diseases.Results A total of 3665697 newborns'screening data were collected including 3019 cases'positive data to establish the 16 artificial intelligence models for 32 inherited metabolic diseases.The single-blind experiment(n=360814)showed that 45 clinically diagnosed infants were detected by both artificial intelligence model and clinicians.A total of 2684 cases were positive in tandem mass spectrometry screening and 1694 cases were with high risk in artificial intelligence prediction model of inherited metabolic diseases,with the positive rates of tandem 0.74%(2684/360814)and 0.46%(1694/360814),respectively.Compared to clinicians,the positive rate of newborns was reduced by 36.89%(990/2684)after the application of the artificial intelligence model,and the positive predictive values of clinicians and artificial intelligence prediction model of inherited metabolic diseases were 1.68%(45/2684)and 2.66%(45/1694)respectively.Conclusion An accurate,fast,and the lower false positive rate auxiliary diagnosis system for neonatal inherited metabolic

关 键 词：婴儿 新生 人工智能 遗传代谢病 疾病风险评估 

分 类 号：R722.1[医药卫生—儿科]