{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T15:50:25Z","timestamp":1753890625735,"version":"3.41.2"},"reference-count":46,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,5,9]],"date-time":"2025-05-09T00:00:00Z","timestamp":1746748800000},"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. Digit. Health"],"abstract":"<jats:p>Data stored within electronic health records (EHRs) offer a valuable source of information for real-world evidence (RWE) studies in oncology. However, many key clinical features are only available within unstructured notes. We present ArcTEX, a novel data enrichment pipeline developed to extract oncological features from NHS unstructured clinical notes with high accuracy, even in resource-constrained environments where availability of GPUs might be limited. By design, the predicted outcomes of ArcTEX are free of patient-identifiable information, making this pipeline ideally suited for use in Trust environments. We compare our pipeline to existing discriminative and generative models, demonstrating its superiority over approaches such as Llama3\/3.1\/3.2 and other BERT based models, with a mean accuracy of 98.67% for several essential clinical features in endometrial and breast cancer. Additionally, we show that as few as 50 annotated training examples are needed to adapt the model to a different oncology area, such as lung cancer, with a different set of priority clinical features, achieving a comparable mean accuracy of 95% on average.<\/jats:p>","DOI":"10.3389\/fdgth.2025.1561358","type":"journal-article","created":{"date-parts":[[2025,5,9]],"date-time":"2025-05-09T05:29:55Z","timestamp":1746768595000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["ArcTEX\u2014a novel clinical data enrichment pipeline to support real-world evidence oncology studies"],"prefix":"10.3389","volume":"7","author":[{"given":"Keiran","family":"Tait","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Joseph","family":"Cronin","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Olivia","family":"Wiper","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jamie","family":"Wallis","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jim","family":"Davies","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Robert","family":"D\u00fcrichen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,5,9]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"376","DOI":"10.1016\/j.annonc.2021.12.015","article-title":"External control arms in oncology: current use and future directions","volume":"33","author":"Mishra-Kalyani","year":"2022","journal-title":"Ann Oncol"},{"key":"B2","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1136\/ijgc-2022-004039","article-title":"Refining adjuvant treatment in endometrial cancer based on molecular features: the RAINBO clinical trial program","volume":"33","year":"2022","journal-title":"Int J Gynecol Cancer"},{"key":"B3","doi-asserted-by":"publisher","first-page":"100511","DOI":"10.1016\/j.cosrev.2022.100511","article-title":"Neural natural language processing for unstructured data in electronic health records: a review","volume":"46","author":"Li","year":"2022","journal-title":"Comput Sci Rev"},{"key":"B4","doi-asserted-by":"publisher","first-page":"106649","DOI":"10.1016\/j.compbiomed.2023.106649","article-title":"Natural language processing in electronic health records in relation to healthcare decision-making: a systematic review","volume":"155","author":"Hossain","year":"2023","journal-title":"Comput Biol Med"},{"key":"B5","doi-asserted-by":"publisher","first-page":"105122","DOI":"10.1016\/j.ijmedinf.2023.105122","article-title":"Clinical named entity recognition and relation extraction using natural language processing of medical free text: a systematic review","volume":"177","author":"Fraile Navarro","year":"2023","journal-title":"Int J Med Inf"},{"article-title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","year":"2019","author":"Devlin","key":"B6"},{"key":"B7","doi-asserted-by":"publisher","first-page":"1234","DOI":"10.1093\/bioinformatics\/btz682","article-title":"BioBERT: a pre-trained biomedical language representation model for biomedical text mining","volume":"36","author":"Lee","year":"2020","journal-title":"Bioinformatics"},{"key":"B8","doi-asserted-by":"publisher","first-page":"57","DOI":"10.3390\/informatics11030057","article-title":"Large