{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T10:54:17Z","timestamp":1793357657521,"version":"4.5.0"},"reference-count":41,"publisher":"MDPI AG","issue":"18","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\/"}],"funder":[{"name":"Agencia Nacional de Investigaci\u00f3n y Desarrollo of Chile"},{"id":[{"id":"https:\/\/ror.org\/05btm5233","id-type":"ROR","asserted-by":"publisher"}]},{"name":"Centro de Prevenci\u00f3n y Control de C\u00e1ncer","award":["FONDAP 152220002"],"award-info":[{"award-number":["FONDAP 152220002"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Cancers"],"abstract":"<jats:p>Background: Class imbalance remains a major limitation in artificial intelligence based oral cancer diagnosis, particularly in small clinical image datasets where malignant lesions are underrepresented. Methods: Using a dataset of 3000 mobile-acquired oral cavity images categorized into four diagnostic classes (healthy, benign, oral potentially malignant disorders, and oral cancer), we systematically evaluated the impact of imbalance mitigation strategies on diagnostic performance. Three backbone architectures, EfficientNet, Vision Transformer, and Swin Transformer, were trained under four conditions: raw imbalanced training, random under-sampling, random over-sampling, and medical-safe data augmentation. Performance was evaluated using stratified five-fold cross-validation, and all reported metrics correspond to the mean performance across the validation folds. Results: In multiclass evaluation, augmentation yielded the strongest overall performance, with EfficientNet achieving the highest macro-F1 score (0.669 \u00b1 0.024) and area under curve (0.881 \u00b1 0.015). Under clinically oriented binary malignant-risk evaluation (high-risk vs. low-risk lesions), sensitivity reached 0.817 \u00b1 0.030 with EfficientNet, while data augmentation maintained competitive malignant-risk detection across the evaluated architectures. In contrast, random undersampling consistently yielded lower descriptive performance across architectures. Conclusions: These findings indicate that augmentation-based imbalance mitigation is associated with improved multiclass and oral cancer specific performance while maintaining competitive overall discrimination. The results provide practical guidance for developing clinically robust artificial intelligence systems for early oral cancer detection in imbalanced real-world datasets.<\/jats:p>","DOI":"10.3390\/cancers18182926","type":"journal-article","created":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T16:43:16Z","timestamp":1788972196000},"page":"2926","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Improving Diagnostic Sensitivity in Imbalanced Oral Cancer Image Classification: A Comparative Study of CNN and Transformer Architectures"],"prefix":"10.3390","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5591-3518","authenticated-orcid":false,"given":"Pablo","family":"Orme\u00f1o-Arriagada","sequence":"first","affiliation":[{"name":"Instituto de Tecnolog\u00eda para la Innovaci\u00f3n en Salud y Bienestar, Facultad de Ingenieria, Universidad Andres Bello, Vi\u00f1a del Mar 2530958, Chile"},{"name":"CECAN, Centro para la Prevenci\u00f3n y Control del Cancer, Santiago 8330077, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"conceptualization","vocabulary":"credit"},{"role":"writing-original-draft","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-4461-748X","authenticated-orcid":false,"given":"Valentina","family":"Z\u00fa\u00f1iga","sequence":"additional","affiliation":[{"name":"Instituto de Tecnolog\u00eda para la Innovaci\u00f3n en Salud y Bienestar, Facultad de Ingenieria, Universidad Andres Bello, Vi\u00f1a del Mar 2530958, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"conceptualization","vocabulary":"credit"},{"role":"writing-original-draft","vocabulary":"credit"},{"role":"data-curation","vocabulary":"credit"},{"role":"formal-analysis","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0304-4343","authenticated-orcid":false,"given":"Carlos","family":"Toro","sequence":"additional","affiliation":[{"name":"Faculty of Engineering, Universidad Andres Bello, Autopista Concepci\u00f3n-Talcahuano CCP-THNO 7100, Talcahuano 4260000, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0167-5969","authenticated-orcid":false,"given":"Gast\u00f3n","family":"M\u00e1rquez","sequence":"additional","affiliation":[{"name":"CECAN, Centro para la Prevenci\u00f3n y Control del Cancer, Santiago 8330077, Chile"},{"name":"Departamento de Ciencias de la Computaci\u00f3n y Tecnolog\u00eda de la Informaci\u00f3n, Universidad del Biobio, Chill\u00e1n 