{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T12:46:30Z","timestamp":1782996390511,"version":"3.54.5"},"reference-count":26,"publisher":"BMJ","issue":"1","license":[{"start":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T00:00:00Z","timestamp":1782950400000},"content-version":"unspecified","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"name":"Faculty of Medicine Ramathibodi Hospital, Mahidol University","award":["N42A690510"],"award-info":[{"award-number":["N42A690510"]}]}],"content-domain":{"domain":["bmj.com"],"crossmark-restriction":true},"short-container-title":["BMJ Health Care Inform"],"accepted":{"date-parts":[[2026,6,21]]},"published-print":{"date-parts":[[2026,7]]},"abstract":"<jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>This study developed and validated monolingual and bilingual sentence-bidirectional encoder representations from transformers (SBERT) models for detecting cancer recurrence within Thai-English electronic medical records (EMRs) from Thai cancer hospitals.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Method<\/jats:title>\n                    <jats:p>A multicentre dataset of 32\u2009436 documents from 1250 patients was used for model development. External validation involved an independent dataset of 9244 documents from 384 patients across two Thai cancer hospitals. Performance was benchmarked against a fine-tuned PubMedBERT (MetBERT).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>The development dataset included breast (43.9%), colorectal (12.1%), cervical (28.0%) and head and neck (16.0%) cancers. MetBERT achieved the highest area under the precision-recall curve (AUPRC) for locoregional versus no recurrence (11.1%) and locoregional versus distant recurrence (91.7%), while monolingual-SBERT excelled at distant versus no recurrence (32.0%). External validation demonstrated MetBERT superiority for locoregional versus no recurrence (9.30%\u201321.50%). For distant versus no recurrence, bilingual-SBERT performed best with AUPRC 17.55%\u201324.39%. While MetBERT led in distinguishing locoregional versus distant recurrence (88.30%\u201394.70%), bilingual-SBERT demonstrated robust external validation performance (AUPRC 85.25%\u201391.80%).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>Low AUPRC values (9%\u201332%) reflect the extreme class imbalance in real-world data (~1% recurrence prevalence). Despite this, fine-tuned MetBERT achieved highest performance, while bilingual-SBERT demonstrated superior robustness during external validation. This validates sentence embedding models for handling mixed Thai-English medical records in multilingual clinical environments.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Sentence embedding frameworks provide a practical, generalisable solution for detecting cancer recurrence within multilingual EMRs. Despite text-length constraints, these models are suitable for clinical integration as a screening tool for cancer registry workflows.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1136\/bmjhci-2025-101997","type":"journal-article","created":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T12:10:20Z","timestamp":1782994220000},"page":"e101997","update-policy":"https:\/\/doi.org\/10.1136\/crossmarkpolicy","source":"Crossref","is-referenced-by-count":0,"title":["Detection of cancer recurrence from Thai-English electronic medical records using sentence embeddings"],"prefix":"10.1136","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1009-7382","authenticated-orcid":false,"given":"Ekapob","family":"Sangariyavanich","sequence":"first","affiliation":[{"name":"National Cancer Institute of Thailand, Bangkok, Thailand"},{"name":"Department of Clinical Epidemiology and Biostatistics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wanchana","family":"Ponthongmak","sequence":"additional","affiliation":[{"name":"Department of Clinical Epidemiology and Biostatistics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nawanan","family":"Theera-Ampornpunt","sequence":"additional","affiliation":[{"name":"Department of Clinical Epidemiology and Biostatistics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nat","family":"Tangchitnob","sequence":"additional","affiliation":[{"name":"Department