Simon Weidlich MD , Diego Mannhart MD , Alan Kennedy PhD , Peter Doggart , Teodor Serban MD , Sven Knecht DSc-PhD , Jeanne Du Fay de Lavallaz MD-PhD , Michael Kühne MD , Christian Sticherling MD , Patrick Badertscher MD
{"title":"通过新型人工智能算法减轻智能设备单导联心电图描记不确定的负担","authors":"Simon Weidlich MD , Diego Mannhart MD , Alan Kennedy PhD , Peter Doggart , Teodor Serban MD , Sven Knecht DSc-PhD , Jeanne Du Fay de Lavallaz MD-PhD , Michael Kühne MD , Christian Sticherling MD , Patrick Badertscher MD","doi":"10.1016/j.cvdhj.2023.12.003","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><p>Multiple smart devices capable of automatically detecting atrial fibrillation (AF) based on single-lead electrocardiograms (SL-ECG) are presently available. The rate of inconclusive tracings by manufacturers’ algorithms is currently too high to be clinically useful.</p></div><div><h3>Method</h3><p>This is a prospective, observational study enrolling patients presenting to a cardiology service at a tertiary referral center. We assessed the clinical value of applying a smart device artificial intelligence (AI)-based algorithm for detecting AF from 4 commercially available smart devices (AliveCor KardiaMobile, Apple Watch 6, Fitbit Sense, and Samsung Galaxy Watch3). Patients underwent a nearly simultaneous 12-lead ECG and 4 smart device SL-ECGs. The novel AI algorithm (PulseAI, Belfast, United Kingdom) was compared with each manufacturer’s algorithm.</p></div><div><h3>Results</h3><p>We enrolled 206 patients (31% female, median age 64 years). AF was present in 60 patients (29%). Sensitivity and specificity for the detection of AF by the novel AI algorithm vs manufacturer algorithm were 88% vs 81% (<em>P</em> = .34) and 97% vs 77% (<em>P</em> < .001) for the AliveCor KardiaMobile, 86% vs 81% (<em>P</em> = .45) and 95% vs 83% (<em>P</em> < .001) for the Apple Watch 6, 91% vs 67% (<em>P</em> < .01) and 94% vs 82% (<em>P</em> < .001) for the Fitbit Sense, and 86% vs 82% (<em>P</em> = .63) and 94% vs 80% (<em>P</em> < .001) for the Samsung Galaxy Watch3, respectively. In addition, the proportion of SL-ECGs with an inconclusive diagnosis (1.2%) was significantly lower for all smart devices using the AI-based algorithm compared to manufacturer’s algorithms (14%–17%), <em>P</em> < .001.</p></div><div><h3>Conclusion</h3><p>A novel AI algorithm reduced the rate of inconclusive SL-ECG diagnosis massively while maintaining sensitivity and improving the specificity compared to the manufacturers’ algorithms.</p></div>","PeriodicalId":72527,"journal":{"name":"Cardiovascular digital health journal","volume":"5 1","pages":"Pages 29-35"},"PeriodicalIF":2.6000,"publicationDate":"2024-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666693623001111/pdfft?md5=50dd82b73a9643b290710c2bd2ea7a1f&pid=1-s2.0-S2666693623001111-main.pdf","citationCount":"0","resultStr":"{\"title\":\"Reducing the burden of inconclusive smart device single-lead ECG tracings via a novel artificial intelligence algorithm\",\"authors\":\"Simon Weidlich MD , Diego Mannhart MD , Alan Kennedy PhD , Peter Doggart , Teodor Serban MD , Sven Knecht DSc-PhD , Jeanne Du Fay de Lavallaz MD-PhD , Michael Kühne MD , Christian Sticherling MD , Patrick Badertscher MD\",\"doi\":\"10.1016/j.cvdhj.2023.12.003\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h3>Background</h3><p>Multiple smart devices capable of automatically detecting atrial fibrillation (AF) based on single-lead electrocardiograms (SL-ECG) are presently available. The rate of inconclusive tracings by manufacturers’ algorithms is currently too high to be clinically useful.</p></div><div><h3>Method</h3><p>This is a prospective, observational study enrolling patients presenting to a cardiology service at a tertiary referral center. We assessed the clinical value of applying a smart device artificial intelligence (AI)-based algorithm for detecting AF from 4 commercially available smart devices (AliveCor KardiaMobile, Apple Watch 6, Fitbit Sense, and Samsung Galaxy Watch3). Patients underwent a nearly simultaneous 12-lead ECG and 4 smart device SL-ECGs. The novel AI algorithm (PulseAI, Belfast, United Kingdom) was compared with each manufacturer’s algorithm.