{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T11:15:12Z","timestamp":1784805312547,"version":"3.55.0"},"reference-count":35,"publisher":"Public Library of Science (PLoS)","issue":"5","license":[{"start":{"date-parts":[[2023,5,9]],"date-time":"2023-05-09T00:00:00Z","timestamp":1683590400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100010564","name":"German Center for Lung Research","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100010564","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS ONE"],"abstract":"<jats:sec id=\"sec001\">\n<jats:title>Purpose<\/jats:title>\n<jats:p>To improve automated lung segmentation on 2D lung MR images using balanced augmentation and artificially-generated consolidations for training of a convolutional neural network (CNN).<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec002\">\n<jats:title>Materials and methods<\/jats:title>\n<jats:p>From 233 healthy volunteers and 100 patients, 1891 coronal MR images were acquired. Of these, 1666 images without consolidations were used to build a binary semantic CNN for lung segmentation and 225 images (187 without consolidations, 38 with consolidations) were used for testing. To increase CNN performance of segmenting lung parenchyma with consolidations, balanced augmentation was performed and artificially-generated consolidations were added to all training images. The proposed CNN (CNN<jats:sub>Bal\/Cons<\/jats:sub>) was compared to two other CNNs: CNN<jats:sub>Unbal\/NoCons<\/jats:sub>\u2014without balanced augmentation and artificially-generated consolidations and CNN<jats:sub>Bal\/NoCons<\/jats:sub>\u2014with balanced augmentation but without artificially-generated consolidations. Segmentation results were assessed using S\u00f8rensen-Dice coefficient (<jats:italic>SDC<\/jats:italic>) and Hausdorff distance coefficient.<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec003\">\n<jats:title>Results<\/jats:title>\n<jats:p>Regarding the 187 MR test images without consolidations, the mean <jats:italic>SDC<\/jats:italic> of CNN<jats:sub>Unbal\/NoCons<\/jats:sub> (92.1 \u00b1 6% (mean \u00b1 standard deviation)) was significantly lower compared to CNN<jats:sub>Bal\/NoCons<\/jats:sub> (94.0 \u00b1 5.3%, <jats:italic>P<\/jats:italic> = 0.0013) and CNN<jats:sub>Bal\/Cons<\/jats:sub> (94.3 \u00b1 4.1%, <jats:italic>P<\/jats:italic> = 0.0001). No significant difference was found between <jats:italic>SDC<\/jats:italic> of CNN<jats:sub>Bal\/Cons<\/jats:sub> and CNN<jats:sub>Bal\/NoCons<\/jats:sub> (<jats:italic>P<\/jats:italic> = 0.54).<\/jats:p>\n<jats:p>For the 38 MR test images with consolidations, <jats:italic>SDC<\/jats:italic> of CNN<jats:sub>Unbal\/NoCons<\/jats:sub> (89.0 \u00b1 7.1%) was not significantly different compared to CNN<jats:sub>Bal\/NoCons<\/jats:sub> (90.2 \u00b1 9.4%, <jats:italic>P<\/jats:italic> = 0.53). <jats:italic>SDC<\/jats:italic> of CNN<jats:sub>Bal\/Cons<\/jats:sub> (94.3 \u00b1 3.7%) was significantly higher compared to CNN<jats:sub>Bal\/NoCons<\/jats:sub> (<jats:italic>P<\/jats:italic> = 0.0146) and CNN<jats:sub>Unbal\/NoCons<\/jats:sub> (<jats:italic>P<\/jats:italic> = 0.001).<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec004\">\n<jats:title>Conclusions<\/jats:title>\n<jats:p>Expanding training datasets via balanced augmentation and artificially-generated consolidations improved the accuracy of CNN<jats:sub>Bal\/Cons<\/jats:sub>, especially in datasets with parenchymal consolidations. This is an important step towards a robust automated postprocessing of lung MRI datasets in clinical routine.