{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T21:37:43Z","timestamp":1792100263774,"version":"4.4.0"},"reference-count":50,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2024,6,3]],"date-time":"2024-06-03T00:00:00Z","timestamp":1717372800000},"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":["Medical Physics"],"published-print":{"date-parts":[[2024,8]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Direction Modulated Brachytherapy (DMBT) enables conformal dose distributions. However, clinicians may face challenges in creating viable treatment plans within a fast\u2010paced clinical setting, especially for a novel technology like DMBT, where cumulative clinical experience is limited. Deep learning\u2010based dose prediction methods have emerged as effective tools for enhancing efficiency.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>To develop a voxel\u2010wise dose prediction model using an attention\u2010gating mechanism and a 3D UNET for cervical cancer high\u2010dose\u2010rate (HDR) brachytherapy treatment planning with DMBT six\u2010groove tandems with ovoids or ring applicators.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      A multi\u2010institutional cohort of 122 retrospective clinical HDR brachytherapy plans treated to a prescription dose in the range of 4.8\u20137.0\u00a0Gy\/fraction was used. A DMBT tandem model was constructed and incorporated onto a research version of BrachyVision Treatment Planning System (BV\u2010TPS) as a 3D solid model applicator and retrospectively re\u2010planned all cases by seasoned experts. Those plans were randomly divided into 64:16:20 as training, validating, and testing cohorts, respectively. Data augmentation was applied to the training and validation sets to increase the size by a factor of 4. An attention\u2010gated 3D UNET architecture model was developed to predict full 3D dose distributions based on high\u2010risk clinical target volume (CTV\n                      <jats:sub>HR<\/jats:sub>\n                      ) and organs at risk (OARs) contour information. The model was trained using the mean absolute error loss function, Adam optimization algorithm, a learning rate of 0.001, 250 epochs, and a batch size of eight. In addition, a baseline UNET model was trained similarly for comparison. The model performance was evaluated on the testing dataset by analyzing the outcomes in terms of mean dose values and derived dose\u2010volume\u2010histogram indices from 3D dose distributions and comparing the generated dose distributions against the ground\u2010truth dose distributions using dose statistics and clinically meaningful dosimetric indices.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The proposed attention\u2010gated 3D UNET model showed competitive accuracy in predicting 3D dose distributions that closely resemble the ground\u2010truth dose distributions. The average values of the mean absolute errors were 1.82\u00a0\u00b1\u00a029.09\u00a0Gy (vs. 6.41\u00a0\u00b1\u00a020.16\u00a0Gy for a baseline UNET) in CTV\n                      <jats:sub>HR<\/jats:sub>\n                      , 0.89\u00a0\u00b1\u00a01.25\u00a0Gy (vs. 0.94\u00a0\u00b1\u00a03.96\u00a0Gy for a baseline UNET) in the bladder, 0.33\u00a0\u00b1\u00a00.67\u00a0Gy (vs. 0.53\u00a0\u00b1\u00a01.66\u00a0Gy for a baseline UNET) in the rectum, and 0.55\u00a0\u00b1\u00a01.57\u00a0Gy (vs. 0.76\u00a0\u00b1\u00a02.89\u00a0Gy for a baseline UNET) in the sigmoid. The results showed that the mean absolute error (MAE) for the bladder, rectum, and sigmoid were 0.22\u00a0\u00b1\u00a01.22\u00a0Gy (3.62%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.015), 0.21\u00a0\u00b1\u00a01.06\u00a0Gy (2.20%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.172), and \u20100.03\u00a0\u00b1\u00a00.54\u00a0Gy (1.13%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.774), respectively. The MAE for\n                      <jats:italic>D<\/jats:italic>\n                      <jats:sub>90<\/jats:sub>\n                      ,\n                      <jats:italic>V<\/jats:italic>\n                      <jats:sub>100%<\/jats:sub>\n                      , and\n                      <jats:italic>V<\/jats:italic>\n                      <jats:sub>150%<\/jats:sub>\n                      of the CTV\n                      <jats:sub>HR<\/jats:sub>\n                      were 0.46\u00a0\u00b1\u00a02.44\u00a0Gy (8.14%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.018), 0.57\u00a0\u00b1\u00a011.25% (5.23%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.283), and \u20100.43\u00a0\u00b1\u00a019.36% (4.62%) (\n                      <jats:italic>p<\/jats:italic>\n                      \u00a0=\u00a00.190), respectively. The proposed model needs less than 5 s to predict a full 3D dose distribution of 64\u00a0\u00d7\u00a064\u00a0\u00d7\u00a064 voxels for any new patient plan, thus making it sufficient for near real\u2010time applications and aiding with decision\u2010making in the clinic.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Attention gated 3D\u2010UNET model demonstrated a capability in predicting voxel\u2010wise dose prediction, in comparison to 3D UNET, for DMBT intracavitary brachytherapy planning. The proposed model could be used to obtain dose distributions for near real\u2010time decision\u2010making before DMBT planning and quality assurance. This will guide future automated planning, making the workflow more efficient and clinically viable.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.17238","type":"journal-article","created":{"date-parts":[[2024,6,3]],"date-time":"2024-06-03T15:48:40Z","timestamp":1717429720000},"page":"5593-5603","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":10,"title":["Attention 3D U\u2010NET for dose distribution prediction of high\u2010dose\u2010rate brachytherapy of cervical cancer: Direction modulated brachytherapy tandem applicator"],"prefix":"10.1002","volume":"51","author":[{"given":"Suman","family":"Gautam","sequence":"first","affiliation":[{"name":"Department of Radiation Oncology Virginia Commonwealth University  Richmond Virginia USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexander F. I.","family":"Osman","sequence":"additional","affiliation":[{"name":"Department of Medical Physics Al\u2010Neelain University  Khartoum Sudan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dylan","family":"Richeson","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Inova Schar