{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,6]],"date-time":"2026-04-06T18:32:28Z","timestamp":1775500348155,"version":"3.50.1"},"reference-count":44,"publisher":"Wiley","issue":"5","license":[{"start":{"date-parts":[[2019,3,12]],"date-time":"2019-03-12T00:00:00Z","timestamp":1552348800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U01 CA195599"],"award-info":[{"award-number":["U01 CA195599"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["81830052"],"award-info":[{"award-number":["81830052"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["11401601"],"award-info":[{"award-number":["11401601"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2019,5]]},"abstract":"Objectives<\/jats:title>The aim of this study was to develop a fully automated deep learning approach for identification of the pectoral muscle on mediolateral oblique (MLO<\/jats:styled-content>) view mammograms and evaluate its performance in comparison to our previously developed texture\u2010field orientation (TFO<\/jats:styled-content>) method using conventional image feature analysis. Pectoral muscle segmentation is an important step for automated image analyses such as breast density or parenchymal pattern classification, lesion detection, and multiview correlation.<\/jats:p><\/jats:sec>Materials and methods<\/jats:title>Institutional Review Board (IRB<\/jats:styled-content>) approval was obtained before data collection. A dataset of 729 MLO<\/jats:styled-content>\u2010view mammograms including 637 digitized film mammograms (DFM<\/jats:styled-content>) and 92 digital mammograms (DM<\/jats:styled-content>) from our previous study was used for the training and validation of our deep convolutional neural network (DCNN<\/jats:styled-content>) segmentation method. In addition, we collected an independent set of 203 DM<\/jats:styled-content>s from 131 patients for testing. The film mammograms were digitized at a pixel size of 50\u00a0\u03bcm\u00a0\u00d7\u00a050\u00a0\u03bcm with a Lumiscan digitizer. All DM<\/jats:styled-content>s were acquired with GE<\/jats:styled-content> systems at a pixel size of 100\u00a0\u03bcm\u00a0\u00d7\u00a0100\u00a0\u03bcm. An experienced MQSA<\/jats:styled-content> radiologist manually drew the pectoral muscle boundary on each mammogram as the reference standard. We trained the DCNN<\/jats:styled-content> to estimate a probability map of the pectoral muscle region on mammograms. The DCNN<\/jats:styled-content> consisted of a contracting path to capture multiresolution image context and a symmetric expanding path for prediction of the pectoral muscle region. Three DCNN<\/jats:styled-content> structures were compared for automated identification of pectoral muscles. Tenfold cross\u2010validation was used in training of the DCNN<\/jats:styled-content>s. After training, we applied the ten trained models during cross\u2010validation to the independent DM<\/jats:styled-content> test set. The predicted pectoral muscle region of each test DM<\/jats:styled-content> was obtained as the mean probability map by averaging the ensemble of probability maps from the ten models. The DCNN<\/jats:styled-content>\u2010segmented pectoral muscle was evaluated by three performance measures relative to the reference standard: (a) the percent overlap area (POA<\/jats:styled-content>) of the pectoral muscle regions, (b) the Hausdorff distance (Hdist), and (c) the average Euclidean distance (AvgDist) between the boundaries. The results were compared to those obtained with the TFO<\/jats:styled-content> method, used as our baseline. A two\u2010tailed paired t<\/jats:italic> test was performed to examine the significance in the differences between the DCNN<\/jats:styled-content> and the baseline.<\/jats:p><\/jats:sec>Results<\/jats:title>In the ten test partitions of the cross\u2010validation set, the DCNN<\/jats:styled-content> achieved a mean POA<\/jats:styled-content> of 96.5\u00a0\u00b1\u00a02.9%, a mean Hdist of 2.26\u00a0\u00b1\u00a01.31\u00a0mm, and a mean AvgDist of 0.78\u00a0\u00b1\u00a00.58\u00a0mm, while the corresponding measures by the baseline method were 94.2\u00a0\u00b1\u00a04.8%, 3.69\u00a0\u00b1\u00a02.48\u00a0mm, and 1.30\u00a0\u00b1\u00a01.22\u00a0mm, respectively. For the independent DM<\/jats:styled-content> test set, the DCNN<\/jats:styled-content> achieved a mean POA<\/jats:styled-content> of 93.7%\u00a0\u00b1\u00a06.9%, a mean Hdist of 3.80\u00a0\u00b1\u00a03.21\u00a0mm, and a mean AvgDist of 1.49\u00a0\u00b1\u00a01.62\u00a0mm comparing to 86.9%\u00a0\u00b1\u00a016.0%, 7.18\u00a0\u00b1\u00a014.22\u00a0mm, and 3.98\u00a0\u00b1\u00a014.13\u00a0mm, respectively, by the baseline method.