{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,27]],"date-time":"2026-06-27T09:17:26Z","timestamp":1782551846692,"version":"3.54.5"},"reference-count":19,"publisher":"Wiley","issue":"5","license":[{"start":{"date-parts":[[2025,2,4]],"date-time":"2025-02-04T00:00:00Z","timestamp":1738627200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"},{"start":{"date-parts":[[2025,2,4]],"date-time":"2025-02-04T00:00:00Z","timestamp":1738627200000},"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":[[2025,5]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Although deep learning (DL) methods for reconstructing 3D magnetic resonance (MR) volumes from 2D MR images yield promising results, they require large amounts of training data to perform effectively. To overcome this challenge, fine\u2010tuning\u2014a transfer learning technique particularly effective for small datasets\u2014presents a robust solution for developing personalized DL models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>A 2D to 3D conditional generative adversarial network (GAN) model with a patient\u2010 and fraction\u2010specific fine\u2010tuning workflow was developed to reconstruct synthetic 3D MR volumes using orthogonal 2D MR images for online dose adaptation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A total of 2473 3D MR volumes were collected from 43 patients. The training and test datasets were separated into 34 and 9 patients, respectively. All patients underwent MR\u2010guided adaptive radiotherapy using the same imaging protocol. The population data contained 2047 3D MR volumes from the training dataset. Population data were used to train the population\u2010based GAN model. For each fraction of the remaining patients, the population model was fine\u2010tuned with the 3D MR volumes acquired before beam irradiation of the fraction, named the fine\u2010tuned model. The performance of the fine\u2010tuned model was tested using the 3D MR volume acquired immediately after the beam delivery of the fraction. The model's input was a pair of axial and sagittal MR images at the isocenter level, and the output was a 3D MR volume. Model performance was evaluated using the structural similarity index measure (SSIM), peak signal\u2010to\u2010noise ratio (PSNR), root mean square error (RMSE), and mean absolute error (MAE). Moreover, the prostate, bladder, and rectum in the predicted MR images were manually segmented. To assess geometric accuracy, the 2D Dice Similarity Coefficient (DSC) and 2D Hausdorff Distance (HD) were calculated.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      A total of 84 3D MR volumes were included in the performance testing. The mean\u00a0\u00b1\u00a0standard deviation (SD) of SSIM, PSNR, RMSE, and MAE were 0.64\u00a0\u00b1\u00a00.10, 93.9\u00a0\u00b1\u00a01.5\u00a0dB, 0.050\u00a0\u00b1\u00a00.009, and 0.036\u00a0\u00b1\u00a00.007 for the population model and 0.72\u00a0\u00b1\u00a00.09, 96.2\u00a0\u00b1\u00a01.8\u00a0dB, 0.041\u00a0\u00b1\u00a00.007, and 0.028\u00a0\u00b1\u00a00.006 for the fine\u2010tuned model, respectively. The image quality of the fine\u2010tuned model was significantly better than that of the population model (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0&lt;\u00a00.05). The mean\u00a0\u00b1\u00a0SD of DSC and HD of the population model were 0.79\u00a0\u00b1\u00a00.08 and 1.70\u00a0\u00b1\u00a02.35\u00a0mm for prostate, 0.81\u00a0\u00b1\u00a00.10 and 2.75\u00a0\u00b1\u00a01.53\u00a0mm for bladder, and 0.72\u00a0\u00b1\u00a00.08 and 1.93\u00a0\u00b1\u00a00.59\u00a0mm for rectum. Contrarily, the mean\u00a0\u00b1\u00a0SD of DSC and HD of the fine\u2010tuned model were 0.83\u00a0\u00b1\u00a00.06 and 1.29\u00a0\u00b1\u00a00.77\u00a0mm for prostate, 0.85\u00a0\u00b1\u00a00.07 and 2.16\u00a0\u00b1\u00a01.09\u00a0mm for bladder, and 0.77\u00a0\u00b1\u00a00.08 and 1.57\u00a0\u00b1\u00a00.52\u00a0mm for rectum. The geometric accuracy of the fine\u2010tuned model was significantly improved than that of the population model (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0&lt;\u00a00.05).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>By employing a patient\u2010 and fraction\u2010specific fine\u2010tuning approach, the GAN model demonstrated promising accuracy despite limited data availability.