{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T16:57:02Z","timestamp":1792083422849,"version":"4.4.0"},"reference-count":38,"publisher":"Wiley","issue":"10","license":[{"start":{"date-parts":[[2026,10,4]],"date-time":"2026-10-04T00:00:00Z","timestamp":1791072000000},"content-version":"vor","delay-in-days":3,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/501100001111","name":"Cancer Australia","doi-asserted-by":"publisher","award":["APP1143778"],"award-info":[{"award-number":["APP1143778"]}],"id":[{"id":"10.13039\/501100001111","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000925","name":"National Health and Medical Research Council","doi-asserted-by":"publisher","award":["2041792"],"award-info":[{"award-number":["2041792"]}],"id":[{"id":"10.13039\/501100000925","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100025520","name":"Medical Research Future Fund","doi-asserted-by":"publisher","award":["NCRI000139"],"award-info":[{"award-number":["NCRI000139"]}],"id":[{"id":"10.13039\/501100025520","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2026,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>During radiation therapy, liver tumor motion can reduce dose delivery accuracy and increase irradiation of adjacent healthy tissues. Because liver tumors are difficult to visualize directly on kilovoltage (kV) X\u2010ray images, fiducial markers are commonly implanted as surrogates for tumor position during treatment. Conventional marker\u2010segmentation approaches such as template matching can lose accuracy when markers are obscured by bone, surgical clips, or stents. Although Convolutional Neural Network (CNN)\u2010based approaches have shown strong performance for image analysis, their application has been constrained to more static organs such as the prostate or has used single\u2010center and vendor\u2010specific datasets, limiting assessment of model transportability across different institutions and anatomic sites.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>The aim of this study was to develop, integrate, and evaluate a CNN\u2010based method for fiducial marker detection in kV X\u2010ray images to support real\u2010time image\u2010guided radiation therapy (IGRT) during liver radiotherapy using a multi\u2010institutional, multi\u2010platform dataset spanning both large and fast free\u2010breathing motion and slow and small breath\u2010hold motion.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A compact Convolutional Neural Network (CNN) was trained on 314,625 kilovoltage (kV) X\u2010ray images encompassing 28 patients from the multi\u2010institutional TROG 17.03 LARK clinical trial (NCT02984566). The model was validated using a hold\u2010out set of 31,463 images (10%) from the same cohort of 12 patients and 55 treatment fractions, spanning three centers and three respiratory motion\u2010management techniques. The CNN was tested on 4184 images from 16 patients and 55 fractions across three centers and three motion\u2010management techniques. The ground truth for testing was manually segmented marker positions from every 10 degrees of gantry rotation for each fraction of each test patient. Based on AAPM guidelines (TG147 and TGB135.B), feasibility for clinical implementation was predefined as\u00a0&gt;\u00a095% of marker positions being within 2\u00a0mm of the ground truth position in each dimension, and processing time less than 150\u00a0ms per image using a simulated real\u2010time Kilovoltage Intrafraction Monitoring (KIM) framework. Additional evaluation metrics included sensitivity, specificity, and the area under the precision\u2010recall curve (AUC).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>For the unseen test patient data, the marker position was segmented by the CNN within 2\u00a0mm in 95.4% of frames on the X axis and 97.6% of frames on the Y axis, meeting AAPM criteria. Sensitivity reached 97.76%, specificity was 99.94%, and the AUC was 0.9964. No statistically significant difference in localization error was observed between breath\u2010hold and free\u2010breathing treatments, although this comparison was limited by the small number of free\u2010breathing patients. Localization error differed significantly between Varian and Elekta linacs due to imbalance within the dataset. The processing time for each image was 50\u201360\u00a0ms using a NVIDIA GeForce RTX 3070 GPU.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>A CNN\u2010based method for fiducial marker detection in the liver was developed and evaluated on multi\u2010institutional kV X\u2010ray images spanning free\u2010breathing and breath\u2010hold motion, as well as Varian and Elekta linac platforms. The method satisfied the predefined feasibility criteria for positional accuracy and computation time, supporting future translation toward clinical implementation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.70691","type":"journal-article","created":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T04:59:35Z","timestamp":1791176375000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Deep learning\u2010based CNN method for fiducial marker detection in kilovoltage X\u2010ray images for liver tumor motion monitoring"],"prefix":"10.1002","volume":"53","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-8226-1403","authenticated-orcid":false,"given":"Frances","family":"Kan","sequence":"first","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Freeman","family":"Jin","sequence":"additional","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Adam","family":"Mylonas","sequence":"additional","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Benjamin","family":"Zwan","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Central Coast Cancer Centre  Gosford