<\/jats:p><\/jats:sec>Conclusion<\/jats:title>In comparison to the TFO<\/jats:styled-content> method, DCNN<\/jats:styled-content> significantly improved the accuracy of pectoral muscle identification on mammograms (P\u00a0<\u00a00.05<\/jats:italic>).<\/jats:p><\/jats:sec>","DOI":"10.1002\/mp.13451","type":"journal-article","created":{"date-parts":[[2019,2,16]],"date-time":"2019-02-16T13:49:28Z","timestamp":1550324968000},"page":"2103-2114","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":16,"title":["Automated pectoral muscle identification on MLO<\/scp>\u2010view mammograms: Comparison of deep neural network to conventional computer vision"],"prefix":"10.1002","volume":"46","author":[{"given":"Xiangyuan","family":"Ma","sequence":"first","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"},{"name":"School of Data and Computer Science Sun Yat\u2010Sen University Guangzhou 510275 P.R. China"},{"name":"Guangdong Province Key Laboratory Computational Science Sun Yat\u2010Sen University Guangzhou 510275 P.R. China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jun","family":"Wei","sequence":"additional","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chuan","family":"Zhou","sequence":"additional","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mark A.","family":"Helvie","sequence":"additional","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Heang\u2010Ping","family":"Chan","sequence":"additional","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lubomir M.","family":"Hadjiiski","sequence":"additional","affiliation":[{"name":"Department of Radiology University of Michigan Ann Arbor MI USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yao","family":"Lu","sequence":"additional","affiliation":[{"name":"School of Data and Computer Science Sun Yat\u2010Sen University Guangzhou 510275 P.R. China"},{"name":"Guangdong Province Key Laboratory Computational Science Sun Yat\u2010Sen University Guangzhou 510275 P.R. China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2019,3,12]]},"reference":[{"key":"e_1_2_6_2_1","first-page":"7","article-title":"Cancer statistics, 2017","volume":"67","author":"Siegel RL","year":"2017","journal-title":"CA"},{"key":"e_1_2_6_3_1","volume-title":"Breast cancer, diagnosis and treatment","author":"Zuckerman HC","year":"1987"},{"key":"e_1_2_6_4_1","doi-asserted-by":"publisher","DOI":"10.1158\/1055-9965.EPI-06-0034"},{"key":"e_1_2_6_5_1","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMoa062790"},{"key":"e_1_2_6_6_1","doi-asserted-by":"publisher","DOI":"10.1186\/bcr2942"},{"key":"e_1_2_6_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2012.10.020"},{"key":"e_1_2_6_8_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.598272"},{"key":"e_1_2_6_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/TITB.2009.2033269"},{"key":"e_1_2_6_10_1","doi-asserted-by":"crossref","unstructured":"MirzaalianH AhmadzadehMR SadriS.Pectoral Muscle Segmentation on Digital Mammograms by Nonlinear Diffusion Filtering. Paper presented at: 2007 IEEE Pacific Rim Conference on Communications Computers and Signal Processing; 22\u201024 Aug. 2007.","DOI":"10.1109\/PACRIM.2007.4313303"},{"key":"e_1_2_6_11_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/43\/2\/011"},{"key":"e_1_2_6_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2006.12.011"},{"key":"e_1_2_6_13_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3395576"},{"key":"e_1_2_6_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/TBME.2017.2649481"},{"key":"e_1_2_6_15_1","doi-asserted-by":"publisher","DOI":"10.1109\/TSMC.1983.6313076"},{"key":"e_1_2_6_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(88)90014-7"},{"key":"e_1_2_6_17_1","doi-asserted-by":"publisher","DOI":"10.1117\/12.154572"},{"key":"e_1_2_6_18_1","first-page":"318","article-title":"Recognition of mammographic microcalcifications with artificial neural network","volume":"189","author":"Chan HP","year":"1993","journal-title":"Radiology"},{"key":"e_1_2_6_19_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.597354"},{"key":"e_1_2_6_20_1","doi-asserted-by":"publisher","DOI":"10.1109\/42.476112"},{"key":"e_1_2_6_21_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.597428"},{"key":"e_1_2_6_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(95)00061-5"},{"key":"e_1_2_6_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/42.538937"},{"key":"e_1_2_6_24_1","unstructured":"KrizhevskyA SutskeverI HintonGE.ImageNet classification with deep convolutional neural networks.Proceedings of the 25th International Conference on Neural Information Processing Systems (Advances in Neural Information Processing Systems 25); 2012: 1097\u20101105."