</p></div><div><h3>Results</h3><p>We enrolled 206 patients (31% female, median age 64 years). AF was present in 60 patients (29%). Sensitivity and specificity for the detection of AF by the novel AI algorithm vs manufacturer algorithm were 88% vs 81% (<em>P</em> = .34) and 97% vs 77% (<em>P</em> < .001) for the AliveCor KardiaMobile, 86% vs 81% (<em>P</em> = .45) and 95% vs 83% (<em>P</em> < .001) for the Apple Watch 6, 91% vs 67% (<em>P</em> < .01) and 94% vs 82% (<em>P</em> < .001) for the Fitbit Sense, and 86% vs 82% (<em>P</em> = .63) and 94% vs 80% (<em>P</em> < .001) for the Samsung Galaxy Watch3, respectively. In addition, the proportion of SL-ECGs with an inconclusive diagnosis (1.2%) was significantly lower for all smart devices using the AI-based algorithm compared to manufacturer’s algorithms (14%–17%), <em>P</em> < .001.</p></div><div><h3>Conclusion</h3><p>A novel AI algorithm reduced the rate of inconclusive SL-ECG diagnosis massively while maintaining sensitivity and improving the specificity compared to the manufacturers’ algorithms.</p></div>\",\"PeriodicalId\":72527,\"journal\":{\"name\":\"Cardiovascular digital health journal\",\"volume\":\"5 1\",\"pages\":\"Pages 29-35\"},\"PeriodicalIF\":2.6000,\"publicationDate\":\"2024-02-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.sciencedirect.com/science/article/pii/S2666693623001111/pdfft?md5=50dd82b73a9643b290710c2bd2ea7a1f&pid=1-s2.0-S2666693623001111-main.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Cardiovascular digital health journal\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2666693623001111\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"CARDIAC & CARDIOVASCULAR SYSTEMS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Cardiovascular digital health journal","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2666693623001111","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CARDIAC & CARDIOVASCULAR SYSTEMS","Score":null,"Total":0}
引用次数: 0
摘要
背景目前有多种智能设备能够根据单导联心电图（SL-ECG）自动检测心房颤动（AF）。方法这是一项前瞻性观察研究，研究对象是在一家三级转诊中心心脏科就诊的患者。我们评估了应用基于智能设备的人工智能（AI）算法检测四种市售智能设备（AliveCor KardiaMobile、Apple Watch 6、Fitbit Sense 和 Samsung Galaxy Watch3）房颤的临床价值。患者几乎同时接受了 12 导联心电图和 4 种智能设备 SL-ECG 检查。新型人工智能算法（PulseAI，英国贝尔法斯特）与各制造商的算法进行了比较。60名患者（29%）存在房颤。新型人工智能算法与制造商算法检测房颤的灵敏度和特异度分别为：AliveCor KardiaMobile 为 88% vs 81% (P = .34) 和 97% vs 77% (P < .001)，AliveCor KardiaMobile 为 86% vs 81% (P = .45）和 95% vs 83% (P < .001)，Fitbit Sense 分别为 91% vs 67% (P < .01) 和 94% vs 82% (P < .001)，三星 Galaxy Watch3 分别为 86% vs 82% (P = .63) 和 94% vs 80% (P < .001)。此外，与制造商的算法（14%-17%）相比，使用基于人工智能算法的所有智能设备中诊断不确定的 SL-ECG 的比例（1.2%）显著降低，P < .001 结论与制造商的算法相比，新型人工智能算法在保持灵敏度和提高特异性的同时，大幅降低了诊断不确定的 SL-ECG 的比例。
Reducing the burden of inconclusive smart device single-lead ECG tracings via a novel artificial intelligence algorithm
Background
Multiple smart devices capable of automatically detecting atrial fibrillation (AF) based on single-lead electrocardiograms (SL-ECG) are presently available. The rate of inconclusive tracings by manufacturers’ algorithms is currently too high to be clinically useful.
Method
This is a prospective, observational study enrolling patients presenting to a cardiology service at a tertiary referral center. We assessed the clinical value of applying a smart device artificial intelligence (AI)-based algorithm for detecting AF from 4 commercially available smart devices (AliveCor KardiaMobile, Apple Watch 6, Fitbit Sense, and Samsung Galaxy Watch3). Patients underwent a nearly simultaneous 12-lead ECG and 4 smart device SL-ECGs. The novel AI algorithm (PulseAI, Belfast, United Kingdom) was compared with each manufacturer’s algorithm.
Results
We enrolled 206 patients (31% female, median age 64 years). AF was present in 60 patients (29%). Sensitivity and specificity for the detection of AF by the novel AI algorithm vs manufacturer algorithm were 88% vs 81% (P = .34) and 97% vs 77% (P < .001) for the AliveCor KardiaMobile, 86% vs 81% (P = .45) and 95% vs 83% (P < .001) for the Apple Watch 6, 91% vs 67% (P < .01) and 94% vs 82% (P < .001) for the Fitbit Sense, and 86% vs 82% (P = .63) and 94% vs 80% (P < .001) for the Samsung Galaxy Watch3, respectively. In addition, the proportion of SL-ECGs with an inconclusive diagnosis (1.2%) was significantly lower for all smart devices using the AI-based algorithm compared to manufacturer’s algorithms (14%–17%), P < .001.
Conclusion
A novel AI algorithm reduced the rate of inconclusive SL-ECG diagnosis massively while maintaining sensitivity and improving the specificity compared to the manufacturers’ algorithms.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