New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Trang","family":"Nguyen","sequence":"additional","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Trevor","family":"Moodie","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Crown Princess Mary Cancer Centre  Westmead New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nicholas","family":"Hardcastle","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Peter MacCallum Cancer Centre  Melbourne Victoria Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sau Fan","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Princess Alexandra Hospital  Woolloongabba Queensland Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daniel","family":"Mason","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Nepean Cancer and Wellness Centre, Nepean Hospital  Kingswood New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tim","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Crown Princess Mary Cancer Centre  Westmead New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yoo Young","family":"Lee","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Princess Alexandra Hospital  Woolloongabba Queensland Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paul","family":"Keall","sequence":"additional","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chandrima","family":"Sengupta","sequence":"additional","affiliation":[{"name":"Faculty of Medicine and Health, Image X Institute The University of Sydney  Sydney New South Wales Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,10,4]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejca.2009.08.011"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.tipsro.2023.100217"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.4103\/tcmj.tcmj_247_20"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2023.110031"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2019.11.008"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.21037\/qims\u201023\u20101489"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1111\/1754\u20109485.13285"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13519"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2018.02.010"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1177\/153303461000900502"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14501"},{"key":"e_1_2_9_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.12978"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2020.07.002"},{"key":"e_1_2_9_15_1","doi-asserted-by":"publisher","DOI":"10.1016\/S1470\u20102045(20)30034\u20106"},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0167\u20108140(16)32139\u20109"},{"key":"e_1_2_9_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2023.06.2525"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2020.03.014"},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2023.1112481"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2014.05.007"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2012.12.017"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3658566"},{"key":"e_1_2_9_23_1","doi-asserted-by":"publisher","DOI":"10.1088\/2057\u20101976\/acb550"},{"key":"e_1_2_9_24_1","doi-asserted-by":"publisher","DOI":"10.4103\/jmp.JMP_117_20"},{"key":"e_1_2_9_25_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/acc77c"},{"key":"e_1_2_9_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.semradonc.2011.05.001"},{"key":"e_1_2_9_27_1","doi-asserted-by":"publisher","DOI":"10.1186\/s40537\u2010019\u20100276\u20102"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3681967"},{"key":"e_1_2_9_29_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.17478"},{"key":"e_1_2_9_30_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14625"},{"key":"e_1_2_9_31_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4898119"},{"key":"e_1_2_9_32_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3524253"},{"key":"e_1_2_9_33_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.2349696"},{"key":"e_1_2_9_34_1","doi-asserted-by":"publisher","DOI":"10.12659\/PJR.890745"},{"key":"e_1_2_9_35_1","article-title":"Detailed image data quality and cleaning practices for artificial intelligence tools for breast cancer","volume":"8","author":"Wu DY","year":"2024","journal-title":"CO Clin Cancer Inform"},{"key":"e_1_2_9_36_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10479\u2010021\u201004006\u20102"},{"key":"e_1_2_9_37_1","doi-asserted-by":"crossref","unstructured":"HuS YuanJ WangS LeiB Cross\u2010modality synthesis from MRI to PET using adversarial U\u2010net with different normalization.2019 International Conference on Medical Imaging Physics and Engineering (ICMIPE).2019:1\u20105.","DOI":"10.1109\/ICMIPE47306.2019.9098219"},{"key":"e_1_2_9_38_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS et\u00a0al Deep residual learning for image recognition.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2015.","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_9_39_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0179676"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70691","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.70691","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70691","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T16:18:25Z","timestamp":1792081105000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.70691"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10]]},"references-count":38,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2026,10]]}},"alternative-id":["10.1002\/mp.70691"],"URL":"https:\/\/doi.org\/10.1002\/mp.70691","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10]]},"assertion":[{"value":"2026-05-21","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-23","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-04","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70691"}}