[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-367571-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-367571-en":131},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","artificial-intelligence-predicts-multiclass-molecular-signatures-and-subtypes-directly-from-breast-cancer-histology-a-multicenter-retrospective-study","Artificial intelligence predicts multiclass molecular signatures and subtypes directly from breast cancer histology - a multicenter retrospective study","","Detection of breast cancer biomarkers creates extra cost and tissue burden. A deep learning approach (BBMIL) is presented to predict classical biomarkers, immunotherapy-associated gene signatures, and prognosis-linked subtypes directly from hematoxylin-and-eosin stained histopathology images. BBMIL achieves the strongest comparative performance for all predicted targets versus multiple alternative algorithms. The study uses whole-slide images and includes tumor region handling through a dedicated ROI prediction model.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/artificial-intelligence-predicts-multiclass-molecular-signatures-and-subtypes-directly-from-breast-cancer-histology-a-multicenter-retrospective-study/367571/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/artificial-intelligence-predicts-multiclass-molecular-signatures-and-subtypes-directly-from-breast-cancer-histology-a-multicenter-retrospective-study/367571.png","ImageObject",300,407,{"name":42,"@type":43},"Bintang","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-27","2026-09-23",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",11,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What problem does the study target in breast cancer biomarker detection?","Question",{"text":63,"@type":64},"Biomarker detection requires additional cost and increases tissue burden, limiting widespread evaluation. The study aims to predict biomarkers directly from standard histopathology images.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"What does BBMIL predict from hematoxylin-and-eosin stained images?",{"text":68,"@type":64},"BBMIL predicts classical biomarkers, immunotherapy-associated gene signatures, and prognosis-associated subtypes. The predicted outputs include established receptors and PAM50 plus immunotherapy-related signatures and novel subtypes.",{"name":70,"@type":61,"acceptedAnswer":71},"How does BBMIL perform compared with other algorithms?",{"text":72,"@type":64},"BBMIL shows the best overall performance among the compared methods for predicting classical biomarkers, immunotherapy-related gene signatures, and subtypes.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},367571,1790198594,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":25,"category_name":129,"show_sort_weight":97,"slug":130},19,"General","general",{"code":4,"msg":82,"data":132},{"doc_id":79,"user_id":133,"nickname":42,"user_avatar":134,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":102,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":145,"preview_page_count":33,"preview_urls":146},962085564381,"https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8","Artificial intelligence predicts multiclass molecular signatures and subtypes directly from breast cancer histology: a multicenter retrospective study  \nXiangyang Zhanga , b , Yang Chenc , Changjing Caia , Yifeng Wangd , Jun Tane , Zijie Fangc , Le Weia ,f, Zhuchen Shaoc , Liwen Wangf, Tiezheng Qif, Yihan Liua ,f , Zhaohui Jianga , Yin Lia , Ying Hana , Tibera Kagemulo Rugambwaa , Shan Zenga , Haoqian Wangc , Hong Shena ,g,*, Yongbing Zhangh,*  \nAbstract  \nDetection of biomarkers of breast cancer incurs additional costs and tissue burden. We propose a deep learning-based algorithm (BBMIL) to predict classical biomarkers, immunotherapy-associated gene signatures, and prognosis-associated subtypes directly from hematoxylin and eosin stained histopathology images. BBMIL showed the best performance among comparative algorithmson the prediction of classical biomarkers, immunotherapy related gene signatures, and subtypes.  \nKeywords: breast cancer, deep learning, digital pathology  \nBreast cancer is one of the most prevalent malignancies worldwide, significantly impacting global life expectancy [1] . Treatment regimens include surgery, chemotherapy, radiotherapy, targeted therapy, and immunotherapy, tailored to patient subtypes due to the cancer’s heterogeneity[2,3] . Numerous subtype systems and biomarkers exist, requiring immunohistochemistry, fluorescence in-situ hybridization, and RNA sequencing for evaluation, which is challenging due to high costs and tissue limitations. To address this, we propose a deep  \naDepartment of Oncology, Xiangya Hospital, Central South University, Changsha, Hunan, China, bDepartment ofBiomedical Sciences, City University of Hong Kong, Hong Kong, China, cTsinghua Shenzhen International Graduate School,Tsinghua University, Shenzhen, China, dSchool of Science, Harbin Institute of Technology (Shenzhen), Shenzhen, China, eDepartment of Neurosurgery, Xiangya Hospital, Central South University, Changsha, China, fXiangya School of Medicine, Central South University, Changsha, China, gNational Clinical Research Center for Geriatric Disorders, Xiangya Hospital, Central South University, Changsha, China and hComputer Science and Technology, Harbin Institute of Technology (Shenzhen), Shenzhen, China  \nXiangyang Zhang, Yang Chen, Changjing Cai, and Yifeng Wang contributed equally to this work.  \nSponsorships or competing interests that may be relevant to content are disclosed at the end of this article.  \n*Corresponding authors. Address: Department of Oncology, Xiangya Hospital, Central  \nSouth University, Changsha, Hunan, [China. E-mail: hongshen2000@csu.edu.cn](China. E-mail: hongshen2000@csu.edu.cn)  \n(H. Shen) and Computer Science and Technology, Harbin Institute of  \nTechnology (Shenzhen), Shenzhen, [China. E-mail: ybzhang08@hit.edu.cn](China. E-mail: ybzhang08@hit.edu.cn)  \n(Y.B. Zhang).  \nCopyright © 2025 The Author(s). Published by Wolters Kluwer Health, Inc. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.  \nInternational Journal of Surgery (2025) 111:3109–3114  \nReceived 23 July 2024; Accepted 21 November 2024  \nSupplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal's website [www.lww.com/international-journal-of-surgery](www.lww.com/international-journal-of-surgery).  \nPublished online 7 January 2025  \n[http://dx.doi.org/10.1097/JS9.0000000000002220](http://dx.doi.org/10.1097/JS9.0000000000002220)  \nlearning-based multiple instance learning (MIL) algorithm to predict these biomarkers.  \nDeep learning, a subset of machine learning, allows neural networks to learn from large datasets[4] . MIL models learn from sets of instances with set-level labels. W","cbCaih2PBTK0GSWP","https://ap.wps.com/l/cbCaih2PBTK0GSWP","pdf",2258101,"English","# Abstract\n## Detection problem and proposed BBMIL framework\n## Predictions for biomarkers, gene signatures, and subtypes\n## Comparative performance and validation datasets","[{\"question\":\"What problem does the study target in breast cancer biomarker detection?\",\"answer\":\"Biomarker detection requires additional cost and increases tissue burden, limiting widespread evaluation. The study aims to predict biomarkers directly from standard histopathology images.\"},{\"question\":\"What does BBMIL predict from hematoxylin-and-eosin stained images?\",\"answer\":\"BBMIL predicts classical biomarkers, immunotherapy-associated gene signatures, and prognosis-associated subtypes. The predicted outputs include established receptors and PAM50 plus immunotherapy-related signatures and novel subtypes.\"},{\"question\":\"How does BBMIL perform compared with other algorithms?\",\"answer\":\"BBMIL shows the best overall performance among the compared methods for predicting classical biomarkers, immunotherapy-related gene signatures, and subtypes.\"}]","Artificial intelligence predicts multiclass molecular signatures and subtypes directly from breast cancer histology - a multicenter retrospective study | PDF",1790171420,15,[147,151,155,159],{"page_no":22,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/adc967ce0fccc532dc2a62cf63e4c3c0.webp",983,1310,{"page_no":26,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/2c518f5ab695ba5a348f7589c9bc81fd.webp",982,1309,{"page_no":30,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/fd5485461f3d28e3f74e29de728b80cc.webp",981,1308,{"page_no":33,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/32cdf32abe053cf79d11d00aaea6e30f.webp",980,1306]