{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T05:08:45Z","timestamp":1788930525480,"version":"build-2803163510"},"reference-count":97,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T00:00:00Z","timestamp":1788912000000},"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. Med."],"abstract":"<jats:p>Deep models for oral squamous cell carcinoma (OSCC) detection from haematoxylin-eosin histopathology degrade sharply across laboratories, scanners and magnifications, yet the target centre is usually unavailable during training, making generalization to unseen domains the decisive requirement for reliable screening. We cast multi-centre OSCC classification as a domain-generalization problem and propose SICA (Style-Invariant Consistency Alignment), a single objective combining three established invariance mechanisms: style randomisation of channel-wise feature statistics, a stain-consistency term forcing two independently stained views of each image to agree in embedding and prediction space, and a class-conditional feature alignment that preserves the decision boundary. Two of them rely only on paired stain views, so SICA remains effective with a single source domain, where classical cross-domain alignment collapses to plain risk minimisation. On a reproducible leave-one-domain-out benchmark spanning a genuine cross-magnification shift and a controlled multi-centre stain-shift simulation, SICA is compared against nine baselines over eight metrics. Among the ten algorithms sharing a common ImageNet backbone it attains the best and most stable out-of-domain AUROC (90.6% on the multi-centre protocol), the best worst-case centre, and a balanced operating point better calibrated than empirical risk minimisation. With confidence intervals, paired tests and multiplicity correction, that margin over the strongest baselines proves consistent but not statistically decisive on five simulated centres. Under the same protocol, a frozen pathology foundation model closes part of the stain gap but not the scale gap, and the proposed objective improves it further (92.3% to 93.4% AUROC on identical frozen features), so better representations and explicit invariance are complementary. An audit of the public corpus uncovered 28 byte-identical images filed under contradictory labels, removed before re-running the benchmark under a group-aware, leakage-controlled protocol. Finally, on a genuinely independent 150-patient cohort we report a negative result: every method, ours included, falls to chance-level discrimination, because the binding constraint is field-of-view scale rather than stain. Stain invariance is therefore necessary but not sufficient, and a simulated stain-shift benchmark must not be read as an estimate of deployment performance.<\/jats:p>","DOI":"10.3389\/fmed.2026.1951350","type":"journal-article","created":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T04:41:20Z","timestamp":1788928880000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Domain-generalizable oral cancer classification: bridging multi-center histopathological datasets"],"prefix":"10.3389","volume":"13","author":[{"given":"Ziying","family":"Hua","sequence":"first","affiliation":[{"name":"School of Stomatology, Harbin Medical University","place":["Harbin, Heilongjiang, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yushi","family":"Zhou","sequence":"additional","affiliation":[{"name":"First Affiliated Hospital of Kunming Medical University","place":["Kunming, Yunnan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"209","DOI":"10.3322\/caac.21660","article-title":"Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries","volume":"71","author":"Sung","year":"2021","journal-title":"CA Cancer J Clin"},{"key":"B2","doi-asserted-by":"publisher","first-page":"13811","DOI":"10.1002\/cam4.6025","article-title":"Global, regional, and national burden of oral cancer and its attributable risk factors from 1990 to 2019","volume":"12","author":"Sun","year":"2023","journal-title":"Cancer Med"},{"key":"B3","doi-asserted-by":"publisher","first-page":"539","DOI":"10.1002\/hed.26006","article-title":"Potentially malignant disorders of the oral cavity and oral dysplasia: a systematic review and meta-analysis of malignant transformation rate by subtype","volume":"42","author":"Iocca","year":"2020","journal-title":"Head