<\/jats:p>\n<\/jats:sec>","DOI":"10.1371\/journal.pone.0285378","type":"journal-article","created":{"date-parts":[[2023,5,9]],"date-time":"2023-05-09T21:19:30Z","timestamp":1683667170000},"page":"e0285378","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":9,"title":["Artificially-generated consolidations and balanced augmentation increase performance of U-net for lung parenchyma segmentation on MR images"],"prefix":"10.1371","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3006-0229","authenticated-orcid":true,"given":"Cristian","family":"Crisosto","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andreas","family":"Voskrebenzev","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marcel","family":"Gutberlet","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2715-0757","authenticated-orcid":true,"given":"Filip","family":"Klime\u0161","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Till F.","family":"Kaireit","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gesa","family":"P\u00f6hler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0779-0688","authenticated-orcid":true,"given":"Tawfik","family":"Moher","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lea","family":"Behrendt","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2137-7355","authenticated-orcid":true,"given":"Robin","family":"M\u00fcller","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maximilian","family":"Zubke","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frank","family":"Wacker","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jens","family":"Vogel-Claussen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2023,5,9]]},"reference":[{"key":"pone.0285378.ref001","doi-asserted-by":"crossref","first-page":"2770","DOI":"10.1007\/s00330-018-5863-7","article-title":"Technical challenges of quantitative chest MRI data analysis in a large cohort pediatric study","volume":"29","author":"A. H. Nguyen","year":"2019","journal-title":"Eur. Radiol"},{"key":"pone.0285378.ref002","doi-asserted-by":"crossref","first-page":"656","DOI":"10.1002\/mrm.22031","article-title":"Non-contrast-enhanced perfusion and ventilation assessment of the human lung by means of fourier decomposition in proton MRI","volume":"62","author":"G. Bauman","year":"2009","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref003","doi-asserted-by":"crossref","first-page":"2306","DOI":"10.1002\/mrm.26893","article-title":"Feasibility of quantitative regional ventilation and perfusion mapping with phase\u2010resolved functional lung (PREFUL) MRI in healthy volunteers and COPD, CTEPH, and CF patients","volume":"79","author":"A. Voskrebenzev","year":"2018","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref004","first-page":"1","article-title":"Very deep convolutional networks for large-scale image recognition","author":"K. Simonyan","year":"2015","journal-title":"3rd Int. Conf. Learn. Represent. ICLR 2015\u2014Conf. Track Proc"},{"key":"pone.0285378.ref005","doi-asserted-by":"crossref","unstructured":"Willers, C. C. et al. Inter-reader variation in lung segmentation of functional lung MRI quantification. in Cystic fibrosis PA333 (European Respiratory Society, 2019). doi:10.1183\/13993003.congress-2019.PA333.","DOI":"10.1183\/13993003.congress-2019.PA333"},{"key":"pone.0285378.ref006","doi-asserted-by":"crossref","first-page":"915","DOI":"10.1002\/jmri.27385","article-title":"Repeatability of Phase\u2010Resolved Functional Lung (PREFUL)\u2010 MRI Ventilation and Perfusion Parameters in Healthy Subjects and COPD Patients","volume":"53","author":"G. H. P\u00f6hler","year":"2021","journal-title":"J. Magn. Reson. Imaging"},{"key":"pone.0285378.ref007","doi-asserted-by":"crossref","first-page":"912","DOI":"10.1002\/mrm.28482","article-title":"3D phase\u2010resolved functional lung ventilation MR imaging in healthy volunteers and patients with chronic pulmonary