},{"key":"e_1_2_6_25_1","doi-asserted-by":"crossref","unstructured":"SzegedyC IoffeS VanhouckeV AlemiAA.Inception\u2010v4 Inception\u2010ResNet and the Impact of Residual Connections on Learning. In:Proceedings of the Thirty\u2010First AAAI Conference on Artificial Intelligence (AAAI\u201017);2017:4278\u20104284.","DOI":"10.1609\/aaai.v31i1.11231"},{"key":"e_1_2_6_26_1","first-page":"7","article-title":"Pectoral muscle segmentation in breast tomosynthesis with deep learning","volume":"10575","author":"Rodriguez\u2010Ruiz A","year":"2018","journal-title":"Proc SPIE"},{"key":"e_1_2_6_27_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4816310"},{"key":"e_1_2_6_28_1","unstructured":"RonnebergerO FischerP BroxT.U\u2010Net: convolutional networks for biomedical image segmentation;2015. arXiv:1505.04597."},{"key":"e_1_2_6_29_1","unstructured":"CholletF.Keras:.https:\/\/github.com\/fchollet\/kerasGitHub repository;2015."},{"key":"e_1_2_6_30_1","unstructured":"NasrGE BadrEA JounC.Cross entropy error function in neural networks: forecasting gasoline demand. In:Proceedings of the Fifteenth International Florida Artificial Intelligence Research Society Conference;2002."},{"key":"e_1_2_6_31_1","doi-asserted-by":"crossref","unstructured":"MilletariF NavabN AhmadiS\u2010A.V\u2010net: fully convolutional neural networks for volumetric medical image segmentation. Paper presented at: 3D Vision (3DV) 2016 Fourth International Conference 2016. IEEE.","DOI":"10.1109\/3DV.2016.79"},{"key":"e_1_2_6_32_1","doi-asserted-by":"crossref","unstructured":"Atiqur RahmanM WangY.Optimizing Intersection\u2010Over\u2010Union in Deep Neural Networks for Image Segmentation. Vol 10072;2016.","DOI":"10.1007\/978-3-319-50835-1_22"},{"key":"e_1_2_6_33_1","unstructured":"PerezL WangJ.The effectiveness of data augmentation in image classification using deep learning. arXiv preprint arXiv:171204621;2017."},{"key":"e_1_2_6_34_1","unstructured":"GhuneimAG.Moore\u2010Neighbor Tracing: _V3\/root_downloads\/tutorials\/contour_tracing_Abeer_George_Ghuneim\/moore.html.http:\/\/www.imageprocessingplace.com\/downloads"},{"key":"e_1_2_6_35_1","doi-asserted-by":"publisher","DOI":"10.1111\/j.1469-8137.1912.tb05611.x"},{"key":"e_1_2_6_36_1","first-page":"117","volume-title":"Variational Analysis","author":"Rockafellar RT","year":"2005"},{"key":"e_1_2_6_37_1","first-page":"170","volume-title":"Elementary Linear Algebra","author":"Anton H","year":"1994"},{"key":"e_1_2_6_38_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4967345"},{"key":"e_1_2_6_39_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/aa93d4"},{"key":"e_1_2_6_40_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2003.823062"},{"key":"e_1_2_6_41_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.1758791"},{"key":"e_1_2_6_42_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278-007-9035-6"},{"key":"e_1_2_6_43_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2016.2572683"},{"key":"e_1_2_6_44_1","doi-asserted-by":"crossref","unstructured":"ZhaoH ShiJ QiX WangX JiaJ.Pyramid scene parsing network. In:2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR);2017:6230\u20106239.","DOI":"10.1109\/CVPR.2017.660"},{"key":"e_1_2_6_45_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2016.2644615"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.wiley.com\/onlinelibrary\/tdm\/v1\/articles\/10.1002%2Fmp.13451","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.13451","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.13451","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.13451","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,9,10]],"date-time":"2023-09-10T22:54:10Z","timestamp":1694386450000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.13451"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2019,3,12]]},"references-count":44,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2019,5]]}},"alternative-id":["10.1002\/mp.13451"],"URL":"https:\/\/doi.org\/10.1002\/mp.13451","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2019,3,12]]},"assertion":[{"value":"2018-06-04","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2019-02-02","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2019-03-12","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}