{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T13:59:15Z","timestamp":1791381555460,"version":"4.3.3"},"reference-count":56,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2025,8,13]],"date-time":"2025-08-13T00:00:00Z","timestamp":1755043200000},"content-version":"vor","delay-in-days":12,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,8,1]],"date-time":"2025-08-01T00:00:00Z","timestamp":1754006400000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2025,8]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Magnetic Resonance (MR)\u2010based synthetic Computed Tomography (sCT) generation is an emerging promising technique, required for the transition from conventional planning workflows to MR\u2010only radiotherapy planning. This shift aims to replace CT acquisition with a sCT improving both cost efficiency and burden to the patient. Generative Adversarial Networks (GANs) have shown some of the best performance in this area.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>This study aims to identify optimal approaches to improve the quality and clinical applicability of MR\u2010based sCT generation for treatment planning by performing an extensive comparison of GAN architectures and parameters thereof. It focuses on the abdominal region, which still lacks certified medical products for sCT generation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>In order to improve the current state of deep learning technologies, we generated sCTs based on abdominal MR images of 154 cancer patients using GANs, varying the following parameters: (1) generator architectures (U\u2010Net, ResNet); (2) GAN architectures trained in paired (Pix2Pix) and unpaired fashion (CycleGAN and CUT); (3) number of input\u2010output channels (2D, 2.5D); (4) training set size. The quality of sCT generation was assessed by using both image similarity and dosimetric metrics; correlation between the two was evaluated. The dosimetric accuracy was evaluated through an automated process that compared the dose distributions of photon treatment plans calculated on sCT and CT images, using Dose\u2010Volume Histogram (DVH) parameters for tumor and organs at risk.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>The Pix2Pix model, trained in paired fashion with 2.5D input\u2010output channels and a ResNet generator emerged as the best\u2010performing model, achieving a mean absolute error (MAE, mean) of 63.21 HU, a planning target volume Dmean difference of \u20100.09%, and no outliers above 2% for other DVH parameters. This configuration addressed prior challenges of Pix2Pix with bone and rigid organ boundary generation, delivering robust results even for cases with significant air pockets. The 2D input\u2010output channel configuration showed beneficial for GANs trained in unpaired fashion, achieving a mean MAE of 66.97 HU for CycleGAN and 69.49 HU for CUT. Both delivered clinically applicable results, with mean DVH discrepancies below 0.8%. Expanding the training set size was essential for minimizing outliers in dosimetric parameters. High correlation was observed between the image similarity metrics\u2014MAE, MAE bones, structural similarity index measure\u2014and target DVH parameters, with Pearson coefficients ranging from 0.77 to 0.9. However, within the clinically relevant range of DVH deviations (\u00b1\u00a02%), stochastic variations obscured linear trends.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The study provided a new benchmark for the abdominal sCT generation task, showing its clinical applicability for treatment planning and further advancing the state\u2010of\u2010the\u2010art. This study also confirmed that image similarity metrics alone can not reliably predict small dosimetric deviations within a clinical threshold; but contributed by identifying specific metrics that correlate with DVH discrepancies above\u00a0\u00b1\u00a05%, offering valuable tools for training, evaluation, and standardization of reporting across studies.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.18038","type":"journal-article","created":{"date-parts":[[2025,8,14]],"date-time":"2025-08-14T04:41:30Z","timestamp":1755146490000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":9,"title":["A comprehensive comparative study of generative adversarial network architectures for synthetic computed tomography generation in the abdomen"],"prefix":"10.1002","volume":"52","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3280-3409","authenticated-orcid":false,"given":"Mariia","family":"Lapaeva","sequence":"first","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"},{"name":"Artificial Intelligence and Machine Learning Group Department of Informatics University of Zurich  Zurich Switzerland"},{"name":"Computer Vision Laboratory