{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T05:17:51Z","timestamp":1792214271171,"version":"4.4.0"},"reference-count":56,"publisher":"Wiley","issue":"7","license":[{"start":{"date-parts":[[2022,5,9]],"date-time":"2022-05-09T00:00:00Z","timestamp":1652054400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["J Applied Clin Med Phys"],"published-print":{"date-parts":[[2022,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>Deep learning\u2013based knowledge\u2010based planning (KBP) methods have been introduced for radiotherapy dose distribution prediction to reduce the planning time and maintain consistent high\u2010quality plans. This paper presents a novel KBP model using an attention\u2010gating mechanism and a three\u2010dimensional (3D) U\u2010Net for intensity\u2010modulated radiation therapy (IMRT) 3D dose distribution prediction in head\u2010and\u2010neck cancer.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>A total of 340 head\u2010and\u2010neck cancer plans, representing the OpenKBP\u20142020 AAPM Grand Challenge data set, were used in this study. All patients were treated with the IMRT technique and a dose prescription of 70\u00a0Gy. The data set was randomly divided into 64%\/16%\/20% as training\/validation\/testing cohorts. An attention\u2010gated 3D U\u2010Net architecture model was developed to predict full 3D dose distribution. The developed model was trained using the mean\u2010squared error loss function, Adam optimization algorithm, a learning rate of 0.001, 120 epochs, and batch size of 4. In addition, a baseline U\u2010Net model was also similarly trained for comparison. The model performance was evaluated on the testing data set by comparing the generated dose distributions against the ground\u2010truth dose distributions using dose statistics and clinical dosimetric indices. Its performance was also compared to the baseline model and the reported results of other deep learning\u2010based dose prediction models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The proposed attention\u2010gated 3D U\u2010Net model showed high capability in accurately predicting 3D dose distributions that closely replicated the ground\u2010truth dose distributions of 68 plans in the test set. The average value of the mean\u00a0absolute dose error was 2.972\u00a0\u00b1\u00a01.220\u00a0Gy (vs. 2.920\u00a0\u00b1\u00a01.476\u00a0Gy for a baseline\u00a0U\u2010Net) in the brainstem, 4.243\u00a0\u00b1\u00a01.791\u00a0Gy (vs. 4.530\u00a0\u00b1\u00a02.295\u00a0Gy for a baseline U\u2010Net) in the left parotid, 4.622\u00a0\u00b1\u00a01.975\u00a0Gy (vs.\u00a04.223\u00a0\u00b1\u00a01.816\u00a0Gy\u00a0for\u00a0a\u00a0baseline U\u2010Net) in the right parotid, 3.346\u00a0\u00b1\u00a01.198\u00a0Gy (vs. 2.958\u00a0\u00b1\u00a00.888\u00a0Gy for a baseline U\u2010Net) in the spinal cord, 6.582\u00a0\u00b1\u00a03.748\u00a0Gy (vs. 5.114\u00a0\u00b1\u00a02.098\u00a0Gy\u00a0for a baseline U\u2010Net) in the esophagus, 4.756\u00a0\u00b1\u00a01.560\u00a0Gy\u00a0(vs. 4.992\u00a0\u00b1\u00a02.030\u00a0Gy for a baseline U\u2010Net) in the mandible, 4.501\u00a0\u00b1\u00a01.784\u00a0Gy (vs.\u00a04.925\u00a0\u00b1\u00a02.347\u00a0Gy for a baseline U\u2010Net) in the larynx, 2.494\u00a0\u00b1\u00a00.953\u00a0Gy (vs. 2.648\u00a0\u00b1\u00a01.247\u00a0Gy for a baseline U\u2010Net) in the PTV_70, and 2.432\u00a0\u00b1\u00a02.272\u00a0Gy (vs. 2.811\u00a0\u00b1\u00a02.896\u00a0Gy for a baseline U\u2010Net) in the body contour. The average difference in predicting the\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>99<\/jats:sub>\n                      value for the targets (PTV_70, PTV_63, and PTV_56) was 2.50\u00a0\u00b1\u00a01.77\u00a0Gy. For the organs at risk, the average difference in predicting the  (brainstem, spinal cord, and mandible) and  (left parotid, right parotid, esophagus, and larynx) values was 1.43\u00a0\u00b1\u00a01.01 and 2.44\u00a0\u00b1\u00a01.73\u00a0Gy, respectively. The average value of the homogeneity index was 7.99\u00a0\u00b1\u00a01.45 for the predicted plans versus 5.74\u00a0\u00b1\u00a02.95 for the ground\u2010truth plans, whereas the average value of the conformity index was 0.63\u00a0\u00b1\u00a00.17 for the predicted plans versus 0.89\u00a0\u00b1\u00a00.19 for the ground\u2010truth plans. The proposed model needs less than 5\u00a0s to predict a full 3D dose distribution of 64\u00a0\u00d7\u00a064\u00a0\u00d7\u00a064 voxels for a new patient that is sufficient for real\u2010time applications.