使用统一的知识蒸馏预训练框架的可推广的病理学基础模型

IF 26.8 1区 医学 Q1 ENGINEERING, BIOMEDICAL
Jiabo Ma, Zhengrui Guo, Fengtao Zhou, Yihui Wang, Yingxue Xu, Jinbang Li, Fang Yan, Yu Cai, Zhengjie Zhu, Cheng Jin, Yi Lin, Xinrui Jiang, Chenglong Zhao, Danyi Li, Anjia Han, Zhenhui Li, Ronald Cheong Kin Chan, Jiguang Wang, Peng Fei, Kwang-Ting Cheng, Shaoting Zhang, Li Liang, Hao Chen
{"title":"使用统一的知识蒸馏预训练框架的可推广的病理学基础模型","authors":"Jiabo Ma, Zhengrui Guo, Fengtao Zhou, Yihui Wang, Yingxue Xu, Jinbang Li, Fang Yan, Yu Cai, Zhengjie Zhu, Cheng Jin, Yi Lin, Xinrui Jiang, Chenglong Zhao, Danyi Li, Anjia Han, Zhenhui Li, Ronald Cheong Kin Chan, Jiguang Wang, Peng Fei, Kwang-Ting Cheng, Shaoting Zhang, Li Liang, Hao Chen","doi":"10.1038/s41551-025-01488-4","DOIUrl":null,"url":null,"abstract":"<p>The generalization ability of foundation models in the field of computational pathology (CPath) is crucial for their clinical success. However, current foundation models have only been evaluated on a limited type and number of tasks, leaving their generalization ability unclear. We establish a comprehensive benchmark to evaluate the performance of off-the-shelf foundation models across six distinct clinical task types, encompassing a total of 72 specific tasks. Our findings reveal that existing foundation models excel at certain task types but struggle to effectively handle the full breadth of clinical tasks. To improve the generalization of pathology foundation models, we propose a unified knowledge distillation framework consisting of both expert and self knowledge distillation, where the former allows the model to learn from the knowledge of multiple expert models, while the latter leverages self distillation to enable image representation learning via local–global alignment. On the basis of this framework, we develop a Generalizable Pathology Foundation Model (GPFM). Evaluated on the established benchmark, GPFM achieves an average rank of 1.6, ranking first in 42 tasks, positioning it as a promising method for feature representation in CPath.</p>","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"59 1","pages":""},"PeriodicalIF":26.8000,"publicationDate":"2025-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A generalizable pathology foundation model using a unified knowledge distillation pretraining framework\",\"authors\":\"Jiabo Ma, Zhengrui Guo, Fengtao Zhou, Yihui Wang, Yingxue Xu, Jinbang Li, Fang Yan, Yu Cai, Zhengjie Zhu, Cheng Jin, Yi Lin, Xinrui Jiang, Chenglong Zhao, Danyi Li, Anjia Han, Zhenhui Li, Ronald Cheong Kin Chan, Jiguang Wang, Peng Fei, Kwang-Ting Cheng, Shaoting Zhang, Li Liang, Hao Chen\",\"doi\":\"10.1038/s41551-025-01488-4\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>The generalization ability of foundation models in the field of computational pathology (CPath) is crucial for their clinical success. However, current foundation models have only been evaluated on a limited type and number of tasks, leaving their generalization ability unclear. We establish a comprehensive benchmark to evaluate the performance of off-the-shelf foundation models across six distinct clinical task types, encompassing a total of 72 specific tasks. Our findings reveal that existing foundation models excel at certain task types but struggle to effectively handle the full breadth of clinical tasks. To improve the generalization of pathology foundation models, we propose a unified knowledge distillation framework consisting of both expert and self knowledge distillation, where the former allows the model to learn from the knowledge of multiple expert models, while the latter leverages self distillation to enable image representation learning via local–global alignment. On the basis of this framework, we develop a Generalizable Pathology Foundation Model (GPFM). Evaluated on the established benchmark, GPFM achieves an average rank of 1.6, ranking first in 42 tasks, positioning it as a promising method for feature representation in CPath.</p>\",\"PeriodicalId\":19063,\"journal\":{\"name\":\"Nature Biomedical Engineering\",\"volume\":\"59 1\",\"pages\":\"\"},\"PeriodicalIF\":26.8000,\"publicationDate\":\"2025-09-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Nature Biomedical Engineering\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://doi.org/10.1038/s41551-025-01488-4\",\"RegionNum\":1,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, BIOMEDICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Nature Biomedical Engineering","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1038/s41551-025-01488-4","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, BIOMEDICAL","Score":null,"Total":0}
引用次数: 0

摘要

基础模型在计算病理学(CPath)领域的泛化能力是其临床成功的关键。然而,目前的基础模型只在有限类型和数量的任务上进行了评估,使其泛化能力不明确。我们建立了一个全面的基准来评估现成的基础模型在六个不同的临床任务类型中的表现,包括总共72个特定的任务。我们的研究结果表明,现有的基础模型在某些任务类型上表现出色,但难以有效地处理临床任务的全部广度。为了提高病理基础模型的泛化能力,我们提出了一个由专家知识蒸馏和自知识蒸馏组成的统一知识蒸馏框架,前者允许模型从多个专家模型的知识中学习,而后者利用自蒸馏通过局部-全局对齐实现图像表示学习。在此框架的基础上,我们开发了一个可推广的病理基础模型(GPFM)。在建立的基准上进行评估,GPFM的平均排名为1.6,在42个任务中排名第一,是一种很有前途的CPath特征表示方法。
本文章由计算机程序翻译,如有差异,请以英文原文为准。

A generalizable pathology foundation model using a unified knowledge distillation pretraining framework

A generalizable pathology foundation model using a unified knowledge distillation pretraining framework

The generalization ability of foundation models in the field of computational pathology (CPath) is crucial for their clinical success. However, current foundation models have only been evaluated on a limited type and number of tasks, leaving their generalization ability unclear. We establish a comprehensive benchmark to evaluate the performance of off-the-shelf foundation models across six distinct clinical task types, encompassing a total of 72 specific tasks. Our findings reveal that existing foundation models excel at certain task types but struggle to effectively handle the full breadth of clinical tasks. To improve the generalization of pathology foundation models, we propose a unified knowledge distillation framework consisting of both expert and self knowledge distillation, where the former allows the model to learn from the knowledge of multiple expert models, while the latter leverages self distillation to enable image representation learning via local–global alignment. On the basis of this framework, we develop a Generalizable Pathology Foundation Model (GPFM). Evaluated on the established benchmark, GPFM achieves an average rank of 1.6, ranking first in 42 tasks, positioning it as a promising method for feature representation in CPath.

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
Nature Biomedical Engineering
Nature Biomedical Engineering Medicine-Medicine (miscellaneous)
CiteScore
45.30
自引率
1.10%
发文量
138
期刊介绍: Nature Biomedical Engineering is an online-only monthly journal that was launched in January 2017. It aims to publish original research, reviews, and commentary focusing on applied biomedicine and health technology. The journal targets a diverse audience, including life scientists who are involved in developing experimental or computational systems and methods to enhance our understanding of human physiology. It also covers biomedical researchers and engineers who are engaged in designing or optimizing therapies, assays, devices, or procedures for diagnosing or treating diseases. Additionally, clinicians, who make use of research outputs to evaluate patient health or administer therapy in various clinical settings and healthcare contexts, are also part of the target audience.
×
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:604180095
Book学术官方微信