发现医疗人工智能预测受保护属性的内在机制

Soham Gadgil, Alex J. DeGrave, Roxana Daneshjou, Su-In Lee
{"title":"发现医疗人工智能预测受保护属性的内在机制","authors":"Soham Gadgil, Alex J. DeGrave, Roxana Daneshjou, Su-In Lee","doi":"10.1101/2024.04.09.24305289","DOIUrl":null,"url":null,"abstract":"Recent advances in Artificial Intelligence (AI) have started disrupting the healthcare industry, especially medical imaging, and AI devices are increasingly being deployed into clinical practice. Such classifiers have previously demonstrated the ability to discern a range of protected demographic attributes (like race, age, sex) from medical images with unexpectedly high performance, a sensitive task which is difficult even for trained physicians. Focusing on the task of predicting sex from dermoscopic images of skin lesions, we are successfully able to train high-performing classifiers achieving a ROC-AUC score of ∼0.78. We highlight how incorrect use of these demographic shortcuts can have a detrimental effect on the performance of a clinically relevant downstream task like disease diagnosis under a domain shift. Further, we employ various explainable AI (XAI) techniques to identify specific signals which can be leveraged to predict sex. Finally, we introduce a technique to quantify how much a signal contributes to the classification performance. Using this technique and the signals identified, we are able to explain ∼44% of the total performance. This analysis not only underscores the importance of cautious AI application in healthcare but also opens avenues for improving the transparency and reliability of AI-driven diagnostic tools.","PeriodicalId":501385,"journal":{"name":"medRxiv - Dermatology","volume":"42 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-04-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Discovering mechanisms underlying medical AI prediction of protected attributes\",\"authors\":\"Soham Gadgil, Alex J. DeGrave, Roxana Daneshjou, Su-In Lee\",\"doi\":\"10.1101/2024.04.09.24305289\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recent advances in Artificial Intelligence (AI) have started disrupting the healthcare industry, especially medical imaging, and AI devices are increasingly being deployed into clinical practice. Such classifiers have previously demonstrated the ability to discern a range of protected demographic attributes (like race, age, sex) from medical images with unexpectedly high performance, a sensitive task which is difficult even for trained physicians. Focusing on the task of predicting sex from dermoscopic images of skin lesions, we are successfully able to train high-performing classifiers achieving a ROC-AUC score of ∼0.78. We highlight how incorrect use of these demographic shortcuts can have a detrimental effect on the performance of a clinically relevant downstream task like disease diagnosis under a domain shift. Further, we employ various explainable AI (XAI) techniques to identify specific signals which can be leveraged to predict sex. Finally, we introduce a technique to quantify how much a signal contributes to the classification performance. Using this technique and the signals identified, we are able to explain ∼44% of the total performance. This analysis not only underscores the importance of cautious AI application in healthcare but also opens avenues for improving the transparency and reliability of AI-driven diagnostic tools.\",\"PeriodicalId\":501385,\"journal\":{\"name\":\"medRxiv - Dermatology\",\"volume\":\"42 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-04-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"medRxiv - Dermatology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1101/2024.04.09.24305289\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"medRxiv - Dermatology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1101/2024.04.09.24305289","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

人工智能(AI)的最新进展已经开始颠覆医疗保健行业,尤其是医学影像行业,而人工智能设备也越来越多地被部署到临床实践中。此类分类器以前曾以出乎意料的高性能展示了从医学影像中辨别一系列受保护的人口属性(如种族、年龄、性别)的能力,而这一敏感任务即使对训练有素的医生来说也是困难重重。我们将重点放在了从皮肤病变的皮肤镜图像中预测性别的任务上,成功地训练出了高性能的分类器,其 ROC-AUC 得分为 0.78。我们强调了在领域转移的情况下,不正确使用这些人口统计学捷径会如何对疾病诊断等临床相关下游任务的性能产生不利影响。此外,我们还采用了各种可解释人工智能(XAI)技术来识别可用于预测性别的特定信号。最后,我们介绍了一种量化信号对分类性能贡献程度的技术。利用这种技术和识别出的信号,我们能够解释总性能的 44%。这项分析不仅强调了在医疗保健领域谨慎应用人工智能的重要性,还为提高人工智能驱动的诊断工具的透明度和可靠性开辟了途径。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Discovering mechanisms underlying medical AI prediction of protected attributes
Recent advances in Artificial Intelligence (AI) have started disrupting the healthcare industry, especially medical imaging, and AI devices are increasingly being deployed into clinical practice. Such classifiers have previously demonstrated the ability to discern a range of protected demographic attributes (like race, age, sex) from medical images with unexpectedly high performance, a sensitive task which is difficult even for trained physicians. Focusing on the task of predicting sex from dermoscopic images of skin lesions, we are successfully able to train high-performing classifiers achieving a ROC-AUC score of ∼0.78. We highlight how incorrect use of these demographic shortcuts can have a detrimental effect on the performance of a clinically relevant downstream task like disease diagnosis under a domain shift. Further, we employ various explainable AI (XAI) techniques to identify specific signals which can be leveraged to predict sex. Finally, we introduce a technique to quantify how much a signal contributes to the classification performance. Using this technique and the signals identified, we are able to explain ∼44% of the total performance. This analysis not only underscores the importance of cautious AI application in healthcare but also opens avenues for improving the transparency and reliability of AI-driven diagnostic tools.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
A Two-step Bayesian Mendelian Randomization Study on Cholecystitis and Dermatitis Risk of Major Adverse Cardiovascular Events following Nicotinamide Exposure: Cohort Study Methodological issues in visible LED therapy dermatological research and reporting Leveraging Machine Learning & Mobile Application Technology for Vitiligo Management: A Proof-of-Concept Nationwide melanoma registry databases in real-world settings: a scoping review protocol
×
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
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 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 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1