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The attention\u2010gated 3D U\u2010Net model demonstrated a capability in predicting accurate 3D dose distributions for head\u2010and\u2010neck IMRT plans with consistent quality. The prediction performance of the proposed model was overall superior to a baseline standard U\u2010Net model, and it was also competitive to the performance of the best state\u2010of\u2010the\u2010art dose prediction method reported in the literature. The proposed model could be used to obtain dose distributions for decision\u2010making before planning, quality assurance of planning, and guiding\u2010automated planning for improved plan consistency, quality, and planning efficiency.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/acm2.13630","type":"journal-article","created":{"date-parts":[[2022,5,9]],"date-time":"2022-05-09T14:05:53Z","timestamp":1652105153000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":46,"title":["Attention\u2010aware 3D U\u2010Net convolutional neural network for knowledge\u2010based planning 3D dose distribution prediction of head\u2010and\u2010neck cancer"],"prefix":"10.1002","volume":"23","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1286-475X","authenticated-orcid":false,"given":"Alexander F. I.","family":"Osman","sequence":"first","affiliation":[{"name":"Department of Medical Physics Al\u2010Neelain University  Khartoum Sudan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nissren M.","family":"Tamam","sequence":"additional","affiliation":[{"name":"Department of Physics College of Science Princess Nourah bint Abdulrahman University  Riyadh Saudi Arabia"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2022,5,9]]},"reference":[{"key":"e_1_2_10_2_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/36\/9\/004"},{"key":"e_1_2_10_3_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.2818738"},{"key":"e_1_2_10_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2011.09.015"},{"key":"e_1_2_10_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.semradonc.2019.02.003"},{"key":"e_1_2_10_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2009.09.040"},{"key":"e_1_2_10_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2005.08.003"},{"key":"e_1_2_10_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2005.06.021"},{"key":"e_1_2_10_9_1","doi-asserted-by":"publisher","DOI":"10.1093\/jnci\/djn020"},{"key":"e_1_2_10_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2010.11.030"},{"key":"e_1_2_10_11_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4757927"},{"key":"e_1_2_10_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2011.11.012"},{"key":"e_1_2_10_13_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2020.580919"},{"key":"e_1_2_10_14_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4923174"},{"key":"e_1_2_10_15_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2017.04.011"},{"key":"e_1_2_10_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2018.06.023"},{"key":"e_1_2_10_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2018.11.002"},{"key":"e_1_2_10_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2018.03.008"},{"key":"e_1_2_10_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.12219"},{"key":"e_1_2_10_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2017.03.002"},{"key":"e_1_2_10_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2018.08.016"},{"key":"e_1_2_10_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2018.04.082"},{"key":"e_1_2_10_23_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13896"},{"key":"e_1_2_10_24_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13597"},{"key":"e_1_2_10_25_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2021.01.006"},{"key":"e_1_2_10_26_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ac0856"},{"key":"e_1_2_10_27_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14114"},{"key":"e_1_2_10_28_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2021.697995"},{"key":"e_1_2_10_29_1","doi-asserted-by":"publisher","DOI":"10.1007\/s12032\u2010020\u201001374\u2010w"},{"key":"e_1_2_10_30_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11684\u2010020\u20100761\u20101"},{"key":"e_1_2_10_31_