language models in healthcare and medical domain: a review","volume":"11","author":"Nazi","year":"2024","journal-title":"Informatics"},{"article-title":"Named clinical entity recognition benchmark","year":"","author":"Abdul","key":"B9"},{"article-title":"Large language models in biomedical natural language processing: benchmarks, baselines, and recommendations","year":"2024","author":"Chen","key":"B10"},{"article-title":"Is larger always better? Evaluating and prompting large language models for non-generative medical tasks","year":"2024","author":"Zhu","key":"B11"},{"key":"B12","doi-asserted-by":"publisher","first-page":"btad557","DOI":"10.1093\/bioinformatics\/btad557","article-title":"An extensive benchmark study on biomedical text generation and mining with ChatGPT","volume":"39","author":"Chen","year":"2023","journal-title":"Bioinformatics"},{"key":"B13","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/D16-1264","article-title":"SQuAD: 100,000+ Questions for Machine Comprehension of Text","author":"Rajpurkar","year":"2016"},{"article-title":"RoBERTa: A Robustly Optimized BERT Pretraining Approach","year":"2019","author":"Liu","key":"B14"},{"key":"B15","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3458754","article-title":"Domain-Specific language model pretraining for biomedical natural language processing","volume":"3","author":"Gu","year":"2022","journal-title":"ACM Trans Comput Healthcare"},{"article-title":"Efficient Few-Shot Learning Without Prompts","year":"2022","author":"Tunstall","key":"B16"},{"article-title":"QLoRA: Efficient Finetuning of Quantized LLMs","year":"2023","author":"Dettmers","key":"B17"},{"article-title":"LoRA: Low-Rank Adaptation of Large Language Models","year":"2021","author":"Hu","key":"B18"},{"article-title":"Awesome Llama Prompts","year":"","author":"LangGPT","key":"B19"},{"year":"","key":"B20","article-title":"Llama 3 | Model Cards and Prompt formats [WWW Document]"},{"year":"","key":"B21","article-title":"Meta Llama 2 | Model Cards and Prompt formats [WWW Document]"},{"article-title":"Scaling Laws for Neural Language Models","year":"2020","author":"Kaplan","key":"B22"},{"key":"B23","doi-asserted-by":"crossref","first-page":"302","DOI":"10.1007\/978-3-031-70242-6_29","article-title":"Assessing generative language models in classification tasks: performance and self-evaluation capabilities in the environmental and climate change domain","volume-title":"Natural Language Processing and Information Systems","author":"Grasso","year":"2024"},{"article-title":"ScholarBERT: Bigger is Not Always Better","year":"2022","author":"Hong","key":"B24"},{"key":"B25","doi-asserted-by":"publisher","first-page":"181","DOI":"10.1186\/s12874-022-01665-y","article-title":"A comparative study on deep learning models for text classification of unstructured medical notes with various levels of class imbalance","volume":"22","author":"Lu","year":"2022","journal-title":"BMC Med Res Methodol"},{"article-title":"Entropic Distribution Matching in Supervised Fine-tuning of LLMs: Less Overfitting and Better Diversity","year":"2024","author":"Li","key":"B26"},{"key":"B27","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2024.emnlp-main.444","article-title":"Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations?","author":"Gekhman","year":"2024"},{"key":"B28","first-page":"1","article-title":"From words to watts: benchmarking the energy costs of large language model inference","author":"Samsi","year":"2023"},{"key":"B29","doi-asserted-by":"publisher","first-page":"128096","DOI":"10.1016\/j.neucom.2024.128096.","article-title":"A review of green artificial intelligence: towards a more sustainable future","volume":"599","author":"Bol\u00f3n-Canedo","year":"2024","journal-title":"Neurocomputing"},{"key":"B30","doi-asserted-by":"publisher","first-page":"26310","DOI":"10.1038\/s41598-024-76682-6","article-title":"Reconciling the contrasting narratives on the environmental impact of large language models","volume":"14","author":"Ren","year":"2024","journal-title":"Sci Rep"},{"key":"B31","doi-asserted-by":"publisher","first-page":"2364","DOI":"10.3390\/app14062364","article-title":"Optimizing large language models on multi-core CPUs: a case study of the BERT model","volume":"14","author":"Zhao","year":"2024","journal-title":"Appl