Comparing code-free deep learning models to expert-designed models for detecting retinal diseases from optical coherence tomography.

IF 1.9 Q2 OPHTHALMOLOGY
Samir Touma, Badr Ait Hammou, Fares Antaki, Marie Carole Boucher, Renaud Duval
{"title":"Comparing code-free deep learning models to expert-designed models for detecting retinal diseases from optical coherence tomography.","authors":"Samir Touma, Badr Ait Hammou, Fares Antaki, Marie Carole Boucher, Renaud Duval","doi":"10.1186/s40942-024-00555-3","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Code-free deep learning (CFDL) is a novel tool in artificial intelligence (AI). This study directly compared the discriminative performance of CFDL models designed by ophthalmologists without coding experience against bespoke models designed by AI experts in detecting retinal pathologies from optical coherence tomography (OCT) videos and fovea-centered images.</p><p><strong>Methods: </strong>Using the same internal dataset of 1,173 OCT macular videos and fovea-centered images, model development was performed simultaneously but independently by an ophthalmology resident (CFDL models) and a postdoctoral researcher with expertise in AI (bespoke models). We designed a multi-class model to categorize video and fovea-centered images into five labels: normal retina, macular hole, epiretinal membrane, wet age-related macular degeneration and diabetic macular edema. We qualitatively compared point estimates of the performance metrics of the CFDL and bespoke models.</p><p><strong>Results: </strong>For videos, the CFDL model demonstrated excellent discriminative performance, even outperforming the bespoke models for some metrics: area under the precision-recall curve was 0.984 (vs. 0.901), precision and sensitivity were both 94.1% (vs. 94.2%) and accuracy was 94.1% (vs. 96.7%). The fovea-centered CFDL model overall performed better than video-based model and was as accurate as the best bespoke model.</p><p><strong>Conclusion: </strong>This comparative study demonstrated that code-free models created by clinicians without coding expertise perform as accurately as expert-designed bespoke models at classifying various retinal pathologies from OCT videos and images. CFDL represents a step forward towards the democratization of AI in medicine, although its numerous limitations must be carefully addressed to ensure its effective application in healthcare.</p>","PeriodicalId":14289,"journal":{"name":"International Journal of Retina and Vitreous","volume":null,"pages":null},"PeriodicalIF":1.9000,"publicationDate":"2024-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11055378/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Retina and Vitreous","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1186/s40942-024-00555-3","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"OPHTHALMOLOGY","Score":null,"Total":0}
引用次数: 0

Abstract

Background: Code-free deep learning (CFDL) is a novel tool in artificial intelligence (AI). This study directly compared the discriminative performance of CFDL models designed by ophthalmologists without coding experience against bespoke models designed by AI experts in detecting retinal pathologies from optical coherence tomography (OCT) videos and fovea-centered images.

Methods: Using the same internal dataset of 1,173 OCT macular videos and fovea-centered images, model development was performed simultaneously but independently by an ophthalmology resident (CFDL models) and a postdoctoral researcher with expertise in AI (bespoke models). We designed a multi-class model to categorize video and fovea-centered images into five labels: normal retina, macular hole, epiretinal membrane, wet age-related macular degeneration and diabetic macular edema. We qualitatively compared point estimates of the performance metrics of the CFDL and bespoke models.

Results: For videos, the CFDL model demonstrated excellent discriminative performance, even outperforming the bespoke models for some metrics: area under the precision-recall curve was 0.984 (vs. 0.901), precision and sensitivity were both 94.1% (vs. 94.2%) and accuracy was 94.1% (vs. 96.7%). The fovea-centered CFDL model overall performed better than video-based model and was as accurate as the best bespoke model.

Conclusion: This comparative study demonstrated that code-free models created by clinicians without coding expertise perform as accurately as expert-designed bespoke models at classifying various retinal pathologies from OCT videos and images. CFDL represents a step forward towards the democratization of AI in medicine, although its numerous limitations must be carefully addressed to ensure its effective application in healthcare.

比较无代码深度学习模型和专家设计的模型,从光学相干断层扫描中检测视网膜疾病。
背景:无代码深度学习(CFDL)是人工智能(AI)领域的一种新型工具。本研究直接比较了没有编码经验的眼科医生设计的 CFDL 模型与人工智能专家设计的定制模型在从光学相干断层扫描(OCT)视频和以眼窝为中心的图像中检测视网膜病变方面的判别性能:使用同一内部数据集(1,173 个 OCT 黄斑视频和以眼窝为中心的图像),由一名眼科住院医师(CFDL 模型)和一名具有人工智能专业知识的博士后研究员(定制模型)同时独立进行模型开发。我们设计了一个多类模型,将视频和以眼窝为中心的图像分为五个标签:正常视网膜、黄斑孔、视网膜外膜、湿性年龄相关性黄斑变性和糖尿病性黄斑水肿。我们对 CFDL 模型和定制模型的性能指标点估计值进行了定性比较:在视频方面,CFDL 模型表现出卓越的判别性能,甚至在某些指标上优于定制模型:精确度-召回曲线下面积为 0.984(vs.0.901),精确度和灵敏度均为 94.1%(vs.94.2%),准确度为 94.1%(vs.96.7%)。以眼窝为中心的 CFDL 模型总体表现优于基于视频的模型,其准确性与最佳定制模型相当:这项比较研究表明,在对 OCT 视频和图像中的各种视网膜病变进行分类时,没有编码专业知识的临床医生创建的无编码模型与专家设计的定制模型一样准确。CFDL代表着人工智能在医学领域的民主化向前迈进了一步,但要确保其在医疗保健领域的有效应用,还必须认真解决其诸多局限性。
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
CiteScore
3.50
自引率
4.30%
发文量
81
审稿时长
19 weeks
期刊介绍: International Journal of Retina and Vitreous focuses on the ophthalmic subspecialty of vitreoretinal disorders. The journal presents original articles on new approaches to diagnosis, outcomes of clinical trials, innovations in pharmacological therapy and surgical techniques, as well as basic science advances that impact clinical practice. Topical areas include, but are not limited to: -Imaging of the retina, choroid and vitreous -Innovations in optical coherence tomography (OCT) -Small-gauge vitrectomy, retinal detachment, chromovitrectomy -Electroretinography (ERG), microperimetry, other functional tests -Intraocular tumors -Retinal pharmacotherapy & drug delivery -Diabetic retinopathy & other vascular diseases -Age-related macular degeneration (AMD) & other macular entities
×
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