4081112, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-6254-4339","authenticated-orcid":false,"given":"David","family":"Araya","sequence":"additional","affiliation":[{"name":"Instituto de Tecnolog\u00eda para la Innovaci\u00f3n en Salud y Bienestar, Facultad de Ingenieria, Universidad Andres Bello, Vi\u00f1a del Mar 2530958, Chile"},{"name":"CECAN, Centro para la Prevenci\u00f3n y Control del Cancer, Santiago 8330077, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8078-253X","authenticated-orcid":false,"given":"Diego","family":"Mellado","sequence":"additional","affiliation":[{"name":"Instituto de Tecnolog\u00eda para la Innovaci\u00f3n en Salud y Bienestar, Facultad de Ingenieria, Universidad Andres Bello, Vi\u00f1a del Mar 2530958, Chile"},{"name":"CECAN, Centro para la Prevenci\u00f3n y Control del Cancer, Santiago 8330077, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8318-4201","authenticated-orcid":false,"given":"Carla","family":"Taramasco","sequence":"additional","affiliation":[{"name":"Instituto de Tecnolog\u00eda para la Innovaci\u00f3n en Salud y Bienestar, Facultad de Ingenieria, Universidad Andres Bello, Vi\u00f1a del Mar 2530958, Chile"},{"name":"CECAN, Centro para la Prevenci\u00f3n y Control del Cancer, Santiago 8330077, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"resources","vocabulary":"credit"},{"role":"funding-acquisition","vocabulary":"credit"},{"role":"project-administration","vocabulary":"credit"}]}],"member":"1968","published-online":{"date-parts":[[2026,9,9]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"100678","DOI":"10.1016\/j.oor.2024.100678","article-title":"Oral cancer: Recent breakthroughs in pathology and therapeutic approaches","volume":"12","author":"Ramachandran","year":"2024","journal-title":"Oral Oncol. Rep."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"6098","DOI":"10.1039\/D4AY01030B","article-title":"A review: Early detection of oral cancer biomarkers using microfluidic colorimetric point-of-care devices","volume":"16","author":"Balapure","year":"2024","journal-title":"Anal. Methods"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"e086588","DOI":"10.1136\/bmjopen-2024-086588","article-title":"Can oral cancer screening reduce late-stage diagnosis, treatment delay and mortality? A population-based study in Taiwan","volume":"14","author":"Tsai","year":"2024","journal-title":"BMJ Open"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Sahoo, R.K., Sahoo, K.C., Dash, G.C., Kumar, G., Baliarsingh, S.K., Panda, B., and Pati, S. (2024). Diagnostic performance of artificial intelligence in detecting oral potentially malignant disorders and oral cancer using medical diagnostic imaging: A systematic review and meta-analysis. Front. Oral Health, 5.","DOI":"10.3389\/froh.2024.1494867"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Gonz\u00e1lez-Ruiz, I., Ramos-Garcia, P., Ruiz-Avila, I., and Gonzalez-Moles, M.A. (2023). Early diagnosis of oral cancer: A complex polyhedral problem with a difficult solution. Cancers, 15.","DOI":"10.3390\/cancers15133270"},{"key":"ref_6","first-page":"101138","article-title":"Automated classification of oral potentially malignant disorders and oral squamous cell carcinoma using a convolutional neural network framework: A cross-sectional study","volume":"47","author":"Pedroso","year":"2025","journal-title":"Lancet Reg. Health\u2013Am."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"100030","DOI":"10.1016\/j.ejrai.2025.100030","article-title":"Evaluation metrics in medical imaging AI: Fundamentals, pitfalls, misapplications, and recommendations","volume":"3","author":"Kocak","year":"2025","journal-title":"Eur. J. Radiol. Artif. Intell."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"273","DOI":"10.1007\/s10462-024-10884-2","article-title":"Handling imbalanced medical datasets: Review of a decade of research","volume":"57","author":"Salmi","year":"2024","journal-title":"Artif. Intell. Rev."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Fan, M.s., Qiu, K.x., Wang, D.y., Wang, H., Zhang, W.w., and Yan, L. (2024). Risk factors associated with false negative rate of sentinel lymph node biopsy in endometrial cancer: A systematic review and meta-analysis. Front. Oncol., 14.","DOI":"10.3389\/fonc.2024.1391267"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Kim, M., and Hwang, K.B. (2022). An empirical evaluation of sampling methods for the classification of imbalanced data. PLoS ONE, 17.","DOI":"10.1371\/journal.pone.0271260"},{"key":"ref_11","first-page":"227","article-title":"The Impactness of SMOTE as Imbalance Class Handling for Myocardial Infarction Complication Classification using Machine Learning Approach with Data Imputation and Hyperparameter","volume":"6","author":"Tajali","year":"2024","journal-title":"Indones. J. Electron. Electromed. Eng. Med. Inform."