揭示神经网络的黑盒用于通过基于闭塞的解释性检测整张幻灯片图像中的前列腺癌症。

IF 4.5 2区 生物学 Q1 BIOCHEMICAL RESEARCH METHODS New biotechnology Pub Date : 2023-10-02 DOI:10.1016/j.nbt.2023.09.008
Matej Gallo , Vojtěch Krajňanský , Rudolf Nenutil , Petr Holub , Tomáš Brázdil
{"title":"揭示神经网络的黑盒用于通过基于闭塞的解释性检测整张幻灯片图像中的前列腺癌症。","authors":"Matej Gallo ,&nbsp;Vojtěch Krajňanský ,&nbsp;Rudolf Nenutil ,&nbsp;Petr Holub ,&nbsp;Tomáš Brázdil","doi":"10.1016/j.nbt.2023.09.008","DOIUrl":null,"url":null,"abstract":"<div><p>Diagnostic histopathology faces increasing demands due to aging populations and expanding healthcare programs. Semi-automated diagnostic systems employing deep learning methods are one approach to alleviate this pressure. The learning models for histopathology are inherently complex and opaque from the user's perspective. Hence different methods have been developed to interpret their behavior. However, relatively limited attention has been devoted to the connection between interpretation methods and the knowledge of experienced pathologists. The main contribution of this paper is a method for comparing morphological patterns used by expert pathologists to detect cancer with the patterns identified as important for inference of learning models. Given the patch-based nature of processing large-scale histopathological imaging, we have been able to show statistically that the VGG16 model could utilize all the structures that are observable by the pathologist, given the patch size and scan resolution. The results show that the neural network approach to recognizing prostatic cancer is similar to that of a pathologist at medium optical resolution. The saliency maps identified several prevailing histomorphological features characterizing carcinoma, e.g., single-layered epithelium, small lumina, and hyperchromatic nuclei with halo. A convincing finding was the recognition of their mimickers in non-neoplastic tissue. The method can also identify differences, i.e., standard patterns not used by the learning models and new patterns not yet used by pathologists. Saliency maps provide added value for automated digital pathology to analyze and fine-tune deep learning systems and improve trust in computer-based decisions.</p></div>","PeriodicalId":19190,"journal":{"name":"New biotechnology","volume":null,"pages":null},"PeriodicalIF":4.5000,"publicationDate":"2023-10-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability\",\"authors\":\"Matej Gallo ,&nbsp;Vojtěch Krajňanský ,&nbsp;Rudolf Nenutil ,&nbsp;Petr Holub ,&nbsp;Tomáš Brázdil\",\"doi\":\"10.1016/j.nbt.2023.09.008\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Diagnostic histopathology faces increasing demands due to aging populations and expanding healthcare programs. Semi-automated diagnostic systems employing deep learning methods are one approach to alleviate this pressure. The learning models for histopathology are inherently complex and opaque from the user's perspective. Hence different methods have been developed to interpret their behavior. However, relatively limited attention has been devoted to the connection between interpretation methods and the knowledge of experienced pathologists. The main contribution of this paper is a method for comparing morphological patterns used by expert pathologists to detect cancer with the patterns identified as important for inference of learning models. Given the patch-based nature of processing large-scale histopathological imaging, we have been able to show statistically that the VGG16 model could utilize all the structures that are observable by the pathologist, given the patch size and scan resolution. The results show that the neural network approach to recognizing prostatic cancer is similar to that of a pathologist at medium optical resolution. The saliency maps identified several prevailing histomorphological features characterizing carcinoma, e.g., single-layered epithelium, small lumina, and hyperchromatic nuclei with halo. A convincing finding was the recognition of their mimickers in non-neoplastic tissue. The method can also identify differences, i.e., standard patterns not used by the learning models and new patterns not yet used by pathologists. Saliency maps provide added value for automated digital pathology to analyze and fine-tune deep learning systems and improve trust in computer-based decisions.</p></div>\",\"PeriodicalId\":19190,\"journal\":{\"name\":\"New biotechnology\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":4.5000,\"publicationDate\":\"2023-10-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"New biotechnology\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1871678423000511\",\"RegionNum\":2,\"RegionCategory\":\"生物学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"BIOCHEMICAL RESEARCH METHODS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"New biotechnology","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1871678423000511","RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BIOCHEMICAL RESEARCH METHODS","Score":null,"Total":0}
引用次数: 0

