{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T06:03:42Z","timestamp":1790661822305,"version":"4.1.0"},"reference-count":21,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Public Health"],"abstract":"<jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>The HER2 immunohistochemical (IHC) score of breast cancer is a critical precursor for determining eligibility for anti-HER2 targeted therapy, as the final HER2 status (positive\/negative\/low) is determined by IHC results combined with FISH confirmation for equivocal 2\u202f+\u202fcases according to ASCO\/CAP guidelines. However, most existing artificial intelligence methods separate the detection of invasive cancer regions from HER2 grading and fail to adequately model the spatial heterogeneity of whole-slide tumors. This study proposes an end-to-end intelligent HER2 grading framework that integrates automated detection of invasive cancer regions with hierarchical graph reasoning, enabling automated and precise evaluation of HER2 IHC in breast cancer.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Method<\/jats:title>\n                    <jats:p>A coarse-to-fine automated HER2 grading framework was developed. First, invasive tumor regions were automatically identified through supervised multi-scale tissue segmentation. Within the detected regions, patch-level HER2 grading integrates Delaunay cell image features with self-supervised visual representations, while WSI-level spatial representations model tumor heterogeneity to achieve fully automated HER2 grading across whole-slide images. The framework was trained using data from two scanning platforms and externally validated on 82 independent WSI samples from a third platform.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In independent external validation, the Dice coefficient for invasive region detection reached 0.89 (on 50 WSI images); the accuracy rate of HER2 grading at the patch level was 96.49% (95% CI, 96.08\u201396.87%) (7,924 patches), while that at the WSI level reached 98.0% (95% CI, 91.5\u201399.9%) (82 WSI images). The quadratic-weighted Cohen\u2019s \u03ba was 0.97 (95% CI, 0.94\u20131.00), and all misclassifications at the WSI level were confined to the distinction between HER2 1\u202f+\u202fand 2\u202f+.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This framework enables fully automated HER2 IHC scoring and interpretation under multicenter and cross-scanner conditions, providing a computational tool for standardized IHC assessment, while final HER2 status determination for 2\u202f+\u202fcases still requires FISH validation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fpubh.2026.1940487","type":"journal-article","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:46:39Z","timestamp":1790660799000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Automatic grading of HER2 immunohistochemical whole-slide images based on a multi-scale hierarchical graph neural network"],"prefix":"10.3389","volume":"14","author":[{"given":"Yingjie","family":"Jiang","sequence":"first","affiliation":[{"name":"Department of Pathology, Changhai Hospital, Naval Medical University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenhao","family":"Sun","sequence":"additional","affiliation":[{"name":"School of Health Science and Engineering, University of Shanghai for Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chao","family":"Ye","sequence":"additional","affiliation":[{"name":"The 902nd Hospital of the Joint Logistic Support Force","place":["Bengbu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chun","family":"Xu","sequence":"additional","affiliation":[{"name":"CellSee(Wuxi) Intelligent Technology Co., Ltd","place":["Wuxi, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiewen","family":"Xue","sequence":"additional","affiliation":[{"name":"School of Basic Medical Sciences, Naval Medical University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yiliang","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Health Science and Engineering, University of Shanghai for Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shiju","family":"Yan","sequence":"additional","affiliation":[{"name":"School of Health Science and Engineering, University of Shanghai for Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yiming","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Pathology, Changhai Hospital, Naval Medical University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chenguang","family":"Bai","sequence":"additional","affiliation":[{"name":"Department of Pathology, Changhai Hospital, Naval Medical University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,29]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"2415","DOI":"10.1016\/S0140-6736(16)32417-5","article-title":"HER2-positive breast cancer","volume":"389","author":"Loibl","year":"2017","journal-title":"Lancet"},{"key":"ref2","volume-title":"Breast Tumours","year":"2019"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"2110","DOI":"10.1056\/nejmoa2407086","article-title":"Trastuzumab deruxtecan after endocrine therapy in metastatic breast cancer","volume":"391","author":"Bardia","year":"2024","journal-title":"N Engl J Med"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"9","DOI":"10.1056\/NEJMoa2203690","article-title":"Trastuzumab