{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,1,16]],"date-time":"2026-01-16T13:08:03Z","timestamp":1768568883163,"version":"3.49.0"},"posted":{"date-parts":[[2025,6,4]]},"group-title":"Nephrology","reference-count":30,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,6,4]],"date-time":"2025-06-04T00:00:00Z","timestamp":1748995200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.medrxiv.org\/about\/FAQ#license"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,6,11]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Introduction<\/jats:title>\n                  <jats:p>Histological analysis of kidney biopsies is crucial in diagnosing kidney diseases and predicting clinical outcomes. Recently developed pathology foundation models, pretrained on large-scale pathology datasets, have demonstrated excellent performance in various downstream applications. This study evaluated the utility of pathology foundation models combined with multiple instance learning (MIL) for kidney pathology analysis.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We used 242 hematoxylin and eosin (H&amp;E)-stained whole slide images (WSIs) from the Kidney Precision Medicine Project (KPMP) and Japan-Pathology Artificial Intelligence Diagnostics Project (JP-AID) databases as the development cohort, comprising 47 healthy controls, 35 acute interstitial nephritis, and 160 diabetic kidney disease (DKD) slides. External validation was performed using 83 WSIs from the University of Tokyo Hospital (UT dataset). Diagnoses were based on adjudicated diagnoses (KPMP) or expert pathologists-derived diagnoses (JP-AID and UT). Pretrained pathology foundation models were utilized as patch encoders and compared with ImageNet-pretrained ResNet50.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>In internal validation, all foundation models outperformed ResNet50, achieving area under the receiver operating characteristic curve (AUROC) over 0.980. In external validation, the performance of ResNet50 markedly dropped (AUROC = 0.768), whereas all foundation models maintained higher performance (AUROC over 0.800). Visualization with attention heatmaps confirmed that foundation models accurately recognized diagnostically relevant structures. Additionally, foundation models outperformed ResNet50 in predicting severe proteinuria among DKD cases from KPMP dataset.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>We successfully integrated pathology foundation models with MIL to achieve robust diagnostic performance, even when trained on a relatively small dataset, highlighting their potential for real-world clinical applications. Key words: artificial intelligence, renal pathology, foundation model, multiple instance learning<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2025.06.03.25328845","type":"posted-content","created":{"date-parts":[[2025,6,4]],"date-time":"2025-06-04T11:30:12Z","timestamp":1749036612000},"source":"Crossref","is-referenced-by-count":0,"title":["Multiple instance learning using pathology foundation models effectively predicts kidney disease diagnosis and clinical classification"],"prefix":"10.64898","author":[{"given":"Yu","family":"Kurata","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Imari","family":"Mimura","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Satoshi","family":"Kodera","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hiroyuki","family":"Abe","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Daisuke","family":"Yamada","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Haruki","family":"Kume","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tetsuo","family":"Ushiku","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tetsuhiro","family":"Tanaka","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Norihiko","family":"Takeda","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Masaomi","family":"Nangaku","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2025061306101237000_2025.06.03.25328845v2.1","doi-asserted-by":"crossref","first-page":"86","DOI":"10.1016\/j.kint.2020.07.044","article-title":"Development and evaluation of deep learning\u2013based segmentation of histologic structures in the kidney cortex with multiple histologic stains","volume":"99","year":"2021","journal-title":"Kidney Int"},{"key":"2025061306101237000_2025.06.03.25328845v2.2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-36173-0"},{"key":"2025061306101237000_2025.06.03.25328845v2.3","doi-asserted-by":"crossref","first-page":"291","DOI":"10.1097\/MNH.0000000000000973","article-title":"Deep learning applications for kidney histology analysis","volume":"33","year":"2024","journal-title":"Curr Opin Nephrol Hypertens"},{"key":"2025061306101237000_2025.06.03.25328845v2.4","unstructured":"Chen T , Kornblith S , Norouzi M , et al.: A Simple Framework for Contrastive Learning of Visual Representations. In Proceedings of the 37th International Conference on Machine Learning (vol 119), edited by Hal D, III , Aarti S , Proceedings of Machine Learning Research, PMLR, 2020, pp 1597\u20131607"},{"key":"2025061306101237000_2025.06.03.25328845v2.5","doi-asserted-by":"crossref","first-page":"471","DOI":"10.1681\/ASN.0000000514","article-title":"Self-Supervised Learning for Feature Extraction from Glomerular Images and Disease Classification with Minimal Annotations","volume":"36","year":"2025","journal-title":"J Am Soc Nephrol"},{"key":"2025061306101237000_2025.06.03.25328845v2.6","unstructured":"Dosovitskiy A , Beyer L , Kolesnikov A , et al. An Image is Worth 16\u00d716 Words: Transformers for Image Recognition at Scale. arXiv 2020; abs\/2010.11929."