disease","volume":"85","author":"F. Klime\u0161","year":"2021","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref008","article-title":"VISCERAL Anatomy3 Organ Segmentation Challenge","volume":"18","author":"O. Goksel","year":"2015","journal-title":"IEEE Int. Symp. Biomed. Imaging"},{"key":"pone.0285378.ref009","article-title":"Automatic lung segmentation in routine imaging is a data diversity problem, not a methodology problem","volume":"2","author":"J. Hofmanninger","year":"2020","journal-title":"arXiv"},{"issue":"2021","key":"pone.0285378.ref010","volume":"2021","author":"S. Tello-mijares","journal-title":"Computed Tomography Image Processing Analysis in COVID-19 Patient Follow-Up Assessment"},{"key":"pone.0285378.ref011","doi-asserted-by":"crossref","first-page":"1079","DOI":"10.1002\/mrm.28476","article-title":"The impact of segmentation on whole-lung functional MRI quantification: Repeatability and reproducibility from multiple human observers and an artificial neural network","volume":"85","author":"C. Willers","year":"2021","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref012","author":"V. Thambawita","year":"2021","journal-title":"SinGAN-Seg: Synthetic Training Data Generation for Medical Image Segmentation"},{"key":"pone.0285378.ref013","doi-asserted-by":"crossref","first-page":"781","DOI":"10.1109\/TMI.2017.2759102","article-title":"End-to-End Adversarial Retinal Image Synthesis","volume":"37","author":"P. Costa","year":"2018","journal-title":"IEEE Trans. Med. Imaging"},{"key":"pone.0285378.ref014","author":"L. Perez","year":"2017","journal-title":"The Effectiveness of Data Augmentation in Image Classification using Deep Learning"},{"key":"pone.0285378.ref015","first-page":"117","article-title":"Data augmentation for improving deep learning in image classification problem","author":"IEEE","year":"2018","journal-title":"in 2018 International Interdisciplinary PhD Workshop (IIPhDW)"},{"key":"pone.0285378.ref016","first-page":"1","author":"G. Mariani","year":"2018","journal-title":"BAGAN: Data Augmentation with Balancing GAN"},{"key":"pone.0285378.ref017","author":"J. I. E Gao","year":"2020","journal-title":"Data Augmentation in Solving Data Imbalance Problems"},{"key":"pone.0285378.ref018","doi-asserted-by":"crossref","first-page":"1086","DOI":"10.1164\/rccm.201805-0995OC","article-title":"Effect of Indacaterol\/Glycopyrronium on Pulmonary Perfusion and Ventilation in Hyperinflated Patients with Chronic Obstructive Pulmonary Disease (CLAIM). A Double-Blind, Randomized, Crossover Trial","volume":"199","author":"J. Vogel-Claussen","year":"2019","journal-title":"Am. J. Respir. Crit. Care Med"},{"key":"pone.0285378.ref019","doi-asserted-by":"crossref","first-page":"1045","DOI":"10.1002\/mrm.27991","article-title":"Effect of intravenously injected gadolinium-based contrast agents on functional lung parameters derived by PREFUL MRI","volume":"83","author":"J. Glandorf","year":"2020","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref020","doi-asserted-by":"crossref","first-page":"103","DOI":"10.1002\/jmri.27027","article-title":"Validation of Automated Perfusion-Weighted Phase-Resolved Functional Lung (PREFUL)-MRI in Patients With Pulmonary Diseases","volume":"52","author":"L. Behrendt","year":"2020","journal-title":"J. Magn. Reson. Imaging"},{"key":"pone.0285378.ref021","doi-asserted-by":"crossref","first-page":"2033","DOI":"10.1016\/j.neuroimage.2010.09.025","article-title":"A reproducible evaluation of ANTs similarity metric performance in brain image registration","volume":"54","author":"B. B. Avants","year":"2011","journal-title":"Neuroimage"},{"key":"pone.0285378.ref022","doi-asserted-by":"crossref","first-page":"697","DOI":"10.1148\/radiol.2462070712","article-title":"Glossary of terms for thoracic imaging","volume":"246","author":"Fleischner Society","year":"2008","journal-title":"Radiology"},{"key":"pone.0285378.ref023","first-page":"234","author":"O. Ronneberger","year":"2015","journal-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation"},{"key":"pone.0285378.ref024","author":"M. Abadi","year":"2016","journal-title":"TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems"},{"key":"pone.0285378.ref025","doi-asserted-by":"crossref","first-page":"1285","DOI":"10.1109\/TMI.2016.2528162","article-title":"Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning","volume":"35","author":"H.