deruxtecan in previously treated HER2-low advanced breast cancer","volume":"387","author":"Modi","year":"2022","journal-title":"N Engl J Med"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"3867","DOI":"10.1200\/jco.22.02864","article-title":"Human epidermal growth factor receptor 2 testing in breast cancer: ASCO\u2013College of American pathologists guideline update","volume":"41","author":"Wolff","year":"2023","journal-title":"J Clin Oncol"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"100009","DOI":"10.1016\/j.modpat.2022.100009","article-title":"Interobserver Variation in the Assessment of Immunohistochemistry Expression Levels in HER2-Negative Breast Cancer: Can We Improve the Identification of Low Levels of HER2 Expression by Adjusting the Criteria? An International Interobserver Study","volume":"36","author":"Baez-Navarro","year":"2023","journal-title":"Mod Pathol"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"100944","DOI":"10.1016\/j.modpat.2025.100944","article-title":"Agreement across 10 artificial intelligence models in assessing human epidermal growth factor receptor 2 (HER2) expression in breast cancer whole-slide images","volume":"39","author":"McKelvey","year":"2026","journal-title":"Mod Pathol"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"263","DOI":"10.3390\/diagnostics13020263","article-title":"Immunohistochemical HER2 recognition and analysis of breast cancer based on deep learning","volume":"13","author":"Che","year":"2023","journal-title":"Diagnostics (Basel)."},{"key":"ref9","doi-asserted-by":"publisher","first-page":"1204","DOI":"10.1186\/s12885-024-12980-6","article-title":"Precision HER2: a comprehensive AI system for accurate and consistent evaluation of HER2 expression in invasive breast cancer","volume":"24","author":"Xiong","year":"2024","journal-title":"BMC Cancer"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"2131","DOI":"10.3390\/diagnostics15172131","article-title":"Transformer-based HER2 scoring in breast Cancer: comparative performance of a foundation and a lightweight model","volume":"15","author":"Wang","year":"2025","journal-title":"Diagnostics"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"102486","DOI":"10.1016\/j.media.2022.102486","article-title":"SlideGraph+: whole slide image level graphs to predict HER2 status in breast cancer","volume":"80","author":"Lu","year":"2022","journal-title":"Med Image Anal"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"0048","DOI":"10.34133\/bmef.0048","article-title":"Automated HER2 scoring in breast Cancer images using deep learning and pyramid sampling","volume":"5","author":"Selcuk","year":"2024","journal-title":"BME Front"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"1868102","DOI":"10.3389\/fphys.2026.1868102","article-title":"Deep learning based automated HER2 score prediction using immunohistochemistry histopathological images: a dual-center study","volume":"17","author":"Ma","year":"2026","journal-title":"Front Physiol"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"h5527","DOI":"10.1136\/bmj.h5527","article-title":"STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies","volume":"351","author":"Bossuyt","year":"2015","journal-title":"BMJ"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"159","DOI":"10.2307\/2529310","article-title":"The measurement of observer agreement for categorical data","volume":"33","author":"Landis","year":"1977","journal-title":"Biometrics"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.2304.07193","article-title":"DINOv2: learning robust visual features without supervision","author":"Oquab","year":"2023"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"850","DOI":"10.1038\/s41591-024-02857-3","article-title":"Towards a general-purpose foundation model for computational pathology","volume":"30","author":"Chen","year":"2024","journal-title":"Nat Med"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2408.00714","article-title":"SAM 2: segment anything in images and videos","author":"Ravi","year":"2024"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"101563","DOI":"10.1016\/j.media.2019.101563","article-title":"HoVer-net: simultaneous segmentation and classification of nuclei in multi-tissue histology images","volume":"58","author":"Graham","year":"2019","journal-title":"Med Image Anal"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"1634","DOI":"10.1038\/s41592-022-01663-4","article-title":"Cellpose 2.0: how to train your own model","volume":"19","author":"Pachitariu","year":"2022","journal-title":"Nat Methods"},{"key":"ref21","author":"Brody","year":"2021"}],"container-title":["Frontiers in Public Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpubh.2026.1940487\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:46:41Z","timestamp":1790660801000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpubh.2026.1940487\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,29]]},"references-count":21,"alternative-id":["10.3389\/fpubh.2026.1940487"],"URL":"https:\/\/doi.org\/10.3389\/fpubh.2026.1940487","relation":{},"ISSN":["2296-2565"],"issn-type":[{"value":"2296-2565","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,29]]},"article-number":"1940487"}}