ETH Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4504-611X","authenticated-orcid":false,"given":"Agustina","family":"La Greca Saint\u2010Esteven","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"},{"name":"Computer Vision Laboratory ETH Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7110-6617","authenticated-orcid":false,"given":"Philipp","family":"Wallimann","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3647-5916","authenticated-orcid":false,"given":"Nicolaus","family":"Andratschke","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7146-9071","authenticated-orcid":false,"given":"Matthias","family":"Guckenberger","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1489-7448","authenticated-orcid":false,"given":"Manuel","family":"G\u00fcnther","sequence":"additional","affiliation":[{"name":"Artificial Intelligence and Machine Learning Group Department of Informatics University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4387-1522","authenticated-orcid":false,"given":"Stephanie","family":"Tanadini\u2010Lang","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8755-377X","authenticated-orcid":false,"given":"Riccardo","family":"Dal Bello","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University Hospital Zurich and University of Zurich  Zurich Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,8,13]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.semradonc.2023.10.015"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2017.08.043"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1120\/jacmp.v15i5.4851"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1186\/s13014\u2010016\u20100747\u2010y"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2015.08.045"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4967480"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.14499"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2014.09.015"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/aae66c"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1186\/1748\u2010717X\u20104\u201054"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15150"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.103046"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.12155"},{"key":"e_1_2_8_15_1","doi-asserted-by":"crossref","unstructured":"IsolaP ZhuJY ZhouT EfrosAA Image\u2010to\u2010image translation with conditional adversarial networks. In:2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR);2017:5967\u20105976. doi:10.1109\/CVPR.2017.632","DOI":"10.1109\/CVPR.2017.632"},{"key":"e_1_2_8_16_1","doi-asserted-by":"crossref","unstructured":"ZhuJY ParkT IsolaP EfrosAA Unpaired image\u2010to\u2010image translation using cycle\u2010consistent adversarial networks. In:2017 IEEE International Conference on Computer Vision (ICCV);2017:2242\u20102251. doi:10.1109\/ICCV.2017.244","DOI":"10.1109\/ICCV.2017.244"},{"key":"e_1_2_8_17_1","doi-asserted-by":"crossref","unstructured":"ParkT EfrosAA ZhangR ZhuJY Contrastive learning for unpaired image\u2010to\u2010image translation. In:VedaldiA BischofH BroxT FrahmJM eds.Computer Vision\u2014ECCV 2020.Springer International Publishing;2020:319\u2010345. doi:10.1007\/978\u20103\u2010030\u201058545\u20107_19","DOI":"10.1007\/978-3-030-58545-7_19"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11263\u2010023\u201001791\u20100"},{"key":"e_1_2_8_19_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2021.07.027"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.105556"},{"key":"e_1_2_8_21_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00530\u2010024\u201001349\u20101"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2020.05.006"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2022.11.011"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13327"},{"key":"e_1_2_8_25_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2023.100471"},{"key":"e_1_2_8_26_1","doi-asserted-by":"crossref","unstructured":"DovletovG L\u00f6rcksS PauliJ GratzM QuickHH Double GRAD\u2010CAM guidance for improved MRI\u2010based Pseudo\u2010CT synthesis. In:DesernoTM HandelsH MaierA Maier\u2010HeinK PalmC TolxdorffT eds.Bildverarbeitung f\u00fcr die Medizin 2023.Springer Fachmedien;2023:45\u201050. doi:10.1007\/978\u20103\u2010658\u201041657\u20107_13","DOI":"10.1007\/978-3-658-41657-7_13"},{"key":"e_1_2_8_27_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2020.10.018"},{"key":"e_1_2_8_28_1","doi-asserted-by":"publisher","DOI":"10.1007\/s13534\u2010021\u201000195\u20108"},{"key":"e_1_2_8_29_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2020.09.029"},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13672"},{"key":"e_1_2_8_31_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13617"},{"key":"e_1_2_8_32_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2951924"},{"key":"e_1_2_8_33_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2024.103276"},{"key":"e_1_2_8_34_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00066\u2010021\u201001771\u20108"},{"key":"e_1_2_8_35_1","doi-asserted-by":"publisher","DOI":"10.3390\/jcm11020364"},{"key":"e_1_2_8_36_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13927"},{"key":"e_1_2_8_37_1","unstructured":"ZhuJY