{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,16]],"date-time":"2026-11-16T06:08:42Z","timestamp":1794809322859,"version":"4.9.1"},"reference-count":31,"publisher":"Wiley","issue":"4","license":[{"start":{"date-parts":[[2022,11,12]],"date-time":"2022-11-12T00:00:00Z","timestamp":1668211200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2023,4]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:sec><jats:title>Purpose<\/jats:title><jats:p>Online adaptive radiotherapy (ART) requires accurate and efficient auto\u2010segmentation of target volumes and organs\u2010at\u2010risk (OARs) in mostly cone\u2010beam computed tomography (CBCT) images, which often have severe artifacts and lack soft\u2010tissue contrast, making direct segmentation very challenging. Propagating expert\u2010drawn contours from the pretreatment planning CT through traditional or deep learning (DL)\u2010based deformable image registration (DIR) can achieve improved results in many situations. Typical DL\u2010based DIR models are population based, that is, trained with a dataset for a population of patients, and so they may be affected by the generalizability problem.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>In this paper, we propose a method called test\u2010time optimization (TTO) to refine a pretrained DL\u2010based DIR population model, first for each individual test patient, and then progressively for each fraction of online ART treatment. Our proposed method is less susceptible to the generalizability problem and thus can improve overall performance of different DL\u2010based DIR models by improving model accuracy, especially for outliers. Our experiments used data from 239 patients with head\u2010and\u2010neck squamous cell carcinoma to test the proposed method. First, we trained a population model with 200 patients and then applied TTO to the remaining 39 test patients by refining the trained population model to obtain 39 individualized models. We compared each of the individualized models with the population model in terms of segmentation accuracy.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>The average improvement of the Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95) of segmentation can be up to 0.04 (5%) and 0.98\u00a0mm (25%), respectively, with the individualized models compared to the population model over 17 selected OARs and a target of 39 patients. Although the average improvement may seem mild, we found that the improvement for outlier patients with structures of large anatomical changes is significant. The number of patients with at least 0.05 DSC improvement or 2\u00a0mm HD95 improvement by TTO averaged over the 17 selected structures for the state\u2010of\u2010the\u2010art architecture VoxelMorph is 10 out of 39 test patients. By deriving the individualized model using TTO from the pretrained population model, TTO models can be ready in about 1\u00a0min. We also generated the adapted fractional models for each of the 39 test patients by progressively refining the individualized models using TTO to CBCT images acquired at later fractions of online ART treatment. When adapting the individualized model to a later fraction of the same patient, the model can be ready in less than a minute with slightly improved accuracy.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>The proposed TTO method is well suited for online ART and can boost segmentation accuracy for DL\u2010based DIR models, especially for outlier patients where the pretrained models fail.<\/jats:p><\/jats:sec>","DOI":"10.1002\/mp.15960","type":"journal-article","created":{"date-parts":[[2022,10,31]],"date-time":"2022-10-31T02:46:36Z","timestamp":1667184396000},"page":"1947-1961","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":11,"title":["Segmentation by test\u2010time optimization for CBCT\u2010based adaptive radiation therapy"],"prefix":"10.1002","volume":"50","author":[{"given":"Xiao","family":"Liang","sequence":"first","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jaehee","family":"Chun","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology, Yonsei Cancer Center Yonsei University College of Medicine  Seoul South Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Howard","family":"Morgan","sequence":"additional","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ti","family":"Bai","sequence":"additional","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dan","family":"Nguyen","sequence":"additional","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Justin","family":"Park","sequence":"additional","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Steve","family":"Jiang","sequence":"additional","affiliation":[{"name":"Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology University of Texas Southwestern Medical Center  Dallas Texas USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2022,11,12]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2020.10.021"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.semradonc.2019.02.001"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11604-016-0604-9"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2007.01.038"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0175906"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/abd953"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2020.12.004"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2009.2035616"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.mri.2012.05.001"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/55\/1\/012"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2008.10.040"},{"key":"e_1_2_8_13_1","unstructured":"JaderbergM SimonyanK ZissermanA KavukcuogluK.Spatial transformer networks. Advances in neural information processing systems.2015;28."},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2019.07.006"},{"key":"e_1_2_8_15_1","doi-asserted-by":"crossref","unstructured":"KuangD SchmahT.FAIM \u2013 a ConvNet method for unsupervised 3D medical image registration.International Workshop on Machine Learning in Medical Imaging.2019;11861:646\u2010654.","DOI":"10.1007\/978-3-030-32692-0_74"},{"key":"e_1_2_8_16_1","doi-asserted-by":"crossref","unstructured":"ZhaoS DongY ChangE XuY.Recursive cascaded networks for unsupervised medical image registration.Paper presented at: 2019 IEEE\/CVF International Conference on Computer Vision (ICCV); 27 Oct\u20132 Nov 2019.2019.","DOI":"10.1109\/ICCV.2019.01070"},{"key":"e_1_2_8_17_1","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2019.2951024"},{"key":"e_1_2_8_18_1","doi-asserted-by":"crossref","unstructured":"BalakrishnanG ZhaoA SabuncuMR DalcaAV GuttagJ.An unsupervised learning model for deformable medical image registration.Paper presented at: 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition; 18\u201323 June 2018.2018.","DOI":"10.1109\/CVPR.2018.00964"},{"key":"e_1_2_8_19_1","doi-asserted-by":"crossref","unstructured":"FerranteE OktayO GlockerB MiloneDH.On the adaptability of unsupervised CNN\u2010based deformable image registration to unseen image domains.Paper presented at: International Workshop on Machine Learning in Medical Imaging.2018.","DOI":"10.1007\/978-3-030-00919-9_34"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14545"},{"key":"e_1_2_8_21_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2020.2972616"},{"key":"e_1_2_8_22_1","unstructured":"KoriA KrishnamurthiG.Zero shot learning for multi\u2010modal real time image registration.2019. arXiv preprint arXiv:190806213."},{"key":"e_1_2_8_23_1","volume-title":"Deep Learning","author":"Goodfellow I","year":"2016"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(89)90020-8"},{"key":"e_1_2_8_25_1","doi-asserted-by":"publisher","DOI":"10.1109\/LSP.2020.3005051"},{"key":"e_1_2_8_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.acha.2019.06.004"},{"key":"e_1_2_8_27_1","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.2.246"},{"key":"e_1_2_8_28_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00365\u2010021\u201009546\u20101"},{"key":"e_1_2_8_29_1","unstructured":"MihalkovaL HuynhT MooneyRJ.Mapping and revising Markov logic networks for transfer learning. In Aaai 2007 Jul 22 (Vol. 7 pp.608\u2010614)."},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2018.2864759"},{"key":"e_1_2_8_31_1","doi-asserted-by":"crossref","unstructured":"TanC SunF KongT ZhangW YangC LiuC.A survey on deep transfer learning.Paper presented at: Artificial Neural Networks and Machine Learning \u2013 ICANN 2018.Cham;2018.","DOI":"10.1007\/978-3-030-01424-7_27"},{"key":"e_1_2_8_32_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/ab22f9"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.15960","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.15960","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.15960","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,8,1]],"date-time":"2025-08-01T10:23:41Z","timestamp":1754043821000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.15960"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,11,12]]},"references-count":31,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2023,4]]}},"alternative-id":["10.1002\/mp.15960"],"URL":"https:\/\/doi.org\/10.1002\/mp.15960","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,11,12]]},"assertion":[{"value":"2022-02-24","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-08-21","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-11-12","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}