Sci"},{"year":"2024","key":"B32","article-title":"The rise of small language models. ObjectBox"},{"key":"B33","doi-asserted-by":"publisher","first-page":"e58158","DOI":"10.2196\/58158","article-title":"Evaluating and enhancing large language Models\u2019 performance in domain-specific medicine: development and usability study with DocOA","volume":"26","author":"Chen","year":"2024","journal-title":"J Med Internet Res"},{"key":"B34","doi-asserted-by":"publisher","first-page":"108189","DOI":"10.1016\/j.compbiomed.2024.108189","article-title":"A comprehensive evaluation of large language models on benchmark biomedical text processing tasks","volume":"171","author":"Jahan","year":"2024","journal-title":"Comput Biol Med"},{"article-title":"Evaluation of GPT and BERT-based models on identifying protein-protein interactions in biomedical text","year":"2023","author":"Rehana","key":"B35"},{"article-title":"Fine-Tuned \u201cSmall\u201d LLMs (Still) Significantly Outperform Zero-Shot Generative AI Models in Text Classification","year":"2024","author":"Bucher","key":"B36"},{"key":"B37","doi-asserted-by":"publisher","first-page":"200308","DOI":"10.1016\/j.iswa.2023.200308","article-title":"ChatGPT and finetuned BERT: a comparative study for developing intelligent design support systems","volume":"21","author":"Qiu","year":"2024","journal-title":"Intell Syst Appl"},{"key":"B38","first-page":"2339","article-title":"It's not just size that matters: small language models are also few-shot learners","author":"Schick","year":"2021"},{"key":"B39","doi-asserted-by":"publisher","first-page":"107","DOI":"10.1007\/s44163-024-00197-2","article-title":"Large language models to process, analyze, and synthesize biomedical texts: a scoping review","volume":"4","author":"Doneva","year":"2024","journal-title":"Discov Artif Intell"},{"key":"B40","doi-asserted-by":"crossref","DOI":"10.1145\/3648471","article-title":"Utilizing BERT for Information Retrieval: Survey, Applications, Resources, and Challenges","author":"Wang","year":"2024"},{"key":"B41","doi-asserted-by":"publisher","first-page":"3703155","DOI":"10.1145\/3703155","article-title":"A survey on hallucination in large language models: principles, taxonomy, challenges, and open questions","volume":"43","author":"Huang","year":"2024","journal-title":"ACM Trans Inf Syst"},{"key":"B42","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3233\/SW-243719","article-title":"Testing prompt engineering methods for knowledge extraction from text","author":"Polat","year":"2024","journal-title":"Semant Web"},{"key":"B43","doi-asserted-by":"publisher","first-page":"39","DOI":"10.1162\/tacl_a_00632","article-title":"Benchmarking large language models for news summarization","volume":"12","author":"Zhang","year":"2024","journal-title":"Trans Assoc Comput Linguist"},{"key":"B44","doi-asserted-by":"publisher","first-page":"244","DOI":"10.3174\/ajnr.A8102","article-title":"AI-assisted summarization of radiological reports: evaluating GPT3davinci, BARTcnn, LongT5booksum, LEDbooksum, LEDlegal, and LEDclinical","volume":"45","author":"Chien","year":"2024","journal-title":"Am J Neuroradiol"},{"key":"B45","doi-asserted-by":"publisher","first-page":"7","DOI":"10.26502\/jrci.2809097","article-title":"Overview of the potentially transforming role of large language models (LlMs) in creating patient-friendly summaries for radiology reports","volume":"2024","author":"Kalidindi","year":"2024","journal-title":"J Radiol Clin Imaging"},{"key":"B46","first-page":"793","article-title":"Retrieval-augmented generation for knowledge-intensive NLP tasks\u2019","author":"Lewis","year":"2020","journal-title":"NIPS'20: Proceedings of the 34th International Conference on Neural Information Processing Systems"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1561358\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,5,9]],"date-time":"2025-05-09T05:29:56Z","timestamp":1746768596000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2025.1561358\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,9]]},"references-count":46,"alternative-id":["10.3389\/fdgth.2025.1561358"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2025.1561358","relation":{},"ISSN":["2673-253X"],"issn-type":[{"type":"electronic","value":"2673-253X"}],"subject":[],"published":{"date-parts":[[2025,5,9]]},"article-number":"1561358"}}