{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T05:52:26Z","timestamp":1791265946855,"version":"4.1.0"},"reference-count":35,"publisher":"Wiley","issue":"10","license":[{"start":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T00:00:00Z","timestamp":1791158400000},"content-version":"vor","delay-in-days":4,"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\/100008363","name":"Kr\u00e6ftens Bek\u00e6mpelse","doi-asserted-by":"publisher","award":["R269\u2010A15840"],"award-info":[{"award-number":["R269\u2010A15840"]}],"id":[{"id":"10.13039\/100008363","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100008363","name":"Kr\u00e6ftens Bek\u00e6mpelse","doi-asserted-by":"publisher","award":["R352\u2010A20676"],"award-info":[{"award-number":["R352\u2010A20676"]}],"id":[{"id":"10.13039\/100008363","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100011958","name":"Danmarks Frie Forskningsfond","doi-asserted-by":"publisher","award":["3101\u201000395B"],"award-info":[{"award-number":["3101\u201000395B"]}],"id":[{"id":"10.13039\/501100011958","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["J Applied Clin Med Phys"],"published-print":{"date-parts":[[2026,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Deep learning (DL) methods have shown strong performance in predicting dose distributions from patient anatomy and target structures; however, accurate dose prediction does not guarantee a deliverable plan after conversion into machine parameters and dose recalculation in the treatment planning system (TPS), motivating training objectives that explicitly couple prediction to deliverability. Lymphoma presents substantial clinical heterogeneity in target location, extent, and prescription\/fractionation, so fast automated methods that produce deliverable plans are particularly valuable.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>To investigate whether a deliverability\u2010aware training framework improves the end\u2010to\u2010end deliverability of DL\u2010generated intensity\u2010modulated radiotherapy (IMRT) plans for lymphoma.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      A retrospective single\u2010institution dataset of 619 plans (571 patients; 2009\u20132021) was used. Clinical volumetric modulated arc therapy (VMAT) plans were converted to standardized 15\u2010field coplanar sliding\u2010window IMRT reference plans in Eclipse using automated scripting. Two DL approaches were compared: a baseline dose\u2010only objective model (DL\u2010DO) and a multi\u2010objective model (DL\u2010MO) trained with dose supervision, fluence supervision, and a dose\u2013fluence\u2013dose cycle\u2010consistency constraint implemented via fixed inverse and forward networks. Predicted fluence maps were imported into the TPS, leaf\u2010sequenced, and recalculated with AcurosXB without further optimization; plan quality was evaluated on a held\u2010out test set (\n                      <jats:italic>n<\/jats:italic>\n                      \u00a0=\u00a061) using planning target volume (PTV)\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>98<\/jats:sub>\n                      ,\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>2<\/jats:sub>\n                      , and\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>max<\/jats:sub>\n                      with paired Wilcoxon testing and Holm correction.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      DL\u2010MO produced significantly better machine\u2010deliverable plans than DL\u2010DO, improving coverage and reducing hotspots (median paired differences: +12.31 percentage points (pp) in\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>98<\/jats:sub>\n                      , \u22123.43 pp in\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>2<\/jats:sub>\n                      , and \u22123.18 pp in\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>max<\/jats:sub>\n                      ; Holm\u2010corrected\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0&lt;\u00a00.0001 for all). Mean\u00a0\u00b1\u00a0SD PTV metrics were\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>98<\/jats:sub>\n                      \u00a0=\u00a090.10\u00a0\u00b1\u00a03.74%\u00a0vs. 78.59\u00a0\u00b1\u00a05.92%,\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>2<\/jats:sub>\n                      \u00a0=\u00a0106.28\u00a0\u00b1\u00a01.13%\u00a0vs. 109.65\u00a0\u00b1\u00a02.40%, and\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>max<\/jats:sub>\n                      \u00a0=\u00a0110.32\u00a0\u00b1\u00a02.20%\u00a0vs. 113.27\u00a0\u00b1\u00a03.03% (DL\u2010MO vs. DL\u2010DO). DL\u2010MO approached the TPS reference IMRT plans (\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>98<\/jats:sub>\n                      \u00a0=\u00a091.62\u00a0\u00b1\u00a03.45%) with residual hotspot elevation.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>\n                      Cycle\u2010consistency training improves the end\u2010to\u2010end deliverability of DL\u2010generated IMRT plans for heterogeneous lymphoma, yielding substantially better TPS\u2010recalculated, machine\u2010deliverable plans than conventional dose\u2010only supervision while maintaining sub\u2010second fluence\u2010map generation. However, statistically significant differences from the TPS reference remained for\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>98<\/jats:sub>\n                      ,\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>2<\/jats:sub>\n                      , and\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>max<\/jats:sub>\n                      .