Accurate detection of atrioventricular septal defect (AVSD) in fetal ultrasound using artificial intelligence

IF 2.3 3区 医学 Q2 CARDIAC & CARDIOVASCULAR SYSTEMS
B. Stos , M. Lévy , E. Héry , I. Durand , E. Askinazi , V. Thorey , M. De Boisredon , C. Gardella
{"title":"Accurate detection of atrioventricular septal defect (AVSD) in fetal ultrasound using artificial intelligence","authors":"B. Stos ,&nbsp;M. Lévy ,&nbsp;E. Héry ,&nbsp;I. Durand ,&nbsp;E. Askinazi ,&nbsp;V. Thorey ,&nbsp;M. De Boisredon ,&nbsp;C. Gardella","doi":"10.1016/j.acvd.2024.07.004","DOIUrl":null,"url":null,"abstract":"<div><h3>Introduction</h3><p>A deep neural network could accurately detect AVSD in 2nd trimester fetal heart ultrasound video clips.</p></div><div><h3>Objective</h3><p>AVSDs are often undetected before birth, with an impact on morbidity and mortality. In addition, detecting AVSD may allow diagnosing genetic disorders (such as Down syndrome), which are frequently associated.</p><p>Here, we aim at evaluating whether a deep neural network (DNN) could identify AVSD in 2nd trimester (2T) fetal ultrasound video clips.</p></div><div><h3>Methods</h3><p>Patients with single pregnancy who had an echocardiography performed at one center (18–25<!--> <!-->weeks GA) were included retrospectively starting from Jan 1, 2021. Based on clinical records, we included consecutive cases of partial or complete AVSD, and consecutive negative cases referred due to family history. This inclusion criterion was used for negative cases to be more representative of the general population, in a center with a high prevalence of CHD since it only receives patients referred for echocardiography.</p><p>Cases were reviewed by one of two fetal echocardiography experts to confirm that the presence or absence of AVSD was documented in at least one video clip. Patients with no such video clip were excluded.</p><p>The DNN takes as input all the recorded video clips of a given examination and outputs the absence or presence of AVSD, or an “inconclusive” output if its confidence is low. The DNN was trained to detect AVSD, as seen on the four-chamber view, on patients not included in the evaluation.</p></div><div><h3>Results</h3><p>We included 26 cases with AVSD and 129 cases without. The DNN achieved an AUC of 97.1%, a sensitivity of 86.4% (95% CI: 66.7–95.3) and a specificity of 95.2% (95% CI: 90.0–97.8), after excluding inconclusive diagnosis. The DNN predicted a conclusive diagnosis in 95.5% of cases (<span><span>Fig. 1</span></span>).</p></div><div><h3>Conclusion</h3><p>A DNN could accurately identify AVSD in 2T fetal echocardiography. These results establish the groundwork for efficient and accurate AI-assisted fetal ultrasound heart screening.</p></div>","PeriodicalId":55472,"journal":{"name":"Archives of Cardiovascular Diseases","volume":"117 8","pages":"Pages S220-S221"},"PeriodicalIF":2.3000,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Archives of Cardiovascular Diseases","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1875213624002250","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CARDIAC & CARDIOVASCULAR SYSTEMS","Score":null,"Total":0}
引用次数: 0

Abstract

Introduction

A deep neural network could accurately detect AVSD in 2nd trimester fetal heart ultrasound video clips.

Objective

AVSDs are often undetected before birth, with an impact on morbidity and mortality. In addition, detecting AVSD may allow diagnosing genetic disorders (such as Down syndrome), which are frequently associated.

Here, we aim at evaluating whether a deep neural network (DNN) could identify AVSD in 2nd trimester (2T) fetal ultrasound video clips.

Methods

Patients with single pregnancy who had an echocardiography performed at one center (18–25 weeks GA) were included retrospectively starting from Jan 1, 2021. Based on clinical records, we included consecutive cases of partial or complete AVSD, and consecutive negative cases referred due to family history. This inclusion criterion was used for negative cases to be more representative of the general population, in a center with a high prevalence of CHD since it only receives patients referred for echocardiography.

Cases were reviewed by one of two fetal echocardiography experts to confirm that the presence or absence of AVSD was documented in at least one video clip. Patients with no such video clip were excluded.

The DNN takes as input all the recorded video clips of a given examination and outputs the absence or presence of AVSD, or an “inconclusive” output if its confidence is low. The DNN was trained to detect AVSD, as seen on the four-chamber view, on patients not included in the evaluation.

Results

We included 26 cases with AVSD and 129 cases without. The DNN achieved an AUC of 97.1%, a sensitivity of 86.4% (95% CI: 66.7–95.3) and a specificity of 95.2% (95% CI: 90.0–97.8), after excluding inconclusive diagnosis. The DNN predicted a conclusive diagnosis in 95.5% of cases (Fig. 1).

Conclusion

A DNN could accurately identify AVSD in 2T fetal echocardiography. These results establish the groundwork for efficient and accurate AI-assisted fetal ultrasound heart screening.

利用人工智能准确检测胎儿超声波中的房室间隔缺损 (AVSD)
导言:深度神经网络可准确检测第二孕期胎儿心脏超声视频片段中的 AVSD。本文旨在评估深度神经网络(DNN)是否能在妊娠后三个月(2T)胎儿超声视频片段中识别出 AVSD。方法回顾性地纳入了自 2021 年 1 月 1 日起在一家中心进行超声心动图检查的单胎妊娠患者(18-25 孕周)。根据临床记录,我们纳入了部分或完全性 AVSD 的连续病例,以及因家族史而转诊的连续阴性病例。阴性病例的纳入标准更能代表普通人群的情况,因为该中心仅接收转诊的超声心动图检查患者,是一个先天性心脏病发病率较高的中心。DNN 将给定检查的所有视频片段作为输入,并输出是否存在房室间SD,如果置信度较低,则输出 "不确定"。对 DNN 进行了训练,以检测未纳入评估的患者四腔切面上的 AVSD。排除不确定诊断后,DNN 的 AUC 为 97.1%,灵敏度为 86.4%(95% CI:66.7-95.3),特异度为 95.2%(95% CI:90.0-97.8)。结论 DNN 可在 2T 胎儿超声心动图中准确识别 AVSD。这些结果为高效、准确的人工智能辅助胎儿超声心脏筛查奠定了基础。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Archives of Cardiovascular Diseases
Archives of Cardiovascular Diseases 医学-心血管系统
CiteScore
4.40
自引率
6.70%
发文量
87
审稿时长
34 days
期刊介绍: The Journal publishes original peer-reviewed clinical and research articles, epidemiological studies, new methodological clinical approaches, review articles and editorials. Topics covered include coronary artery and valve diseases, interventional and pediatric cardiology, cardiovascular surgery, cardiomyopathy and heart failure, arrhythmias and stimulation, cardiovascular imaging, vascular medicine and hypertension, epidemiology and risk factors, and large multicenter studies. Archives of Cardiovascular Diseases also publishes abstracts of papers presented at the annual sessions of the Journées Européennes de la Société Française de Cardiologie and the guidelines edited by the French Society of Cardiology.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信