Neck"},{"key":"B4","doi-asserted-by":"publisher","first-page":"948","DOI":"10.1111\/jop.13056","article-title":"Intra-observer and inter-observer variability in two grading systems for oral epithelial dysplasia: a multi-centre study in India","volume":"49","author":"Ranganathan","year":"2020","journal-title":"J Oral Pathol Med"},{"key":"B5","doi-asserted-by":"publisher","first-page":"838","DOI":"10.1111\/odi.15078","article-title":"Understanding interobserver variability of pathologists to improve oral epithelial dysplasia grading","volume":"31","author":"Ng","year":"2025","journal-title":"Oral Dis"},{"key":"B6","doi-asserted-by":"publisher","first-page":"418","DOI":"10.1111\/jop.13365","article-title":"Oral epithelial dysplasia grading: comparing the binary system to the traditional 3-tier system, an actuarial study with malignant transformation as outcome","volume":"52","author":"Sperandio","year":"2023","journal-title":"J Oral Pathol Med"},{"key":"B7","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1016\/j.neunet.2020.05.003","article-title":"Automated classification of cells into multiple classes in epithelial tissue of oral squamous cell carcinoma using transfer learning and convolutional neural network","volume":"128","author":"Das","year":"2020","journal-title":"Neural Netw"},{"key":"B8","doi-asserted-by":"publisher","first-page":"1899","DOI":"10.3390\/diagnostics12081899","article-title":"Early diagnosis of oral squamous cell carcinoma based on histopathological images using deep and hybrid learning approaches","volume":"12","author":"Fati","year":"2022","journal-title":"Diagnostics"},{"key":"B9","doi-asserted-by":"publisher","first-page":"2131","DOI":"10.3390\/ijerph20032131","article-title":"Automatic detection of oral squamous cell carcinoma from histopathological images of oral mucosa using deep convolutional neural network","volume":"20","author":"Das","year":"2023","journal-title":"Int J Environ Res Public Health"},{"key":"B10","doi-asserted-by":"publisher","first-page":"1321","DOI":"10.1177\/00220345221089858","article-title":"Histopathology-based diagnosis of oral squamous cell carcinoma using deep learning","volume":"101","author":"Yang","year":"2022","journal-title":"J Dent Res"},{"key":"B11","doi-asserted-by":"publisher","first-page":"3833","DOI":"10.3390\/s22103833","article-title":"Histopathologic oral cancer prediction using oral squamous cell carcinoma biopsy empowered with transfer learning","volume":"22","author":"Rahman","year":"2022","journal-title":"Sensors"},{"key":"B12","doi-asserted-by":"publisher","first-page":"6364102","DOI":"10.1155\/2022\/6364102","article-title":"Diagnosis of oral squamous cell carcinoma using deep neural networks and binary particle swarm optimization on histopathological images: an AIoMT approach","volume":"2022","author":"Deif","year":"2022","journal-title":"Comput Intell Neurosci"},{"key":"B13","doi-asserted-by":"publisher","first-page":"12598","DOI":"10.1038\/s41598-020-69250-1","article-title":"Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data","volume":"10","author":"Sheller","year":"2020","journal-title":"Sci Rep"},{"key":"B14","doi-asserted-by":"publisher","first-page":"1259","DOI":"10.1093\/jamia\/ocaa341","article-title":"Federated learning improves site performance in multicenter deep learning without data sharing","volume":"28","author":"Sarma","year":"2021","journal-title":"J Am Med Inform Assoc"},{"key":"B15","doi-asserted-by":"publisher","first-page":"325","DOI":"10.1109\/JBHI.2020.3032060","article-title":"Measuring domain shift for deep learning in histopathology","volume":"25","author":"Stacke","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"B16","doi-asserted-by":"publisher","first-page":"101813","DOI":"10.1016\/j.media.2020.101813","article-title":"Deep neural network models for computational histopathology: a survey","volume":"67","author":"Srinidhi","year":"2021","journal-title":"Med Image Anal"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.1109\/TBME.2021.3117407","article-title":"Domain adaptation for medical image analysis: a survey","volume":"69","author":"Guan","year":"2022","journal-title":"IEEE Trans Biomed Eng"},{"key":"B18","doi-asserted-by":"publisher","first-page":"822","DOI":"10.1007\/s11263-023-01913-8","article-title":"MixStyle neural networks for domain generalization and adaptation","volume":"132","author":"Zhou","year":"2024","journal-title":"Int