},{"key":"ref_12","first-page":"1310","article-title":"Addressing class imbalance of health data: A systematic literature review on modified synthetic minority oversampling technique (SMOTE) strategies","volume":"8","author":"Hairani","year":"2024","journal-title":"JOIV Int. J. Inform. Vis."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"27","DOI":"10.1186\/s40537-019-0192-5","article-title":"Survey on deep learning with class imbalance","volume":"6","author":"Johnson","year":"2019","journal-title":"J. Big Data"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"321","DOI":"10.1613\/jair.953","article-title":"SMOTE: Synthetic minority over-sampling technique","volume":"16","author":"Chawla","year":"2002","journal-title":"J. Artif. Intell. Res."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","article-title":"A survey on image data augmentation for deep learning","volume":"6","author":"Shorten","year":"2019","journal-title":"J. Big Data"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"23","DOI":"10.1007\/s12065-024-01012-8","article-title":"Addressing class imbalance in remote sensing using deep learning approaches: A systematic literature review","volume":"18","author":"Sharma","year":"2025","journal-title":"Evol. Intell."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"863","DOI":"10.1613\/jair.1.11192","article-title":"SMOTE for learning from imbalanced data: Progress and challenges, marking the 15-year anniversary","volume":"61","author":"Garcia","year":"2018","journal-title":"J. Artif. Intell. Res."},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Goodmaker, C., Bhandari, R., Tappuni, A., and Pham, T. (2026). AI for Classifying Oral Cancer and Precursor Lesions Using Visible-Light Photography. Preprint.","DOI":"10.21203\/rs.3.rs-8865303\/v1"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"25","DOI":"10.32996\/jmhs.2026.7.5.5","article-title":"Swin Transformer\u2013Driven Cervical Cell Classification with Explainable AI and Web-Based Screening","volume":"7","author":"Shakil","year":"2026","journal-title":"J. Med. Health Stud."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Mirfendereski, P., Li, G.Y., Pearson, A.T., and Kerr, A.R. (2025). Artificial intelligence and the diagnosis of oral cavity cancer and oral potentially malignant disorders from clinical photographs: A narrative review. Front. Oral Health, 6.","DOI":"10.3389\/froh.2025.1569567"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"188","DOI":"10.1007\/s00784-025-06268-7","article-title":"Assessing the performance of an artificial intelligence based chatbot in the differential diagnosis of oral mucosal lesions: Clinical validation study","volume":"29","author":"Grinberg","year":"2025","journal-title":"Clin. Oral Investig."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"105001","DOI":"10.1117\/1.JBO.26.10.105001","article-title":"Classification of imbalanced oral cancer image data from high-risk population","volume":"26","author":"Song","year":"2021","journal-title":"J. Biomed. Opt."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Tasci, E., Zhuge, Y., Camphausen, K., and Krauze, A.V. (2022). Bias and class imbalance in oncologic data\u2014Towards inclusive and transferrable AI in large scale oncology data sets. Cancers, 14.","DOI":"10.3390\/cancers14122897"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"129004","DOI":"10.1016\/j.eswa.2025.129004","article-title":"A hybrid literature review on handling imbalanced medical data: AI models and open issues","volume":"296","author":"Siddavatam","year":"2025","journal-title":"Expert Syst. Appl."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"2401","DOI":"10.1007\/s10278-024-01013-0","article-title":"A novel data augmentation method for radiomics analysis using image perturbations","volume":"37","author":"Maragna","year":"2024","journal-title":"J. Imaging Inform. Med."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Yang, Y., Khorshidi, H.A., and Aickelin, U. (2024). A review on over-sampling techniques in classification of multi-class imbalanced datasets: Insights for medical problems. Front. Digit. Health, 6.","DOI":"10.3389\/fdgth.2024.1430245"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"13686","DOI":"10.1109\/ACCESS.2025.3531662","article-title":"Imbalanced data problem in machine learning: A review","volume":"13","author":"Altalhan","year":"2025","journal-title":"IEEE Access"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"100340","DOI":"10.1016\/j.health.2024.100340","article-title":"A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions","volume":"5","author":"Islam","year":"2024","journal-title":"Healthc. Anal."