\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/acm2.70815","type":"journal-article","created":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T05:05:01Z","timestamp":1791263101000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Improving deep\u2010learning IMRT plan quality for lymphoma using deliverability\u2010aware training"],"prefix":"10.1002","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2347-6637","authenticated-orcid":false,"given":"Denis","family":"Kutn\u00e1r","sequence":"first","affiliation":[{"name":"Department of Computer Science University of Copenhagen  Copenhagen \u00d8 Denmark"},{"name":"Department of Oncology Rigshospitalet  Copenhagen \u00d8 Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ivan Richter","family":"Vogelius","sequence":"additional","affiliation":[{"name":"Department of Oncology Rigshospitalet  Copenhagen \u00d8 Denmark"},{"name":"Faculty of Health and Medical Sciences University of Copenhagen  Copenhagen N Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hristo Atanasov","family":"Georgiev","sequence":"additional","affiliation":[{"name":"Department of Computer Science University of Copenhagen  Copenhagen \u00d8 Denmark"},{"name":"Department of Oncology Rigshospitalet  Copenhagen \u00d8 Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6902-2190","authenticated-orcid":false,"given":"Lena","family":"Specht","sequence":"additional","affiliation":[{"name":"Department of Oncology Rigshospitalet  Copenhagen \u00d8 Denmark"},{"name":"Faculty of Health and Medical Sciences University of Copenhagen  Copenhagen N Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jens","family":"Petersen","sequence":"additional","affiliation":[{"name":"Department of Computer Science University of Copenhagen  Copenhagen \u00d8 Denmark"},{"name":"Department of Oncology Rigshospitalet  Copenhagen \u00d8 Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,10,5]]},"reference":[{"key":"e_1_2_11_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2016.04.005"},{"key":"e_1_2_11_3_1","doi-asserted-by":"publisher","DOI":"10.1259\/bjr.20180270"},{"key":"e_1_2_11_4_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13526"},{"key":"e_1_2_11_5_1","doi-asserted-by":"publisher","DOI":"10.1186\/1748\u2010717X\u20105\u201033"},{"key":"e_1_2_11_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2020.03.019"},{"key":"e_1_2_11_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.metrad.2024.100057"},{"key":"e_1_2_11_8_1","unstructured":"MahmoodR BabierA McNivenA DiamantA ChanTC Automated treatment planning in radiation therapy using generative adversarial networks inMachine learning for healthcare conference pages484\u2010499 PMLR;2018."},{"key":"e_1_2_11_9_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15461"},{"key":"e_1_2_11_10_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ab039b"},{"key":"e_1_2_11_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13896"},{"key":"e_1_2_11_12_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.70353"},{"key":"e_1_2_11_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13271"},{"key":"e_1_2_11_14_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/aa71f8"},{"key":"e_1_2_11_15_1","doi-asserted-by":"publisher","DOI":"10.3389\/frai.2020.00068"},{"key":"e_1_2_11_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2022.05.008"},{"key":"e_1_2_11_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2020.03.016"},{"key":"e_1_2_11_18_1","doi-asserted-by":"publisher","DOI":"10.1186\/s13014\u2010023\u201002287\u20104"},{"key":"e_1_2_11_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13955"},{"key":"e_1_2_11_20_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/abc12c"},{"key":"e_1_2_11_21_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13617"},{"key":"e_1_2_11_22_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ab4d8c"},{"key":"e_1_2_11_23_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2016.04.236"},{"key":"e_1_2_11_24_1","doi-asserted-by":"publisher","DOI":"10.1093\/jrr\/rrac063"},{"key":"e_1_2_11_25_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13845"},{"key":"e_1_2_11_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.adro.2021.100672"},{"key":"e_1_2_11_27_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13953"},{"key":"e_1_2_11_28_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/abca05"},{"key":"e_1_2_11_29_1","doi-asserted-by":"crossref","unstructured":"SmithLN Cyclical learning rates for training neural networks in2017 IEEE winter conference on applications of computer vision (WACV) pages464\u2010472 IEEE 2017. doi:10.1109\/WACV.2017.58","DOI":"10.1109\/WACV.2017.58"},{"key":"e_1_2_11_30_1","doi-asserted-by":"publisher","DOI":"10.23915\/distill.00003"},{"key":"e_1_2_11_31_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031\u20109155\/58\/1\/87"},{"key":"e_1_2_11_32_1","doi-asserted-by":"crossref","unstructured":"NguyenD BarkousaraieAS ShenC JiaX JiangS Generating Pareto optimal dose distributions for radiation therapy treatment planning inInternational Conference on Medical Image Computing and Computer\u2010Assisted Intervention pages59\u201067 Springer;2019.","DOI":"10.1007\/978-3-030-32226-7_7"},{"key":"e_1_2_11_33_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14374"},{"key":"e_1_2_11_34_1","doi-asserted-by":"publisher","DOI":"10.1259\/bjr.20210303"},{"key":"e_1_2_11_35_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2011.09.020"},{"key":"e_1_2_11_36_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2017.07.033"}],"container-title":["Journal of Applied Clinical Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.70815","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/acm2.70815","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.70815","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T05:05:05Z","timestamp":1791263105000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/acm2.70815"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10]]},"references-count":35,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2026,10]]}},"alternative-id":["10.1002\/acm2.70815"],"URL":"https:\/\/doi.org\/10.1002\/acm2.70815","archive":["Portico"],"relation":{},"ISSN":["1526-9914","1526-9914"],"issn-type":[{"value":"1526-9914","type":"print"},{"value":"1526-9914","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10]]},"assertion":[{"value":"2026-05-18","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-09-01","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-05","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70815"}}