节点文献
评价智能心电检测仪诊断心律失常的准确性
Diagnostic accuracy of intelligent electrocardiogram monitor in the diagnosis of arrhythmia
【摘要】 目的评价智能手机APP操作的单导联手持心电检测仪(智能1L-ECG)在心房颤动(atrial fibrillation, AF)、心房扑动(atrial flutter, AFL)等心律失常筛查中的准确性,探讨其人工智能算法和远程预警的优势。方法选择2019年11月至2020年9月上海市徐汇区凌云街道65岁以上自愿接受12L-ECG诊断和智能1L-ECG的体检患者375例。采用随机原则由2位不同医生分别完成12L-ECG及智能1L-ECG的操作。智能1L-ECG采集的图谱由智能集成算法和盲法的心脏病专家评估。结果最终纳入了350例老年患者,年龄为(71.68±5.14)岁,女性占57.1%。智能1L-ECG的智能算法对诊断AF/AFL的受试者工作曲线下面积(area under curve, AUC)为0.851(95%CI,0.649~1),灵敏度、特异度、准确度、阳性似然比和阴性似然比分别为71.4%、98.8%、98.3%、59.5、0.29。心脏病专家诊断AF/AFL的AUC为0.927(95%CI,0.774~1),灵敏度、特异度、准确度、阳性似然比和阴性似然比分别为85.7%、99.7%、99.4%、285.7、0.14;对其他心律异常的AUC为0.791(95%CI,0.719~0.863),灵敏度、特异度、准确度、阳性似然比和阴性似然比分别为63.2%、95%、88.9%、12.6、0.39;对房室传导阻滞的AUC为0.853(95%CI,0.709~0.996),灵敏度、特异度、准确度、阳性似然比和阴性似然比分别为71.4%、99.1%、98%、79.3、0.29。结论智能1L-ECG的智能算法对AF/AFL和房室传导阻滞具有较高的诊断价值,能用于诊断房颤等心律失常,提高筛检率。
【Abstract】 Objective To evaluate the accuracy of a smart phone-operated applications, single-lead ECG monitor(intelligent 1 L-ECG) in screening of atrial fibrillation(AF)/atrial flutter(AFL) and other arrhythmias, and explore the advantages of its intelligent artificial algorithm and remote warning.Methods A total of 375 patients over 65 years old who voluntarily received 12 L-ECG and intelligent 1 L-ECG from Nov.2019 to Sep.2020 were selected from Lingyun Street, Xuhui District, Shanghai.All intelligent 1 L-ECG atlas of electrocardiograms which were collected through smart phone-operated apps were estimated by blinded cardiologists as well as by the smart phone-integrated algorithm.Results We identified a total of 350 elderly patients in Shanghai.Mean±SD age was 71.68 ± 5.14 years, and 57.1% of the patients were female.The algorithm for intelligent 1 L-ECG to diagnose AF/AFL achieved an area under the receiver operating curve(AUC) of 0.851(95% CI,0.649 to 1) with sensitivity, specificity, accuracy, positive likelihood ratio(+LR) and negative likelihood ratio(-LR) of 71.4%,98.8%,98.3%,59.5 and 0.29 respectively, while the cardiologists to diagnose AF/AF had an AUC of 0.927(95% CI,0.774 to 1) with sensitivity, specificity, accuracy, +LR and-LR of 85.7%,99.7%,285.7 and 0.39 respectively.The cardiologists to diagnose other arrhythmias had an AUC of 0.791(95%CI,0.719 to 0.14)with sensitivity, specificity, accuracy, +LR and-LR of 63.2%,95%,88.9%,12.6 and 0.39 respectively. and diagnose atrioventricular block(AV block) had an AUC of 0.853(95%CI,0.709 to 0.996)with sensitivity, specificity, accuracy, +LR and-LR of 71.4%,99.1%,98%,79.3 and 0.29 respectively.Conclusions The algorithm of intelligent 1 L-ECG has high diagnostic value for AF/AFL and AV block.It can be used to diagnose arrhythmias such as atrial fibrillation and improve the screening rate.
【Key words】 electrocardiogram(ECG); arrhythmia; diagnostic accuracy; artificial intelligence; remote warning;
- 【文献出处】 北京生物医学工程 ,Beijing Biomedical Engineering , 编辑部邮箱 ,2022年01期
- 【分类号】R541.7
- 【下载频次】103