of Clinical Epidemiology and Biostatistics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gareth J","family":"McKay","sequence":"additional","affiliation":[{"name":"Centre for Public Health, Queen\u2019s University Belfast, Belfast, Northern Ireland, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ammarin","family":"Thakkinstian","sequence":"additional","affiliation":[{"name":"Department of Clinical Epidemiology and Biostatistics, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"239","published-online":{"date-parts":[[2026,7,2]]},"reference":[{"key":"2026070205101413000_33.1.e101997.1","doi-asserted-by":"crossref","first-page":"420","DOI":"10.1001\/jamaoncol.2021.6987","article-title":"Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability-Adjusted Life Years for 29 Cancer Groups From 2010 to 2019: A Systematic Analysis for the Global Burden of Disease Study 2019","volume":"8","author":"Kocarnik","year":"2022","journal-title":"JAMA Oncol"},{"key":"2026070205101413000_33.1.e101997.2","doi-asserted-by":"publisher","DOI":"10.1038\/s41571-019-0167-7"},{"key":"2026070205101413000_33.1.e101997.3","doi-asserted-by":"crossref","first-page":"657","DOI":"10.1002\/jso.23883","article-title":"The quest for population-level cancer recurrence data; current deficiencies and targets for improvement","volume":"111","author":"In","year":"2015","journal-title":"J Surg Oncol"},{"key":"2026070205101413000_33.1.e101997.4","doi-asserted-by":"crossref","DOI":"10.1186\/s12859-018-2466-x","article-title":"Using natural language processing and machine learning to identify breast cancer local recurrence","volume":"19","author":"Zeng","year":"2018","journal-title":"BMC Bioinformatics"},{"key":"2026070205101413000_33.1.e101997.5","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1200\/CCI.19.00034","article-title":"Natural Language Processing Approaches to Detect the Timeline of Metastatic Recurrence of Breast Cancer","volume":"3","author":"Banerjee","year":"2019","journal-title":"JCO Clin Cancer Inform"},{"key":"2026070205101413000_33.1.e101997.6","doi-asserted-by":"crossref","first-page":"283","DOI":"10.1007\/s41666-019-00046-3","article-title":"Identifying Breast Cancer Distant Recurrences from Electronic Health Records Using Machine Learning","volume":"3","author":"Zeng","year":"2019","journal-title":"J Healthc Inform Res"},{"key":"2026070205101413000_33.1.e101997.7","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1148\/radiol.2021210043","article-title":"Patterns of Metastatic Disease in Patients with Cancer Derived from Natural Language Processing of Structured CT Radiology Reports over a 10-year Period","volume":"301","author":"Do","year":"2021","journal-title":"Radiology"},{"key":"2026070205101413000_33.1.e101997.8","doi-asserted-by":"crossref","first-page":"469","DOI":"10.1200\/CCI.20.00165","article-title":"Development and Use of Natural Language Processing for Identification of Distant Cancer Recurrence and Sites of Distant Recurrence Using Unstructured Electronic Health Record Data","volume":"5","author":"Karimi","year":"2021","journal-title":"JCO Clinical Cancer Informatics"},{"key":"2026070205101413000_33.1.e101997.9","doi-asserted-by":"crossref","unstructured":"Kaka HM G , Subendran S , Decker K , et al . Pretrained Neural Networks Accurately Identify Cancer Recurrence in Medical Record. 2022;93\u20137. doi:10.3233\/SHTI220403","DOI":"10.3233\/SHTI220403"},{"key":"2026070205101413000_33.1.e101997.10","unstructured":"Liu KK O , Witteveen-Lane M , Chen B , et al . MetBERT: a generalizable and pre-trained deep learning model for the prediction of metastatic cancer from clinical notes. AMIA Annual Symposium proceedings AMIA Symposium; 2022:331\u20138."},{"key":"2026070205101413000_33.1.e101997.11","doi-asserted-by":"publisher","DOI":"10.1126\/science.aaa8685"},{"key":"2026070205101413000_33.1.e101997.12","doi-asserted-by":"crossref","first-page":"101326","DOI":"10.1016\/j.imu.2023.101326","article-title":"Systematic review of natural language processing for recurrent cancer detection from electronic medical records","volume":"41","author":"Sangariyavanich","year":"2023","journal-title":"Informatics in Medicine Unlocked"},{"key":"2026070205101413000_33.1.e101997.13","unstructured":"Devlin J , Chang M-W , Lee K , et al . Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv 2018;181004805."},{"key":"2026070205101413000_33.1.e101997.14","doi-asserted-by":"publisher","DOI":"10.1145\/3458754"},{"key":"2026070205101413000_33.1.e101997.15","doi-asserted-by":"crossref","DOI":"10.1038\/sdata.2016.35","article-title":"MIMIC-III, a freely accessible critical care database","volume":"3","author":"Johnson","year":"2016","journal-title":"Sci Data"},{"key":"2026070205101413000_33.1.e101997.16","doi-asserted-by":"crossref","unstructured":"Reimers N , Gurevych I . Sentence-bert: sentence embeddings using siamese bert-networks. arXiv 2019;190810084. doi:10.18653\/v1\/D19-1410","DOI":"10.18653\/v1\/D19-1410"},{"key":"2026070205101413000_33.1.e101997.17","doi-asserted-by":"crossref","unstructured":"Conneau A , Kiela D , Schwenk H , et al . Supervised learning of universal sentence representations from natural language inference data. arXiv 2017. doi:10.18653\/v1\/D17-1070","DOI":"10.18653\/v1\/D17-1070"},{"key":"2026070205101413000_33.1.e101997.18","doi-asserted-by":"crossref","unstructured":"Cer D , Yang Y , Kong S , et al . Universal sentence encoder. arXiv 2018;180311175. doi:10.18653\/v1\/D18-2029","DOI":"10.18653\/v1\/D18-2029"},{"key":"2026070205101413000_33.1.e101997.19","doi-asserted-by":"crossref","DOI":"10.2196\/37770","article-title":"Identifying the Perceived Severity of Patient-Generated Telemedical Queries Regarding COVID: Developing and Evaluating a Transfer Learning-Based Solution","volume":"10","author":"Gatto","year":"2022","journal-title":"JMIR Med Inform"},{"key":"2026070205101413000_33.1.e101997.20","doi-asserted-by":"crossref","first-page":"340","DOI":"10.1016\/j.jad.2024.07.102","article-title":"Computational analysis of linguistic features in speech samples of first-episode bipolar disorder and psychosis","volume":"363","author":"Arslan","year":"2024","journal-title":"J Affect Disord"},{"key":"2026070205101413000_33.1.e101997.21","doi-asserted-by":"crossref","unstructured":"Akiba T , Sano S , Yanase T , et al . Optuna: a next-generation hyperparameter optimization framework. Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining; 2019:2623\u201331.","DOI":"10.1145\/3292500.3330701"},{"key":"2026070205101413000_33.1.e101997.22","doi-asserted-by":"crossref","unstructured":"Feng F , Yang Y , Cer D , et al . Language-agnostic bert sentence embedding. Proceedings of the 60th annual meeting of the association for computational linguistics (volume 1: Long papers); 2022 Available: https:\/\/aclanthology.org\/2022.acl-long","DOI":"10.18653\/v1\/2022.acl-long.62"},{"key":"2026070205101413000_33.1.e101997.23","unstructured":"Lundberg S . A unified approach to interpreting model predictions. arXiv 2017;170507874."},{"key":"2026070205101413000_33.1.e101997.24","doi-asserted-by":"crossref","DOI":"10.1200\/CCI.24.00107","article-title":"Automated Identification of Breast Cancer Relapse in Computed Tomography Reports Using Natural Language Processing","volume":"8","author":"Lee","year":"2024","journal-title":"JCO Clin Cancer Inform"},{"key":"2026070205101413000_33.1.e101997.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-02855-5"},{"key":"2026070205101413000_33.1.e101997.26","unstructured":"G\u00fcnther M , Mohr I , Williams DJ , et al . Late chunking: contextual chunk embeddings using long-context embedding models. arXiv 2024;240904701."}],"container-title":["BMJ Health &amp; Care Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1136\/bmjhci-2025-101997","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T12:10:41Z","timestamp":1782994241000},"score":1,"resource":{"primary":{"URL":"https:\/\/informatics.bmj.com\/lookup\/doi\/10.1136\/bmjhci-2025-101997"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7]]},"references-count":26,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,7,2]]},"published-print":{"date-parts":[[2026,7]]}},"alternative-id":["10.1136\/bmjhci-2025-101997"],"URL":"https:\/\/doi.org\/10.1136\/bmjhci-2025-101997","relation":{},"ISSN":["2632-1009"],"issn-type":[{"value":"2632-1009","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7]]}}}