J Comput Vis"},{"key":"B19","doi-asserted-by":"publisher","first-page":"746307","DOI":"10.3389\/fmed.2021.746307","article-title":"StainNet: a fast and robust stain normalization network","volume":"8","author":"Kang","year":"2021","journal-title":"Front Med"},{"key":"B20","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1109\/JBHI.2020.2983206","article-title":"Stain standardization capsule for application-driven histopathological image normalization","volume":"25","author":"Zheng","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"B21","author":"Shen","year":""},{"key":"B22","doi-asserted-by":"publisher","first-page":"100183","DOI":"10.1016\/j.jpi.2022.100183","article-title":"Data-driven color augmentation for H&E stained images in computational pathology","volume":"14","author":"Marini","year":"2023","journal-title":"J Pathol Inform"},{"key":"B23","doi-asserted-by":"publisher","first-page":"108136","DOI":"10.1016\/j.engappai.2024.108136","article-title":"StainSwin: vision transformer-based stain normalization for histopathology image analysis","volume":"133","author":"Kablan","year":"2024","journal-title":"Eng Appl Artif Intell"},{"key":"B24","doi-asserted-by":"publisher","first-page":"277","DOI":"10.1016\/j.neucom.2021.07.005","article-title":"Towards histopathological stain invariance by unsupervised domain augmentation using generative adversarial networks","volume":"460","author":"Vasiljevi\u0107","year":"2021","journal-title":"Neurocomputing"},{"key":"B25","doi-asserted-by":"publisher","first-page":"4396","DOI":"10.1109\/TPAMI.2022.3195549","article-title":"Domain generalization: a survey","volume":"45","author":"Zhou","year":"2023","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"B26","doi-asserted-by":"publisher","first-page":"8052","DOI":"10.1109\/TKDE.2022.3178128","article-title":"Generalizing to unseen domains: a survey on domain generalization","volume":"35","author":"Wang","year":"2023","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"B27","doi-asserted-by":"publisher","first-page":"105114","DOI":"10.1016\/j.dib.2020.105114","article-title":"Histopathological imaging database for oral cancer analysis","volume":"29","author":"Rahman","year":"2020","journal-title":"Data Brief"},{"key":"B28","author":"Caron","year":""},{"key":"B29","doi-asserted-by":"publisher","first-page":"100133","DOI":"10.1016\/j.jpi.2022.100133","article-title":"A self-supervised contrastive learning approach for whole slide image representation in digital pathology","volume":"13","author":"Fashi","year":"2022","journal-title":"J Pathol Inform"},{"key":"B30","doi-asserted-by":"publisher","first-page":"100198","DOI":"10.1016\/j.mlwa.2021.100198","article-title":"Self supervised contrastive learning for digital histopathology","volume":"7","author":"Ciga","year":"2022","journal-title":"Mach Learn Appl"},{"key":"B31","doi-asserted-by":"publisher","first-page":"102256","DOI":"10.1016\/j.media.2021.102256","article-title":"Self-supervised driven consistency training for annotation efficient histopathology image analysis","volume":"75","author":"Srinidhi","year":"2022","journal-title":"Med Image Anal"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1145\/3724391","article-title":"Domain generalization in computational pathology: survey and guidelines","volume":"57","author":"Jahanifar","year":"2025","journal-title":"ACM Comput Surv"},{"key":"B33","doi-asserted-by":"publisher","first-page":"151","DOI":"10.1038\/s41698-024-00652-4","article-title":"Learning generalizable AI models for multi-center histopathology image classification","volume":"8","author":"Asadi-Aghbolaghi","year":"2024","journal-title":"npj Precis Oncol"},{"key":"B34","doi-asserted-by":"publisher","first-page":"e0262838","DOI":"10.1371\/journal.pone.0262838","article-title":"Deep learning model calibration for improving performance in class-imbalanced medical image classification tasks","volume":"17","author":"Rajaraman","year":"2022","journal-title":"PLoS One"},{"key":"B35","doi-asserted-by":"publisher","first-page":"132677","DOI":"10.1109\/ACCESS.2020.3010180","article-title":"Automated detection and classification of oral lesions using deep learning for early detection of oral cancer","volume":"8","author":"Welikala","year":"2020","journal-title":"IEEE Access"},{"key":"B36","doi-asserted-by":"publisher","first-page":"699","DOI":"10.1016\/j.ijom.2021.09.001","article-title":"Performance