{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T02:49:46Z","timestamp":1782269386482,"version":"3.54.5"},"reference-count":36,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2021,4,14]],"date-time":"2021-04-14T00:00:00Z","timestamp":1618358400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,4,14]],"date-time":"2021-04-14T00:00:00Z","timestamp":1618358400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"DOI":"10.13039\/100007210","name":"Varian Medical Systems","doi-asserted-by":"crossref","id":[{"id":"10.13039\/100007210","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2021,6,1]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>Monte Carlo (MC) simulation is considered the gold standard method for radiotherapy dose calculation. However, achieving high precision requires a large number of simulation histories, which is time-consuming. The use of computer graphics processing units (GPUs) has greatly accelerated MC simulation and allows dose calculation within a few minutes for a typical radiotherapy treatment plan. However, some clinical applications demand real-time efficiency for MC dose calculation. To tackle this problem, we have developed a real-time, deep learning (DL)-based dose denoiser that can be plugged into a current GPU-based MC dose engine to enable real-time MC dose calculation. We used two different acceleration strategies to achieve this goal: (1) we applied voxel unshuffle and voxel shuffle operators to decrease the input and output sizes without any information loss, and (2) we decoupled the 3D volumetric convolution into a 2D axial convolution and a 1D slice convolution. In addition, we used a weakly supervised learning framework to train the network, which greatly reduces the size of the required training dataset and thus enables fast fine-tuning-based adaptation of the trained model to different radiation beams. Experimental results show that the proposed denoiser can run in as little as 39 ms, which is 11.6 times faster than the baseline model. As a result, the whole MC dose calculation pipeline can be finished within \u223c\u20090.15\u2009s, including both GPU MC dose calculation and DL-based denoising, achieving the real-time efficiency needed for some radiotherapy applications, such as online adaptive radiotherapy.<\/jats:p>","DOI":"10.1088\/2632-2153\/abdbfe","type":"journal-article","created":{"date-parts":[[2021,1,15]],"date-time":"2021-01-15T03:30:53Z","timestamp":1610681453000},"page":"025033","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":33,"title":["Deep dose plugin: towards real-time Monte Carlo dose calculation through a deep learning-based denoising algorithm"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6697-7434","authenticated-orcid":false,"given":"Ti","family":"Bai","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Biling","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9590-0655","authenticated-orcid":false,"given":"Dan","family":"Nguyen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Steve","family":"Jiang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"266","published-online":{"date-parts":[[2021,4,14]]},"reference":[{"key":"mlstabdbfebib2","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2018.00742","article-title":"Cascaded pyramid network for multi-person pose estimation","author":"Chen","year":"2018"},{"key":"mlstabdbfebib3","doi-asserted-by":"publisher","first-page":"1359","DOI":"10.1088\/0031-9155\/37\/6\/012","article-title":"The generation of intensity-modulated fields for conformal radiotherapy by dynamic collimation","volume":"37","author":"Convery","year":"1992","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib4","first-page":"3123","article-title":"BinaryConnect: training deep neural networks with binary weights during propagations","author":"Courbariaux","year":"2015"},{"key":"mlstabdbfebib5","article-title":"Denoising Monte Carlo dose calculations using a deep neural network","author":"Fornander","year":"2019"},{"key":"mlstabdbfebib6","doi-asserted-by":"publisher","first-page":"6287","DOI":"10.1088\/0031-9155\/54\/20\/017","article-title":"GPU-based ultra-fast dose calculation using a finite size pencil beam model","volume":"54","author":"Gu","year":"2009","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib7","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2016.90","article-title":"Deep residual learning for image recognition","author":"He","year":"2016"},{"key":"mlstabdbfebib8","article-title":"MobileNets: efficient convolutional neural networks for mobile vision applications","author":"Howard","year":"2017"},{"key":"mlstabdbfebib9","article-title":"Batch normalization: accelerating deep network training by reducing internal covariate shift","author":"Ioffe","year":"2015"},{"key":"mlstabdbfebib10","doi-asserted-by":"publisher","first-page":"5790","DOI":"10.1002\/mp.13856","article-title":"Mitigating inherent noise in Monte Carlo dose distributions using dilated U-Net","volume":"46","author":"Javaid","year":"2019","journal-title":"Med. Phys."},{"key":"mlstabdbfebib11","doi-asserted-by":"publisher","first-page":"7017","DOI":"10.1088\/0031-9155\/56\/22\/002","article-title":"GPU-based fast Monte Carlo simulation for radiotherapy dose calculation","volume":"56","author":"Jia","year":"2011","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib12","doi-asserted-by":"publisher","first-page":"3077","DOI":"10.1088\/0031-9155\/55\/11\/006","article-title":"Development of a GPU-based Monte Carlo dose calculation code for coupled electron\u2013photon transport","volume":"55","author":"Jia","year":"2010","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib13","doi-asserted-by":"publisher","first-page":"485","DOI":"10.1118\/1.598917","article-title":"Accurate condensed history Monte Carlo simulation of electron transport. i. EGSnrc, the new EGS4 version","volume":"27","author":"Kawrakow","year":"2000","journal-title":"Med. Phys."