Cancer Institute  Fairfax Virginia USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Somayeh","family":"Gholami","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology University of Utah  Salt Lake City Utah USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Binod","family":"Manandhar","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Virginia Commonwealth University  Richmond Virginia USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sharmin","family":"Alam","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Virginia Commonwealth University  Richmond Virginia USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"William Y.","family":"Song","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Virginia Commonwealth University  Richmond Virginia USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2024,6,3]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.3322\/caac.21660"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/S2214-109X(19)30482-6"},{"issue":"1","key":"e_1_2_9_4_1","first-page":"NP","article-title":"Prescribing, recording, and reporting brachytherapy for cancer of the cervix","volume":"13","year":"2013","journal-title":"J ICRU"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2015.03.010"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2015.12.003"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2003.09.001"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3193682"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2012.09.037"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2006.01.036"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2011.04.044"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2013.11.011"},{"key":"e_1_2_9_13_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ac344d"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1200\/JCO.2009.27.0942"},{"key":"e_1_2_9_15_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4933244"},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4828778"},{"key":"e_1_2_9_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2014.02.039"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2016.06.015"},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13030"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15666"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1259\/bjr.20220500"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/0167-8140(88)90167-3"},{"key":"e_1_2_9_23_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/47\/14\/309"},{"key":"e_1_2_9_24_1","doi-asserted-by":"publisher","DOI":"10.5114\/jcb.2018.78832"},{"key":"e_1_2_9_25_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.semradonc.2020.05.009"},{"key":"e_1_2_9_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2020.04.008"},{"key":"e_1_2_9_27_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010018\u201037741\u2010x"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1093\/jrr\/rrz051"},{"key":"e_1_2_9_29_1","doi-asserted-by":"publisher","DOI":"10.1186\/s13014-021-01864-9"},{"key":"e_1_2_9_30_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2017.04.011"},{"key":"e_1_2_9_31_1","unstructured":"RonnebergerO FischerP BroxT.U\u2010NET: convolutional networks for biomedical image segmentation.2015.http:\/\/arxiv.org\/abs\/1505.04597"},{"key":"e_1_2_9_32_1","doi-asserted-by":"crossref","unstructured":"\u00c7i\u00e7ek\u00d6 AbdulkadirA LienkampSS BroxT RonnebergerO 3D U\u2010Net: learning dense volumetric segmentation from sparse annotation. In:Medical Image Computing and Computer\u2010Assisted Intervention\u2013MICCAI 2016: 19th International Conference Athens Greece October 17\u201321 2016 Proceedings Part II 19.Springer;2016.424\u2010432.","DOI":"10.1007\/978-3-319-46723-8_49"},{"key":"e_1_2_9_33_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2021.3086020"},{"key":"e_1_2_9_34_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13630"},{"key":"e_1_2_9_35_1","doi-asserted-by":"publisher","DOI":"10.1109\/TRPMS.2021.3098507"},{"key":"e_1_2_9_36_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14827"},{"key":"e_1_2_9_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2023.06.136"},{"key":"e_1_2_9_38_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3703892"},{"key":"e_1_2_9_39_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2023.06.120"},{"key":"e_1_2_9_40_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2023.06.1810"},{"key":"e_1_2_9_41_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2020.580919"},{"key":"e_1_2_9_42_1","unstructured":"YangS XiaoW ZhangM GuoS ZhaoJ ShenF.Image data augmentation for deep learning: a survey. arXiv Prepr arXiv220408610 2022"},{"key":"e_1_2_9_43_1","doi-asserted-by":"publisher","DOI":"10.3389\/fnhum.2021.765525"},{"key":"e_1_2_9_44_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.brachy.2021.10.003"},{"key":"e_1_2_9_45_1","volume-title":"Deep learning with Python: Develop Deep Learning models on Theano and TensorFlow Using Keras","author":"Brownlee J","year":"2016"},{"key":"e_1_2_9_46_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_9_47_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS SunJDelving deep into rectifiers: surpassing human\u2010level performance on imagenet classification2015.","DOI":"10.1109\/ICCV.2015.123"},{"key":"e_1_2_9_48_1","unstructured":"KingmaDP BaJL.Adam: a method for stochastic optimization. In:3rd international conference learning representation ICLR 2015\u2010conference track proceeding.2015."},{"key":"e_1_2_9_49_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.14015"},{"key":"e_1_2_9_50_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/acecd2"},{"key":"e_1_2_9_51_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.598248"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.17238","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,29]],"date-time":"2024-10-29T15:07:51Z","timestamp":1730214471000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.17238"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,6,3]]},"references-count":50,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2024,8]]}},"alternative-id":["10.1002\/mp.17238"],"URL":"https:\/\/doi.org\/10.1002\/mp.17238","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,6,3]]},"assertion":[{"value":"2023-10-24","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-05-21","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-06-03","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}