of deep convolutional neural network for classification and detection of oral potentially malignant disorders in photographic images","volume":"51","author":"Warin","year":"2022","journal-title":"Int J Oral Maxillofac Surg"},{"key":"B37","doi-asserted-by":"publisher","first-page":"212","DOI":"10.1186\/s12903-024-03993-5","article-title":"Deep learning in oral cancer \u2013 a systematic review","volume":"24","author":"Warin","year":"2024","journal-title":"BMC Oral Health"},{"key":"B38","doi-asserted-by":"publisher","first-page":"551","DOI":"10.1111\/jop.13578","article-title":"Deep learning-based image classification and segmentation on digital histopathology for oral squamous cell carcinoma: a systematic review and meta-analysis","volume":"53","author":"Pirayesh","year":"2024","journal-title":"J Oral Pathol Med"},{"key":"B39","doi-asserted-by":"publisher","first-page":"2368","DOI":"10.1111\/odi.15330","article-title":"Diagnosis of oral cancer with deep learning. A comparative test accuracy systematic review","volume":"31","author":"Nieri","year":"2025","journal-title":"Oral Dis"},{"key":"B40","doi-asserted-by":"publisher","first-page":"1","DOI":"10.59275\/j.melba.2022-f9a1","article-title":"Learning representations with contrastive self-supervised learning for histopathology applications","volume":"1","author":"Stacke","year":"2022","journal-title":"Mach Learn Biomed Imaging"},{"key":"B41","first-page":"291","article-title":"Quantification of histochemical staining by color deconvolution","volume":"23","author":"Ruifrok","year":"2001","journal-title":"Anal Quant Cytol Histol"},{"key":"B42","author":"Macenko","year":""},{"key":"B43","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1109\/38.946629","article-title":"Color transfer between images","volume":"21","author":"Reinhard","year":"2001","journal-title":"IEEE Comput Graph Appl"},{"key":"B44","doi-asserted-by":"publisher","first-page":"1962","DOI":"10.1109\/TMI.2016.2529665","article-title":"Structure-preserving color normalization and sparse stain separation for histological images","volume":"35","author":"Vahadane","year":"2016","journal-title":"IEEE Trans Med Imaging"},{"key":"B45","doi-asserted-by":"publisher","first-page":"2126","DOI":"10.1109\/TMI.2018.2820199","article-title":"Whole-slide mitosis detection in H&E breast histology using PHH3 as a reference to train distilled stain-invariant convolutional networks","volume":"37","author":"Tellez","year":"2018","journal-title":"IEEE Trans Med Imaging"},{"key":"B46","doi-asserted-by":"publisher","first-page":"101544","DOI":"10.1016\/j.media.2019.101544","article-title":"Quantifying the effects of data augmentation and stain color normalization in convolutional neural networks for computational pathology","volume":"58","author":"Tellez","year":"2019","journal-title":"Med Image Anal"},{"key":"B47","author":"Chang","year":""},{"key":"B48","author":"Sun","year":""},{"key":"B49","author":"Long","year":""},{"key":"B50","first-page":"1","article-title":"Domain-adversarial training of neural networks","volume":"17","author":"Ganin","year":"2016","journal-title":"J Mach Learn Res"},{"key":"B51","author":"Arjovsky","year":""},{"key":"B52","author":"Sagawa","year":""},{"key":"B53","author":"Zhang","year":""},{"key":"B54","author":"Zhou","year":""},{"key":"B55","author":"Zhou","year":""},{"key":"B56","doi-asserted-by":"publisher","first-page":"8008","DOI":"10.1109\/TIP.2021.3112012","article-title":"Domain adaptive ensemble learning","volume":"30","author":"Zhou","year":"2021","journal-title":"IEEE Trans Image Process"},{"key":"B57","author":"Huang","year":""},{"key":"B58","author":"Ulyanov","year":""},{"key":"B59","doi-asserted-by":"publisher","first-page":"102559","DOI":"10.1016\/j.media.2022.102559","article-title":"Transformer-based unsupervised contrastive learning for histopathological image classification","volume":"81","author":"Wang","year":"2022","journal-title":"Med Image Anal"},{"key":"B60","author":"Grill","year":""},{"key":"B61","author":"Chen","year":""},{"key":"B62","author":"Hinton","year":""},{"key":"B63","author":"Sajjadi","year":""},{"key":"B64","author":"Laine","year":""},{"key":"B65","author":"Tarvainen","year":""},{"key":"B66","author":"Xie","year":""},{"key":"B67","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":"B68","doi-asserted-by":"publisher","first-page":"2924","DOI":"10.1038\/s41591-024-03141-0","article-title":"A