摘要

由于人口老龄化和医疗保健项目的扩大,诊断组织病理学面临着越来越多的需求。采用深度学习方法的半自动化诊断系统是缓解这种压力的一种方法。从用户的角度来看,组织病理学的学习模型本质上是复杂和不透明的。因此,人们开发了不同的方法来解释它们的行为。然而,对解释方法与经验丰富的病理学家的知识之间的联系的关注相对有限。本文的主要贡献是一种将病理学家用于检测癌症的形态学模式与被确定为对学习模型的推断很重要的模式进行比较的方法。考虑到处理大规模组织病理学成像的基于斑块的性质,我们已经能够从统计上表明,VGG16模型可以利用病理学家可以观察到的所有结构,考虑到斑块的大小和扫描分辨率。结果表明,在中等光学分辨率下,神经网络识别前列腺癌症的方法与病理学家的方法相似。显著性图确定了几种主要的癌症组织形态学特征,例如单层上皮、小管腔和带晕染的深染细胞核。一个令人信服的发现是在非肿瘤组织中识别出了它们的拟态物。该方法还可以识别差异,即学习模型未使用的标准模式和病理学家尚未使用的新模式。显著性图为自动化数字病理学分析和微调深度学习系统提供了附加值,并提高了对基于计算机的决策的信任。
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
Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability

Diagnostic histopathology faces increasing demands due to aging populations and expanding healthcare programs. Semi-automated diagnostic systems employing deep learning methods are one approach to alleviate this pressure. The learning models for histopathology are inherently complex and opaque from the user's perspective. Hence different methods have been developed to interpret their behavior. However, relatively limited attention has been devoted to the connection between interpretation methods and the knowledge of experienced pathologists. The main contribution of this paper is a method for comparing morphological patterns used by expert pathologists to detect cancer with the patterns identified as important for inference of learning models. Given the patch-based nature of processing large-scale histopathological imaging, we have been able to show statistically that the VGG16 model could utilize all the structures that are observable by the pathologist, given the patch size and scan resolution. The results show that the neural network approach to recognizing prostatic cancer is similar to that of a pathologist at medium optical resolution. The saliency maps identified several prevailing histomorphological features characterizing carcinoma, e.g., single-layered epithelium, small lumina, and hyperchromatic nuclei with halo. A convincing finding was the recognition of their mimickers in non-neoplastic tissue. The method can also identify differences, i.e., standard patterns not used by the learning models and new patterns not yet used by pathologists. Saliency maps provide added value for automated digital pathology to analyze and fine-tune deep learning systems and improve trust in computer-based decisions.

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
New biotechnology
New biotechnology 生物-生化研究方法
CiteScore
11.40
自引率
1.90%
发文量
77
审稿时长
1 months
期刊介绍: New Biotechnology is the official journal of the European Federation of Biotechnology (EFB) and is published bimonthly. It covers both the science of biotechnology and its surrounding political, business and financial milieu. The journal publishes peer-reviewed basic research papers, authoritative reviews, feature articles and opinions in all areas of biotechnology. It reflects the full diversity of current biotechnology science, particularly those advances in research and practice that open opportunities for exploitation of knowledge, commercially or otherwise, together with news, discussion and comment on broader issues of general interest and concern. The outlook is fully international. The scope of the journal includes the research, industrial and commercial aspects of biotechnology, in areas such as: Healthcare and Pharmaceuticals; Food and Agriculture; Biofuels; Genetic Engineering and Molecular Biology; Genomics and Synthetic Biology; Nanotechnology; Environment and Biodiversity; Biocatalysis; Bioremediation; Process engineering.
期刊最新文献
Beyond viability: Advancing CHO cell culture process strategies to modulate host cell protein levels. Elevated CO2 enhances growth and cyanide assimilation in nitrogen-deficient rice: A transcriptome and metabolomic perspective Knock-out of the major regulator Flo8 in Komagataella phaffii results in unique host strain performance for methanol-free recombinant protein production Impact of hydromechanical stress on CHO cells’ metabolism and productivity: Insights from shake flask cultivations with online monitoring of the respiration activity Assessing the potential of olive mill solid waste as feedstock for methane and volatile fatty acids production via anaerobic bioprocesses
×
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