-C. Shin","year":"2016","journal-title":"IEEE Trans. Med. Imaging"},{"key":"pone.0285378.ref026","first-page":"1","article-title":"Adam: A method for stochastic optimization","author":"D. P. Kingma","year":"2015","journal-title":"3rd Int. Conf. Learn. Represent. ICLR 2015\u2014Conf. Track Proc"},{"key":"pone.0285378.ref027","doi-asserted-by":"crossref","unstructured":"Biederer, J. et al. MRI of the lung (2\/3). Why \u2026 when \u2026 how? Insights Imaging 3, 355\u2013371 (2012).","DOI":"10.1007\/s13244-011-0146-8"},{"key":"pone.0285378.ref028","doi-asserted-by":"crossref","first-page":"412","DOI":"10.1016\/j.acra.2018.08.003","article-title":"Convolutional Neural Networks with Template-Based Data Augmentation for Functional Lung Image Quantification","volume":"26","author":"N. J. Tustison","year":"2019","journal-title":"Acad. Radiol"},{"key":"pone.0285378.ref029","doi-asserted-by":"crossref","first-page":"2135","DOI":"10.1002\/mrm.27527","article-title":"A framework for Fourier\u2010decomposition free\u2010breathing pulmonary 1 H MRI ventilation measurements","volume":"81","author":"F. Guo","year":"2019","journal-title":"Magn. Reson. Med"},{"key":"pone.0285378.ref030","unstructured":"Pal et al. 2021, Computer Science, https:\/\/doi.org\/10.48550\/arXiv.2103.01532, https:\/\/arxiv.org\/pdf\/2103.01532.pdf"},{"key":"pone.0285378.ref031","unstructured":"Han et al. 2018, International Symposium on Biomedical Imaging, 10.1109\/ISBI.2018.8363678, http:\/\/www.nlab.ci.i.u-tokyo.ac.jp\/pdf\/isbi2018.pdf)."},{"key":"pone.0285378.ref032","unstructured":"Shin, H. C. et al. Medical image synthesis for data augmentation and anonymization using generative adversarial networks. Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics) 11037 LNCS, 1\u201311 (2018)."},{"key":"pone.0285378.ref033","doi-asserted-by":"crossref","first-page":"571","DOI":"10.1002\/jmri.26853","article-title":"Deep semantic lung segmentation for tracking potential pulmonary perfusion biomarkers in chronic obstructive pulmonary disease (COPD): The multi\u2010ethnic study of atherosclerosis COPD study","volume":"51","author":"H. B. Winther","year":"2020","journal-title":"J. Magn. Reson. Imaging"},{"key":"pone.0285378.ref034","unstructured":"Pusterla et al. 2022, Magnetic Resonance in Medicine,https:\/\/doi.org\/10.1002\/mrm.29184, https:\/\/onlinelibrary.wiley.com\/doi\/full\/10.1002\/mrm.29184"},{"key":"pone.0285378.ref035","unstructured":"Klime\u0161 et al. 2022, NMR in Biomedicine, https:\/\/doi.org\/10.1002\/nbm.4860, https:\/\/analyticalsciencejournals.onlinelibrary.wiley.com\/doi\/10.1002\/nbm.4860"}],"container-title":["PLOS ONE"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0285378","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,5,9]],"date-time":"2023-05-09T21:19:51Z","timestamp":1683667191000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0285378"}},"subtitle":[],"editor":[{"given":"Kumaradevan","family":"Punithakumar","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2023,5,9]]},"references-count":35,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2023,5,9]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0285378","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,5,9]]}}}