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"487","DOI":"10.11591\/ijai.v13.i1.pp487-499","article-title":"An efficient convolutional neural network-based classifier for an imbalanced oral squamous carcinoma cell dataset","volume":"13","author":"Mohapatra","year":"2024","journal-title":"IAES Int. J. Artif. Intell."},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Zhang, X., and Liu, Q. (2026, January 18\u201320). Breast cancer radiogenomics analysis via computational perturbation on AI-driven multi-omics guided image synthesis. Proceedings of the 2026 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), Montreal, QC, Canada.","DOI":"10.1109\/CCECE68150.2026.11610465"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Huang, W. (2026). Artificial intelligence and its application in early oral cancer screening: A systematic review. Front. Oncol., 16.","DOI":"10.3389\/fonc.2026.1789708"},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Song, B., Kc, D.R., Yang, R.Y., Li, S., Zhang, C., and Liang, R. (2024). Classification of mobile-based oral cancer images using the vision transformer and the Swin transformer. Cancers, 16.","DOI":"10.3390\/cancers16050987"},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"465","DOI":"10.3233\/THC-248041","article-title":"Intelligent deep learning supports biomedical image detection and classification of oral cancer","volume":"32","author":"Chen","year":"2024","journal-title":"Technol. Health Care"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"364","DOI":"10.1007\/s00784-024-05762-8","article-title":"Advancements in diagnosing oral potentially malignant disorders: Leveraging Vision transformers for multi-class detection","volume":"28","author":"Vinayahalingam","year":"2024","journal-title":"Clin. Oral Investig."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"2253","DOI":"10.1002\/hed.27843","article-title":"Detection of oral cancer and oral potentially malignant disorders using artificial intelligence-based image analysis","volume":"46","author":"Kouketsu","year":"2024","journal-title":"Head Neck"},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Rabinovici-Cohen, S., Fridman, N., Weinbaum, M., Melul, E., Hexter, E., Rosen-Zvi, M., Aizenberg, Y., and Porat Ben Amy, D. (2024). From pixels to diagnosis: Algorithmic analysis of clinical oral photos for early detection of oral squamous cell carcinoma. Cancers, 16.","DOI":"10.3390\/cancers16051019"},{"key":"ref_37","unstructured":"Piyarathne, N.S., Liyanage, S.N., Rasnayaka, R.M.S.G.K., Hettiarachchi, P.V.K.S., Devindi, G.A.I., Francis, F.B.A.H., Dissanayake, D.M.D.R., Ranasinghe, R.A.N.S., Pavithya, M.B.D., and Nawinne, I.B. (2026, August 25). Dataset of Annotated Oral Cavity Images for Oral Cancer Detection. Zenodo. Available online: https:\/\/zenodo.org\/records\/10664056."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"106946","DOI":"10.1016\/j.oraloncology.2024.106946","article-title":"A comprehensive dataset of annotated oral cavity images for diagnosis of oral cancer and oral potentially malignant disorders","volume":"156","author":"Piyarathne","year":"2024","journal-title":"Oral Oncol."},{"key":"ref_39","unstructured":"Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., and Gelly, S. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv."},{"key":"ref_40","doi-asserted-by":"crossref","unstructured":"Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., and Guo, B. (2021, January 10\u201317). Swin transformer: Hierarchical vision transformer using shifted windows. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Montreal, QC, Canada.","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref_41","unstructured":"Tan, M., and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. Proceedings of the International Conference on Machine Learning, PMLR."}],"container-title":["Cancers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2072-6694\/18\/18\/2926\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T08:18:38Z","timestamp":1789028318000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2072-6694\/18\/18\/2926"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,9]]},"references-count":41,"journal-issue":{"issue":"18","published-online":{"date-parts":[[2026,9]]}},"alternative-id":["cancers18182926"],"URL":"https:\/\/doi.org\/10.3390\/cancers18182926","relation":{},"ISSN":["2072-6694"],"issn-type":[{"value":"2072-6694","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,9]]}}}