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.17668","type":"journal-article","created":{"date-parts":[[2025,2,4]],"date-time":"2025-02-04T21:34:56Z","timestamp":1738704896000},"page":"2785-2795","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Patient\u2010 and fraction\u2010specific magnetic resonance volume reconstruction from orthogonal images with generative adversarial networks"],"prefix":"10.1002","volume":"52","author":[{"given":"Hideaki","family":"Hirashima","sequence":"first","affiliation":[{"name":"Department of Radiation Oncology and Image\u2010Applied Therapy Graduate School of Medicine Kyoto University  Kyoto Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dejun","family":"Zhou","sequence":"additional","affiliation":[{"name":"Department of Advanced Medical Physics Graduate School of Medicine Kyoto University  Kyoto Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nobutaka","family":"Mukumoto","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haruo","family":"Inokuchi","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nobunari","family":"Hamaura","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mutsumi","family":"Yamagishi","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mai","family":"Sakagami","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Naoki","family":"Mukumoto","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mitsuhiro","family":"Nakamura","sequence":"additional","affiliation":[{"name":"Department of Advanced Medical Physics Graduate School of Medicine Kyoto University  Kyoto Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Keiko","family":"Shibuya","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Graduate School of Medicine Osaka Metropolitan University  Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,2,4]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1186\/s13014\u2010019\u20101308\u2010y"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2019.08.027"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2020.06.010"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2020.12.013"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ctro.2020.05.002"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1111\/1754\u20109485.12713"},{"key":"e_1_2_9_8_1","doi-asserted-by":"crossref","unstructured":"YingX GuoH MaK WuJ WengZ ZhengY X2CT\u2010gan: reconstructing CT from biplanar x\u2010rays with generative adversarial networks.Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2019;2019\u2010June;Dl:10611\u201010620. doi:10.1109\/CVPR.2019.01087","DOI":"10.1109\/CVPR.2019.01087"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11548\u2010021\u201002425\u2010x"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010024\u201065795\u20107"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11760\u2010024\u201003334\u20107"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2024.103283"},{"key":"e_1_2_9_13_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15793"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2018.2872043"},{"key":"e_1_2_9_15_1","doi-asserted-by":"crossref","unstructured":"PathakD KrahenbuhlP DonahueJ DarrellT EfrosAA Context encoders: feature learning by inpainting.Proceedings of the CVPR IEEE;2016 p.2536\u20102544. doi:10.1109\/CVPR.2016.278","DOI":"10.1109\/CVPR.2016.278"},{"key":"e_1_2_9_16_1","doi-asserted-by":"crossref","unstructured":"JiangL ShiSS QiXJ JiaJY.GAL: geometric adversarial loss for single\u2010view 3D\u2010object reconstruction.Computer vision\u2014Eccv 2018.2018;11212:820\u2010834. doi:10.1007\/978\u20103\u2010030\u201001237\u20103_49","DOI":"10.1007\/978-3-030-01237-3_49"},{"key":"e_1_2_9_17_1","doi-asserted-by":"crossref","unstructured":"IsolaP ZhuJY ZhouTH EfrosAA Image\u2010to\u2010image translation with conditional adversarial networks.30th IEEE Conference on Computer Vision and Pattern Recognition(Cvpr 2017);2017 p.5967\u20105976.doi:10.1109\/CVPR.2017.632","DOI":"10.1109\/CVPR.2017.632"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2018.09.013"},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.1120\/jacmp.v16i2.5013"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2014.09.040"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17668","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.17668","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17668","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,2]],"date-time":"2026-05-02T10:00:24Z","timestamp":1777716024000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17668"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,4]]},"references-count":19,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2025,5]]}},"alternative-id":["10.1002\/mp.17668"],"URL":"https:\/\/doi.org\/10.1002\/mp.17668","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,4]]},"assertion":[{"value":"2024-07-07","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-01-22","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-02-04","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}