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13337"},{"key":"e_1_2_10_32_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2004.08.037"},{"key":"e_1_2_10_33_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2009.03.079"},{"key":"e_1_2_10_34_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13262"},{"key":"e_1_2_10_35_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13271"},{"key":"e_1_2_10_36_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ab039b"},{"key":"e_1_2_10_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2020.10.028"},{"key":"e_1_2_10_38_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14827"},{"key":"e_1_2_10_39_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15034"},{"key":"e_1_2_10_40_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13490"},{"key":"e_1_2_10_41_1","unstructured":"MahmoodR BabierA McNivenA DiamantA ChanTCY.Automated treatment planning in radiation therapy using generative adversarial networks. In:FinaleD\u2010V JimF KenJ et\u00a0al. eds.Proceedings of the 3rd Machine Learning for Healthcare Conference. Proceedings of Machine Learning Research: PMLR.2018:484\u2010499."},{"key":"e_1_2_10_42_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14774"},{"key":"e_1_2_10_43_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14770"},{"key":"e_1_2_10_44_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/aabd14"},{"key":"e_1_2_10_45_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2019.01.012"},{"key":"e_1_2_10_46_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010020\u201068062\u20107"},{"key":"e_1_2_10_47_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14845"},{"key":"e_1_2_10_48_1","first-page":"10","volume-title":"Feasibility of CT\u2010Only 3D Dose Prediction for VMAT Prostate Plans Using Deep Learning","author":"Willems S","year":"2019"},{"key":"e_1_2_10_49_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010019\u201052262\u2010x"},{"key":"e_1_2_10_50_1","doi-asserted-by":"crossref","unstructured":"RonnebergerO FischerP BroxT.U\u2010net: convolutional networks for biomedical image segmentation. In:MICCAI.Springer;2015:234\u2010241.","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_10_51_1","unstructured":"NairV HintonGE.Rectified linear units improve restricted Boltzmann machines. In:Proc. 27th Int Conf Mach Learn.2010:807\u2010814."},{"key":"e_1_2_10_52_1","first-page":"1929","article-title":"Dropout: a simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava N","year":"2014","journal-title":"J Mach Learn Res"},{"key":"e_1_2_10_53_1","unstructured":"OktayO SchlemperJ FolgocL et\u00a0al.Attention U\u2010Net: learning where to look for the pancreas.2018:10. arXiv preprint arXiv:1804.03999."},{"key":"e_1_2_10_54_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS SunJ.Delving deep into rectifiers: surpassing human\u2010level performance on ImageNet classification. In:ICCV.2015:1026\u20101034.","DOI":"10.1109\/ICCV.2015.123"},{"key":"e_1_2_10_55_1","unstructured":"KingmaD BaJ.Adam: a method for stochastic optimization. In:Proc. Int Conf Learning Representations.2014:1\u201013."},{"key":"e_1_2_10_56_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0360-3016(02)04617-5"},{"key":"e_1_2_10_57_1","doi-asserted-by":"publisher","DOI":"10.1016\/0360-3016(93)90548-A"}],"container-title":["Journal of Applied Clinical Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.13630","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/acm2.13630","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.13630","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,23]],"date-time":"2023-08-23T21:15:31Z","timestamp":1692825331000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/acm2.13630"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,5,9]]},"references-count":56,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2022,7]]}},"alternative-id":["10.1002\/acm2.13630"],"URL":"https:\/\/doi.org\/10.1002\/acm2.13630","archive":["Portico"],"relation":{},"ISSN":["1526-9914","1526-9914"],"issn-type":[{"value":"1526-9914","type":"print"},{"value":"1526-9914","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,5,9]]},"assertion":[{"value":"2022-01-17","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-04-20","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-05-09","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e13630"}}