junyanz\/pytorch\u2010CycleGAN\u2010and\u2010pix2pix. Published online October 1 2024. Accessed October 1 2024.https:\/\/github.com\/junyanz\/pytorch\u2010CycleGAN\u2010and\u2010pix2pix"},{"key":"e_1_2_8_38_1","unstructured":"ParkT taesungp\/contrastive\u2010unpaired\u2010translation. Published online September 30 2024. Accessed October 1 2024.https:\/\/github.com\/taesungp\/contrastive\u2010unpaired\u2010translation"},{"key":"e_1_2_8_39_1","doi-asserted-by":"crossref","unstructured":"RonnebergerO FischerP BroxT U\u2010net: convolutional networks for biomedical image segmentation. In:NavabN HorneggerJ WellsWM FrangiAF eds.Medical Image Computing and Computer\u2010Assisted Intervention\u2014MICCAI 2015.Springer International Publishing;2015:234\u2010241. doi:10.1007\/978\u20103\u2010319\u201024574\u20104_28","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_8_40_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS SunJ Deep residual learning for image recognition. In:2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR);2016:770\u2010778. doi:10.1109\/CVPR.2016.90","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_8_41_1","unstructured":"LapaevaM MR\u2010based synthetic CT generation for MR\u2010guided radiotherapy.Master's Thesis. Accessed October 1 2024.https:\/\/www.merlin.uzh.ch\/contributionDocument\/download\/15045"},{"key":"e_1_2_8_42_1","unstructured":"PaszkeA GrossS MassaF et\u00a0al.PyTorch: an imperative style high\u2010performance deep learning library. In:Proceedings of the 33rd International Conference on Neural Information Processing Systems.Curran Associates Inc.;2019:8026\u20108037."},{"key":"e_1_2_8_43_1","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.1896.0007"},{"key":"e_1_2_8_44_1","doi-asserted-by":"publisher","DOI":"10.2307\/1412159"},{"key":"e_1_2_8_45_1","doi-asserted-by":"publisher","DOI":"10.1213\/ANE.0000000000002864"},{"key":"e_1_2_8_46_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2022.969463"},{"key":"e_1_2_8_47_1","unstructured":"HeuselM RamsauerH UnterthinerT NesslerB HochreiterS GANs trained by a two time\u2010scale update rule converge to a local nash equilibrium. In:Proceedings of the 31st International Conference on Neural Information Processing Systems. NIPS\u201917.Curran Associates Inc.;2017:6629\u20106640."},{"key":"e_1_2_8_48_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.12251"},{"key":"e_1_2_8_49_1","unstructured":"ICRU Report 83 Prescribing Recording and Reporting Intensity\u2010Modulated Photon\u2010Beam Therapy (IMRT)\u2014ICRU. Accessed October 18 2024.https:\/\/www.icru.org\/report\/prescribing\u2010recording\u2010and\u2010reporting\u2010intensity\u2010modulated\u2010photon\u2010beam\u2010therapy\u2010imrticru\u2010report\u201083\/"},{"key":"e_1_2_8_50_1","doi-asserted-by":"publisher","DOI":"10.1186\/1748\u2010717X\u20109\u201016"},{"key":"e_1_2_8_51_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-00536-8_4"},{"key":"e_1_2_8_52_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2020.2990339"},{"key":"e_1_2_8_53_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-66535-6_13"},{"key":"e_1_2_8_54_1","first-page":"25400","volume-title":"Advances in Neural Information Processing Systems","author":"Sangalli S","year":"2021"},{"key":"e_1_2_8_55_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.compmedimag.2024.102353"},{"key":"e_1_2_8_56_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.16847"},{"key":"e_1_2_8_57_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2022.105151"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.18038","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.18038","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.18038","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T13:36:33Z","timestamp":1791380193000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.18038"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8]]},"references-count":56,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2025,8]]}},"alternative-id":["10.1002\/mp.18038"],"URL":"https:\/\/doi.org\/10.1002\/mp.18038","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8]]},"assertion":[{"value":"2024-12-14","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-07-11","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-08-13","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e18038"}}