},{"key":"mlstabdbfebib14","doi-asserted-by":"publisher","first-page":"3087","DOI":"10.1088\/0031-9155\/47\/17\/304","article-title":"On the de-noising of Monte Carlo calculated dose distributions","volume":"47","author":"Kawrakow","year":"2002","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib15","doi-asserted-by":"publisher","first-page":"1315","DOI":"10.1016\/S0360-3016(99)00324-7","article-title":"Intensity modulation with the \u2018step and shoot\u2019 technique using a commercial MLC: a planning study","volume":"45","author":"Keller-Reichenbecher","year":"1999","journal-title":"Int. J. Radiat. Oncol.* Biol.* Phys."},{"key":"mlstabdbfebib16","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2014"},{"key":"mlstabdbfebib17","article-title":"Noise2noise: learning image restoration without clean data","author":"Lehtinen","year":"2018"},{"key":"mlstabdbfebib18","doi-asserted-by":"publisher","first-page":"157","DOI":"10.1007\/s41095-020-0167-7","article-title":"A detail preserving neural network model for Monte Carlo denoising","volume":"6","author":"Lin","year":"2020","journal-title":"Computational Visual Media"},{"key":"mlstabdbfebib19","first-page":"pp 3431","article-title":"Fully convolutional networks for semantic segmentation","author":"Long","year":"2015"},{"key":"mlstabdbfebib20","article-title":"Nvidia 8-bit inference with TensorRT","author":"Migacz","year":"2017"},{"key":"mlstabdbfebib21","doi-asserted-by":"crossref","DOI":"10.2172\/1453993","article-title":"EGS4 code system","author":"Nelson","year":"1985"},{"key":"mlstabdbfebib22","first-page":"pp 8026","article-title":"PyTorch: an imperative style, high-performance deep learning library","author":"Paszke","year":"2019"},{"key":"mlstabdbfebib23","doi-asserted-by":"publisher","first-page":"76680","DOI":"10.1109\/ACCESS.2019.2921013","article-title":"MCDNET \u2013 a denoising convolutional neural network to accelerate Monte Carlo radiation transport simulations: a proof of principle with patient dose from x-ray CT imaging","volume":"7","author":"Peng","year":"2019","journal-title":"IEEE Access"},{"key":"mlstabdbfebib24","first-page":"1","article-title":"Faster R-CNN: towards real-time object detection with region proposal networks","author":"Ren","year":"2016","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"mlstabdbfebib25","doi-asserted-by":"publisher","first-page":"80","DOI":"10.1016\/j.radonc.2012.10.010","article-title":"Adaptive radiotherapy for head and neck cancer\u2013dosimetric results from a prospective clinical trial","volume":"106","author":"Schwartz","year":"2013","journal-title":"Radiother. Oncol."},{"key":"mlstabdbfebib26","doi-asserted-by":"publisher","first-page":"1177","DOI":"10.1016\/j.ijrobp.2009.03.013","article-title":"A comparison of volumetric modulated arc therapy and conventional intensity-modulated radiotherapy for frontal and temporal high-grade gliomas","volume":"76","author":"Shaffer","year":"2010","journal-title":"Int. J. Radiat. Oncol. Biol. Phys."},{"key":"mlstabdbfebib27","first-page":"pp 1874","article-title":"Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network","author":"Shi","year":"2016"},{"key":"mlstabdbfebib28","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014"},{"key":"mlstabdbfebib29","doi-asserted-by":"crossref","DOI":"10.1609\/aaai.v31i1.11231","article-title":"Inception-v4, inception-ResNet and the impact of residual connections on learning","author":"Szegedy","year":"2017"},{"key":"mlstabdbfebib30","article-title":"Instance normalization: the missing ingredient for fast stylization","author":"Ulyanov","year":"2016"},{"key":"mlstabdbfebib31","doi-asserted-by":"publisher","first-page":"678","DOI":"10.1259\/bjr\/65676879","article-title":"The physical basis of IMRT and inverse planning","volume":"76","author":"Webb","year":"2003","journal-title":"Br. J. Radiol."},{"key":"mlstabdbfebib32","doi-asserted-by":"publisher","first-page":"1647","DOI":"10.1118\/1.2721657","article-title":"A fast algorithm for gamma evaluation in 3D","volume":"34","author":"Wendling","year":"2007","journal-title":"Med. Phys."},{"key":"mlstabdbfebib33","first-page":"pp 3","article-title":"Group normalization","author":"Wu","year":"2018"},{"key":"mlstabdbfebib34","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2017.634","article-title":"Aggregated residual transformations for deep neural networks","author":"Xie","year":"2017"},{"key":"mlstabdbfebib35","doi-asserted-by":"publisher","first-page":"1333","DOI":"10.1088\/0031-9155\/48\/10\/307","article-title":"Aperture modulated arc therapy","volume":"48","author":"Xing","year":"2003","journal-title":"Phys. Med. Biol."},{"key":"mlstabdbfebib36","doi-asserted-by":"publisher","first-page":"224","DOI":"10.1145\/3355089.3356547","article-title":"Adversarial Monte Carlo denoising with conditioned auxiliary feature modulation","volume":"38","author":"Xu","year":"2019","journal-title":"ACM Trans. Graph."},{"key":"mlstabdbfebib37","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2018.00716","article-title":"ShuffleNet: an extremely efficient convolutional neural network for mobile","author":"Zhang","year":"2018"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,22]],"date-time":"2024-08-22T11:55:04Z","timestamp":1724327704000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abdbfe"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,4,14]]},"references-count":36,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2021,4,14]]},"published-print":{"date-parts":[[2021,6,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/abdbfe","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,4,14]]},"assertion":[{"value":"Deep dose plugin: towards real-time Monte Carlo dose calculation through a deep learning-based denoising algorithm","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2021 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2020-10-11","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-01-14","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-04-14","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}