[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-461606-105":59,"doc-detail-461606-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","clinically-aware-learning-ordinal-loss-improves-medical-image-classifiers","Clinically Aware Learning - Ordinal Loss Improves Medical Image Classifiers","","BI-RADS (Breast Imaging Reporting and Data System) mammogram classification underpins early breast cancer detection, yet most models treat its ordinal class levels as nominal labels by optimizing cross-entropy. This study evaluates whether ordinal-aware loss functions better match clinical severity by accounting for the ordered structure and dataset imbalance. Using fixed architectures and a unified training pipeline across multiple datasets, ordinal losses such as Earth Mover Distance (EMD) outperform cross-entropy in AUROC and macro-F1, improving robustness, particularly by reducing severe misclassifications.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/clinically-aware-learning-ordinal-loss-improves-medical-image-classifiers/461606/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/clinically-aware-learning-ordinal-loss-improves-medical-image-classifiers/461606.png","ImageObject",300,407,{"name":92,"@type":93},"nayy☆","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-07","2026-09-30",true,{"@type":102,"interactionType":103,"userInteractionCount":81},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"Why is treating BI-RADS as a nominal task problematic for medical image classifiers?","Question",{"text":112,"@type":113},"BI-RADS categories are ordered by malignancy suspicion, but cross-entropy penalizes all misclassifications equally. This fails to reflect that clinically meaningful errors can differ in severity depending on how far the prediction deviates from the true class.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"What method does the study use to test ordinal-aware learning for BI-RADS?",{"text":117,"@type":113},"The work compares ordinal loss functions against standard cross-entropy under controlled, architecture-fixed conditions. It uses a unified training pipeline across multiple datasets and evaluates effects of dataset and label balancing using AUROC and macro-F1 averaged over five seeds.",{"name":119,"@type":110,"acceptedAnswer":120},"What were the main results of using ordinal loss functions such as EMD?",{"text":121,"@type":113},"Balanced sampling across datasets during training produced statistically significant improvements. Ordinal losses, including Earth Mover Distance (EMD), consistently achieved higher performance across multiple metrics and were especially effective at reducing severe misclassifications.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},461606,1790919778,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":81,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145},962090893153,"https://ap-avatar.wpscdn.com/davatar_9964176cb1d06d4a9deccf72a44ae3dc","Article  \nClinically Aware Learning: Ordinal Loss Improves Medical Image Classifiers  \nArsenii Litvinov 1, Egor Ushakov 1, Sofia Senotrusova 1, Kirill Lukianov 1, Yury Markin 1, Liudmila Mikhailova 2,3 and Evgeny Karpulevich 1, *  \nAcademic Editor: Andrea Ciarmiello  \nReceived: 16 November 2025  \nRevised: 6 December 2025  \nAccepted: 11 December 2025  \nPublished: 3 January 2026  \nCopyright: © 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.  \n1 Trusted AI Research Center, RAS, 109004 Moscow, Russia; [filashkov@gmail.com](filashkov@gmail.com) (A.L.)  \n2 Economic Faculty, Lomonosov Moscow State University, 119991 Moscow, Russia  \n3 Higher School of Management, Financial University Under the Government of the Russian Federation, 125167 Moscow, Russia  \n* Correspondence: [karpulevich@mail.ru](karpulevich@mail.ru)  \nAbstract  \nBackground: BI-RADS (Breast Imaging Reporting and Data System) mammogram classification is central to early breast cancer detection. Despite being an ordinal scale that reflects increasing levels of malignancy suspicion, most models treat BI-RADS as a nominal task using cross-entropy loss, thereby disregarding the inherent class order. This mismatch between the clinical severity of misclassification and the model’s optimization objective remains underexplored. Methods: We systematically evaluate whether incorporating ordinal-aware loss functions improves BI-RADS classification performance under controlled, architecture-fixed conditions and dataset imbalance. Using a unified training pipeline across multiple datasets, we compare ordinal losses to standard cross-entropy, analyzing the effect of dataset-and label-level balancing. Area under the receiver operating characteristic curve (AUROC) and macro-F1 scores are reported as averages over five seeds. Results: Balanced sampling across datasets during training led to statistically significant improvements. Ordinal loss functions, such as Earth Mover Distance (EMD), consistently achieved higher performance across multiple metrics compared to conventional cross-entropy approaches commonly reported in the literature. Improvements were particularly evident in reducing severe misclassifications, demonstrating that aligning the learning objective with the ordinal structure of BI-RADS enhances robustness and clinical relevance. Conclusions: Aligning the learning objective with the ordinal BI-RADS structure substantially improves classification accuracy without changing the underlying architecture. These findings emphasize the importance of loss design, regularization, and data-balancing strategies in medical AI, supporting more reliable breast cancer screening.  \nKeywords: breast imaging risk classification; ordinal classification; loss functions; deep learning; breast cancer screening  \n1. Introduction  \nBreast cancer is the second leading cause of mortality among women worldwide, despite advances in its early diagnosis and treatment [1] . Combating breast cancer remains a highly important challenge of modern medicine [2] . Timely detection of malignant changes significantly increases the chances of successful therapy [3,4] . In this regard, mammographic screening based on the interpretation of breast X-ray images plays a key role in clinical practice [5] . Breast Imaging Reporting and Data System (BI-RADS), developed by  \nthe American College of Radiology (ACR), is used worldwide to standardize the description of findings and unify clinical decision-making [6] . BI-RADS classifies detected changes by degree of suspicion for malignancy: from category 1 (no cancer) to category 6 (proven malignancy confirmed by biopsy) [7] . It is important to emphasize that this scale is ordinal and meaning-ordered, and classification errors of different magnitudes have different clinical significance [8] . At the same time, many researchers analy","cbCaibomAr4isb2n","https://ap.wps.com/l/cbCaibomAr4isb2n","pdf",1307910,22,"English","# Abstract\n# 1. Introduction","[{\"question\":\"Why is treating BI-RADS as a nominal task problematic for medical image classifiers?\",\"answer\":\"BI-RADS categories are ordered by malignancy suspicion, but cross-entropy penalizes all misclassifications equally. This fails to reflect that clinically meaningful errors can differ in severity depending on how far the prediction deviates from the true class.\"},{\"question\":\"What method does the study use to test ordinal-aware learning for BI-RADS?\",\"answer\":\"The work compares ordinal loss functions against standard cross-entropy under controlled, architecture-fixed conditions. It uses a unified training pipeline across multiple datasets and evaluates effects of dataset and label balancing using AUROC and macro-F1 averaged over five seeds.\"},{\"question\":\"What were the main results of using ordinal loss functions such as EMD?\",\"answer\":\"Balanced sampling across datasets during training produced statistically significant improvements. Ordinal losses, including Earth Mover Distance (EMD), consistently achieved higher performance across multiple metrics and were especially effective at reducing severe misclassifications.\"}]","Clinically Aware Learning - Ordinal Loss Improves Medical Image Classifiers | PDF",1790761959,55]