[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-363288-105":59,"doc-detail-363288-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","fine-tuned-lightweight-language-models-for-structured-extraction-of-liver-cancer-imaging-free-text-report-a-comparative-analysis-with-existing-large-language-models","Fine-tuned lightweight language models for structured extraction of liver cancer imaging free-text report - a comparative analysis with existing large language models","","Organizing free-text patient records into structured data remains labor-intensive and slow in liver cancer imaging workflows. This study evaluates a fine-tuned lightweight language model for structuring liver cancer imaging reports, using 2,780 retrospective reports from 2012–2022. Three radiologists annotated key fields (NMT, DLT, VI) and a senior oncologist calibrated reliability. The AT5-based Liver-T5 model (250M) is fine-tuned and assessed by EM and Macro-F1, then compared with larger LLMs. Results show higher EM and strong accuracy for NMT and VI, with improved schema compliance and better efficiency than larger models.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/fine-tuned-lightweight-language-models-for-structured-extraction-of-liver-cancer-imaging-free-text-report-a-comparative-analysis-with-existing-large-language-models/363288/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/fine-tuned-lightweight-language-models-for-structured-extraction-of-liver-cancer-imaging-free-text-report-a-comparative-analysis-with-existing-large-language-models/363288.png","ImageObject",300,407,{"name":92,"@type":93},"Stanley","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-25","2026-09-23",true,{"@type":102,"interactionType":103,"userInteractionCount":44},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What is the main goal of the fine-tuned lightweight language model in this study?","Question",{"text":112,"@type":113},"To evaluate how effectively a fine-tuned lightweight language model can transform liver cancer imaging free-text reports into structured data by extracting key fields.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How were the evaluation targets and dataset created?",{"text":117,"@type":113},"A retrospective dataset of 2,780 liver imaging reports was collected, and three radiologists annotated Number of Malignant Tumors (NMT), Diameter of the Largest Tumor (DLT), and Vascular Invasion (VI), followed by calibration by a senior oncologist.",{"name":119,"@type":110,"acceptedAnswer":120},"How did Liver-T5 perform compared with larger LLMs?",{"text":121,"@type":113},"Liver-T5 outperformed larger LLMs in Exact Match (EM) rate and key metrics, achieving EM of 0.8907 and high extraction accuracy for NMT and VI, while generating more schema-compliant outputs.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},363288,1790805939,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":44,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":19,"preview_urls":146},2336477405376,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","Luo etal. BMC Medical Imaging (2026) 26:225 [https://doi.org/10.1186/s12880-026-02351-x](https://doi.org/10.1186/s12880-026-02351-x)  \nBMC Medical Imaging  \nRESEARCH Open Access  \nFine-tuned lightweight language models   \nfor structured extraction of liver cancer imaging free-text report: a comparative analysis with existing large language models  \nYang-Di Luo 1†, Yong Sun2†, Shi-Le Tang1†, Lu-Jun Shen 1, Xuxiazi Zou3, Hong-Liang Zou4, Hua-Lin Shen 1, Tong-Tong Yang1, Jia-Yun Wu1, Rong-Yu Lu 1, Chao-Feng Li5, Jin-Hua Huang1, Jing Cai6, Yi-Quan Jiang1*† and Ge Ren6*†  \nAbstract  \nBackground Organizing free-text patient data into a structured format is labor-intensive and time-consuming.  \nThis study aims to evaluate the effectiveness of a fine-tuned lightweight language model in structuring liver cancer imaging reports.  \nMethods A retrospective dataset of 2,780 liver imaging reports from SunYat-sen University Cancer Center (2012– 2022), including cases of primary liver cancer and benign liver disease, was collected. Three key entries—Number of Malignant Tumors (NMT), Diameter of the Largest Tumor (DLT), and Vascular Invasion (VI)—were annotated by three radiologists and subsequently reviewed and calibrated by a senior oncologist to ensure data reliability. The annotated dataset was randomly split into training, validation, and test sets at a ratio of 7:1:2 . AT5-based lightweight model with 250 M parameters (Liver-T5) was fine-tuned using these data. Performance was evaluated using Accuracy and Macro-F1 metrics. Comparative analysis with LLMs such as ChatGLM4, Qianwen2 . 0, and Llama3 .1 was conducted. Results The fine-tuned Liver-T5 model outperformed larger LLMs in Exact Match (EM) rate and key evaluation metrics, achieving an EM of 0.8907 and high accuracy for NMT (0 . 9355) and VI (0 . 9910) . Specifically, for NMT extraction, Liver-T5 achieved an accuracy of 0 . 9355, outperforming large models such as Qianwen72B (accuracy 0 . 9140), LLaMA3 (accuracy 0 . 8961), and ChatGLM4 (accuracy 0 . 8226) . In the VI extraction, Liver-T5 achieved the highest accuracy of 0 . 9910, significantly surpassing other models, with Qianwen72B, LLaMA3, and ChatGLM4 achieving accuracies of 0 . 9606, 0 . 9462, and 0 . 7581, respectively. A higher proportion of schema-nonconforming outputs was observed  \n\n| †Yang-Di Luo, Yong Sun and Shi-Le Tang contributed equally to this work. |\n| --- |\n| †Yi-Quan Jiang and Ge Ren are the co-senior authors. |\n\n*Correspondence:  \nYi-Quan Jiang [jiangyq@sysucc.org.cn](jiangyq@sysucc.org.cn)[ ](jiangyq@sysucc.org.cn)Ge Ren [gary-ge. ren@polyu.edu.hk](gary-ge. ren@polyu.edu.hk)  \nFull list of author information is available at the end of the article  \n© The Author(s) 2026. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit [http://creati](http://creati)[vecommons.org/licenses/by-nc-nd/4.0/](vecommons.org/licenses/by-nc-nd/4.0/.)[.](vecommons.org/licenses/by-nc-nd/4.0/.)  \nLuo et al. BMC Medical Imaging (2026) 26:225 Page 2 of 11  \nin large general-purpose models (e. g., LLaMA3), while Liver-T5 more consistently ","cbCailWFnH20g697","https://ap.wps.com/l/cbCailWFnH20g697","pdf",3040506,11,"English","# Abstract\n# Background\n# Methods\n## Dataset and annotation\n## Model and evaluation metrics\n# Results\n# Conclusions","[{\"question\":\"What is the main goal of the fine-tuned lightweight language model in this study?\",\"answer\":\"To evaluate how effectively a fine-tuned lightweight language model can transform liver cancer imaging free-text reports into structured data by extracting key fields.\"},{\"question\":\"How were the evaluation targets and dataset created?\",\"answer\":\"A retrospective dataset of 2,780 liver imaging reports was collected, and three radiologists annotated Number of Malignant Tumors (NMT), Diameter of the Largest Tumor (DLT), and Vascular Invasion (VI), followed by calibration by a senior oncologist.\"},{\"question\":\"How did Liver-T5 perform compared with larger LLMs?\",\"answer\":\"Liver-T5 outperformed larger LLMs in Exact Match (EM) rate and key metrics, achieving EM of 0.8907 and high extraction accuracy for NMT and VI, while generating more schema-compliant outputs.\"}]","Fine-tuned lightweight language models for structured extraction of liver cancer imaging free-text report - a comparative analysis with existing large language models | PDF",1790150637,28,[147,151,155,159],{"page_no":8,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/911aa36519bf721365e0da6e8d76536f.webp",983,1310,{"page_no":14,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/677feabc2fc2a029b5e688b70edd169c.webp",982,1309,{"page_no":81,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/e6ebfe25169fe3263ea9465ae444cb23.webp",981,1308,{"page_no":19,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/dbf157045334421cf6f2df1d4b4050e0.webp",980,1306]