[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-370686-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-370686-en":131},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","tuning-vision-foundation-models-for-rectal-cancer-segmentation-from-ct-scans","Tuning vision foundation models for rectal cancer segmentation from CT scans","","Rectal cancer segmentation on CT is essential for timely diagnosis, yet existing methods face hurdles due to complex rectal anatomy and limited comprehensive annotated datasets. A new CARE dataset was built from 33,024 CT slice pairs of 398 patients, with pixel-level labels for normal and tumor rectum. A novel prompt-based model, U-SAM, was evaluated using IoU, Dice, and NSD, with benchmarking against 46 clinicians. Results show state-of-the-art accuracy and diagnostic outputs comparable to experts within seconds of inference.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/healthcare/","Healthcare",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/tuning-vision-foundation-models-for-rectal-cancer-segmentation-from-ct-scans/370686/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/tuning-vision-foundation-models-for-rectal-cancer-segmentation-from-ct-scans/370686.png","ImageObject",300,407,{"name":42,"@type":43},"Chumphorn","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-31","2026-09-23",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",12,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"Why is rectal cancer segmentation on CT considered crucial?","Question",{"text":63,"@type":64},"Accurate segmentation supports diagnosis and treatment planning, improves access to anatomical and diagnostic imaging features, and assists downstream tasks such as automated staging and radiotherapy delineation.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"What dataset and annotations were used in the study?",{"text":68,"@type":64},"The study used 33,024 CT slice pairs from 398 rectal cancer patients in a new source center, forming the CARE dataset with pixel-level annotations for both normal and cancerous rectum tissue.",{"name":70,"@type":61,"acceptedAnswer":71},"How did the U-SAM model perform compared with prior state-of-the-art methods?",{"text":72,"@type":64},"U-SAM achieved Dice scores of 71.23% for normal rectum and 76.38% for rectal tumors, with IoU and NSD results that surpassed state-of-the-art approaches, alongside strong observer-study validation.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},370686,1790792514,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,110,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":108,"slug":109},7,40,"healthcare",{"id":111,"doc_module":4,"doc_module_name":25,"category_name":112,"show_sort_weight":113,"slug":114},8,"Research & Report",30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":25,"category_name":129,"show_sort_weight":97,"slug":130},19,"General","general",{"code":4,"msg":82,"data":132},{"doc_id":79,"user_id":133,"nickname":42,"user_avatar":134,"doc_module":4,"category_id":107,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":140,"language":141,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":12,"update_tm":145,"read_time":146},2336475401981,"https://ap-avatar.wpscdn.com/avatar/22000c94efd8d5204d?x-image-process=image/resize,m_fixed,w_180,h_180&k=1786935347598174694","communications medicine Article  \nA Nature Portfolio journal  \n[https://doi.org/10.1038/s43856-025-00953-0](https://doi.org/10.1038/s43856-025-00953-0)  \nTuning vision foundation models for rectal cancer segmentation from CT scans  \n Check for updates  \n\n| Hantao Zhang 1,2, Weidong Guo1,2, Shouhong Wan 1,2 , Bingbing Zou2,3,5, Wanqin Wang2,4,5, Chenyang Qiu2,3,5, Kaige Liu2,3,5, Peiquan Jin1 & Jiancheng Yang 6 |  |  |\n| --- | --- | --- |\n| Abstract |  | Plain Language Summary |\n| Background Rectal cancer segmentation in CT is crucial for timely diagnosis. Despite promising methods, challenges remain duetothe rectum’s complex anatomy andthe lack of a comprehensive annotated dataset.\u003Cbr>Methods A total of 33,024 slice pairs from 398 rectal cancer patients in a new source center are enrolled into our dataset, named CARE Dataset, with pixel-level annotations for both normal and cancerous rectum tissue. We split it into 317cases for training and 81for testing. Additionally, we introduce a segmentation model, U-SAM, which, to the best of our knowledge, is a novel approach designed to handle the complex anatomy of the rectum by incorporating prompt information. Segmentation performance for both normal and cancerous rectum was evaluated using Intersection-over-Union (IoU), Dice Coefficient (Dice), and Normalized Surface Distance (NSD) . With the assistance of 46 clinical practitioners, an observer study is conducted to benchmark the U-SAM with human performance and evaluate its clinical applicability. The original new source 398 CT scansand our code are openly available for research.\u003Cbr>Results Our method achieves Dice of 71 .23% for normal rectum and 76 .38% for rectal tumor, with IoU of 55.32% and 61.78%, and NSD values of 83.63% and 58.59%, respectively, surpassing state-of-the-art methods. The observer study validates that\u003Cbr>U-SAM can produce diagnostic results comparable to those of highly experienced doctors in just 3 seconds of inference time (with about 5 minutes for prompt acquisition) in clinical settings.\u003Cbr>Conclusions The proposed U-SAM offers an efficient and reliable method for segmenting rectal cancer and normal tissue, significantly reducing time in clinical settingsand effectively assisting radiologists. We believe this initial exploration in CT-based rectal cancer segmentation will be instrumental for future diagnosis. |  | Establishing the location of tumors from scans taken of people with cancer is helpful for diagnosis and surgical treatment. In the rectum this is challenging due to the rectum structure and the lack of comparative scans from other people. We combined a large number of rectal images from people with cancer to create an image dataset. We also developed a computational method to assist in the identification of normal tissues and areas of rectal cancer. Our work highlights that computational methods can be used to improve the efficiency and accuracy of rectal cancer diagnosis in clinical practice. |\n| Colorectal cancer ranks as the third most prevalent cancer worldwide and the second leading cause of cancer-related mortality. Notably, rectal cancer accounts for one-third of these cases1–3. Timely detection and treatment can effectively halt the further deterioration ofthe patient’s condition. Clinically, treatment for rectal cancer involves a multidisciplinary approach1,4,5, combining surgery with chemoradiation, tailored to the tumor’s size and location, among other factors, to optimize outcomes. Accurate segmentation of both the tumor and the surrounding normal rectal tissue is of significant | importance6, as it provides valuable imaging anatomical and diagnostic information7,8, such as the tumor’s density differences, the extent of tumor invasion into the various layers of the rectal wall, the structure of the serosal surface infiltration, and inflammatory changes in the surrounding tissue. These imaging features offer clinical reference value for radiologists in diagnostic practice9. Furthermore, segme","cbCaisoVVkczNTsg","https://ap.wps.com/l/cbCaisoVVkczNTsg","pdf",1564318,11,"English","# Background\n## Clinical importance of rectal cancer segmentation\n# Methods\n## CARE dataset construction\n## U-SAM prompt-based segmentation model\n## Evaluation metrics and observer study","[{\"question\":\"Why is rectal cancer segmentation on CT considered crucial?\",\"answer\":\"Accurate segmentation supports diagnosis and treatment planning, improves access to anatomical and diagnostic imaging features, and assists downstream tasks such as automated staging and radiotherapy delineation.\"},{\"question\":\"What dataset and annotations were used in the study?\",\"answer\":\"The study used 33,024 CT slice pairs from 398 rectal cancer patients in a new source center, forming the CARE dataset with pixel-level annotations for both normal and cancerous rectum tissue.\"},{\"question\":\"How did the U-SAM model perform compared with prior state-of-the-art methods?\",\"answer\":\"U-SAM achieved Dice scores of 71.23% for normal rectum and 76.38% for rectal tumors, with IoU and NSD results that surpassed state-of-the-art approaches, alongside strong observer-study validation.\"}]","Tuning vision foundation models for rectal cancer segmentation from CT scans | PDF",1790189793,28]