foundation model for clinical-grade computational pathology and rare cancers detection","volume":"30","author":"Vorontsov","year":"2024","journal-title":"Nat Med"},{"key":"B69","author":"Zimmermann","year":""},{"key":"B70","author":"Filiot","year":""},{"key":"B71","doi-asserted-by":"publisher","first-page":"970","DOI":"10.1038\/s41586-024-07894-z","article-title":"A pathology foundation model for cancer diagnosis and prognosis prediction","volume":"634","author":"Wang","year":"2024","journal-title":"Nature"},{"key":"B72","doi-asserted-by":"publisher","first-page":"863","DOI":"10.1038\/s41591-024-02856-4","article-title":"A visual-language foundation model for computational pathology","volume":"30","author":"Lu","year":"2024","journal-title":"Nat Med"},{"key":"B73","doi-asserted-by":"publisher","first-page":"3640","DOI":"10.1038\/s41467-025-58796-1","article-title":"A clinical benchmark of public self-supervised pathology foundation models","volume":"16","author":"Campanella","year":"2025","journal-title":"Nat Commun"},{"key":"B74","author":"Xu","year":""},{"key":"B75","author":"Zhang","year":""},{"key":"B76","author":"Liang","year":""},{"key":"B77","author":"Sun","year":""},{"key":"B78","author":"Zellinger","year":""},{"key":"B79","author":"Tzeng","year":""},{"key":"B80","author":"Motiian","year":""},{"key":"B81","author":"Li","year":""},{"key":"B82","author":"Long","year":""},{"key":"B83","author":"Kang","year":""},{"key":"B84","author":"Gulrajani","year":""},{"key":"B85","author":"Bussola","year":""},{"key":"B86","doi-asserted-by":"publisher","first-page":"100804","DOI":"10.1016\/j.patter.2023.100804","article-title":"Leakage and the reproducibility crisis in machine-learning-based science","volume":"4","author":"Kapoor","year":"2023","journal-title":"Patterns"},{"key":"B87","doi-asserted-by":"publisher","first-page":"48","DOI":"10.1038\/s41746-022-00592-y","article-title":"Machine learning for medical imaging: methodological failures and recommendations for the future","volume":"5","author":"Varoquaux","year":"2022","journal-title":"npj Digit Med"},{"key":"B88","author":"Naeini","year":""},{"key":"B89","author":"Guo","year":""},{"key":"B90","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1175\/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2","article-title":"Verification of forecasts expressed in terms of probability","volume":"78","author":"Brier","year":"1950","journal-title":"Mon Weather Rev"},{"key":"B91","doi-asserted-by":"publisher","first-page":"230","DOI":"10.1186\/s12916-019-1466-7","article-title":"Calibration: the achilles heel of predictive analytics","volume":"17","author":"Van Calster","year":"2019","journal-title":"BMC Med"},{"key":"B92","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1214\/aos\/1176344552","article-title":"Bootstrap methods: another look at the jackknife","volume":"7","author":"Efron","year":"1979","journal-title":"Ann Stat"},{"key":"B93","doi-asserted-by":"publisher","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach","volume":"44","author":"DeLong","year":"1988","journal-title":"Biometrics"},{"key":"B94","first-page":"65","article-title":"A simple sequentially rejective multiple test procedure","volume":"6","author":"Holm","year":"1979","journal-title":"Scand J Stat"},{"key":"B95","first-page":"1","article-title":"Statistical comparisons of classifiers over multiple data sets","volume":"7","author":"Dem\u0161ar","year":"2006","journal-title":"J Mach Learn Res"},{"key":"B96","doi-asserted-by":"publisher","first-page":"1050","DOI":"10.1038\/s41597-024-03836-6","article-title":"High-resolution AI image dataset for diagnosing oral submucous fibrosis and squamous cell carcinoma","volume":"11","author":"Chaudhary","year":"2024","journal-title":"Sci Data"},{"key":"B97","author":"Koh","year":""}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1951350\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T04:41:27Z","timestamp":1788928887000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1951350\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,9]]},"references-count":97,"alternative-id":["10.3389\/fmed.2026.1951350"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1951350","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,9]]},"article-number":"1951350"}}