},{"key":"2025061306101237000_2025.06.03.25328845v2.7","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-02857-3"},{"key":"2025061306101237000_2025.06.03.25328845v2.8","unstructured":"Vorontsov E , Bozkurt A , Casson A , et al. Virchow: A Million-Slide Digital Pathology Foundation Model. arXiv 2023; abs\/2309.07778."},{"key":"2025061306101237000_2025.06.03.25328845v2.9","unstructured":"Zimmermann E , Vorontsov E , Viret J , et al. Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology. arXiv 2024; abs\/2408.00738."},{"key":"2025061306101237000_2025.06.03.25328845v2.10","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-024-07441-w"},{"key":"2025061306101237000_2025.06.03.25328845v2.11","doi-asserted-by":"crossref","unstructured":"Filiot A , Ghermi R , Olivier A , et al. Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling. medRxiv 2023.07.21.23292757.","DOI":"10.1101\/2023.07.21.23292757"},{"key":"2025061306101237000_2025.06.03.25328845v2.12","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.102762"},{"key":"2025061306101237000_2025.06.03.25328845v2.13","doi-asserted-by":"crossref","unstructured":"He K , Zhang X , Ren S , et al. Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2015: 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"2025061306101237000_2025.06.03.25328845v2.14","doi-asserted-by":"publisher","DOI":"10.1016\/j.kint.2020.08.039"},{"key":"2025061306101237000_2025.06.03.25328845v2.15","first-page":"40","article-title":"Development of artificial intelligence to help pathological diagnosis-Japan pathology AI diagnostics (JP-AID) project","volume":"6","year":"2019","journal-title":"Impact"},{"key":"2025061306101237000_2025.06.03.25328845v2.16","doi-asserted-by":"publisher","DOI":"10.1186\/s12859-024-05758-x"},{"key":"2025061306101237000_2025.06.03.25328845v2.17","doi-asserted-by":"publisher","DOI":"10.1109\/TSMC.1979.4310076"},{"key":"2025061306101237000_2025.06.03.25328845v2.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41379-021-00859-x"},{"key":"2025061306101237000_2025.06.03.25328845v2.19","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-03141-0"},{"key":"2025061306101237000_2025.06.03.25328845v2.20","unstructured":"Wolflein G , Ferber D , Meneghetti AR , et al. Benchmarking Pathology Feature Extractors for Whole Slide Image Classification. arXiv 2024; arXiv:2311.11772v5."},{"key":"2025061306101237000_2025.06.03.25328845v2.21","unstructured":"TorchVision: PyTorch\u2019s Computer Vision library. In, GitHub repository, GitHub, 2016"},{"key":"2025061306101237000_2025.06.03.25328845v2.22","doi-asserted-by":"crossref","unstructured":"Wolf T , Debut L , Sanh V , et al. HuggingFace\u2019s Transformers: State-of-the-art Natural Language Processing. arXiv 2019; abs\/1910.03771.","DOI":"10.18653\/v1\/2020.emnlp-demos.6"},{"key":"2025061306101237000_2025.06.03.25328845v2.23","unstructured":"Maximilian Ilse JMT , Max Welling . Attention-based Deep Multiple Instance Learning. arXiv 2018; 1802.04712v4."},{"key":"2025061306101237000_2025.06.03.25328845v2.24","unstructured":"Zhuchen Shao HB , Yang Chen , Yifeng Wang , Jian Zhang , Xiangyang Ji , Yongbing Zhang . TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classification. arXiv 2021; 2106.00908v2."},{"key":"2025061306101237000_2025.06.03.25328845v2.25","doi-asserted-by":"crossref","first-page":"555","DOI":"10.1038\/s41551-020-00682-w","article-title":"Data-efficient and weakly supervised computational pathology on whole-slide images","volume":"5","year":"2021","journal-title":"Nat Biomed Eng"},{"key":"2025061306101237000_2025.06.03.25328845v2.26","unstructured":"KDIGO Clinical Practice Guideline for Acute Kidney Injury. Kidney Int Suppl 2012; 2."},{"key":"2025061306101237000_2025.06.03.25328845v2.27","doi-asserted-by":"crossref","unstructured":"Macenko M , Niethammer M , Marron JS , et al. A method for normalizing histology slides for quantitative analysis. 2009 IEEE International Symposium on Biomedical Imaging: From Nano to Macro 2009: 1107\u20131110.","DOI":"10.1109\/ISBI.2009.5193250"},{"key":"2025061306101237000_2025.06.03.25328845v2.28","doi-asserted-by":"crossref","first-page":"15","DOI":"10.1002\/path.5797","article-title":"The utility of color normalization for AI-based diagnosis of hematoxylin and eosin-stained pathology images","volume":"256","year":"2022","journal-title":"J Pathol"},{"key":"2025061306101237000_2025.06.03.25328845v2.29","unstructured":"Edwin D. de Jong EM, Jonas Teuwen. Current Pathology Foundation Models are unrobust to Medical Center Differences. arXiv 2025; arXiv:2501.18055."},{"key":"2025061306101237000_2025.06.03.25328845v2.30","doi-asserted-by":"crossref","first-page":"714","DOI":"10.1016\/j.kint.2024.12.007","article-title":"Multi-stain deep learning prediction model of treatment response in lupus nephritis based on renal histopathology","volume":"107","year":"2025","journal-title":"Kidney Int"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.06.03.25328845","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:37:35Z","timestamp":1768509455000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.06.03.25328845"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,4]]},"references-count":30,"URL":"https:\/\/doi.org\/10.1101\/2025.06.03.25328845","relation":{},"subject":[],"published":{"date-parts":[[2025,6,4]]},"subtype":"preprint"}}