{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T21:28:05Z","timestamp":1790890085782,"version":"4.1.0"},"reference-count":36,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2022,1,12]],"date-time":"2022-01-12T00:00:00Z","timestamp":1641945600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2022,2]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:sec><jats:title>Background<\/jats:title><jats:p>Intraventricular hemorrhaging (IVH) within cerebral lateral ventricles affects 20\u201330% of very low birth weight infants (&lt;1500\u00a0g). As the ventricles increase in size, the intracranial pressure increases, leading to post\u2010hemorrhagic ventricle dilatation (PHVD), an abnormal enlargement of the head. The most widely used imaging tool for measuring IVH and PHVD is cranial two\u2010dimensional (2D) ultrasound (US). Estimating volumetric changes over time with 2D US is unreliable due to high user variability when locating the same anatomical location at different scanning sessions. Compared to 2D US, three\u2010dimensional (3D) US is more sensitive to volumetric changes in the ventricles and does not suffer from variability in slice acquisition. However, 3D US images require segmentation of the ventricular surface, which is tedious and time\u2010consuming when done manually.<\/jats:p><\/jats:sec><jats:sec><jats:title>Purpose<\/jats:title><jats:p>A fast, automated ventricle segmentation method for 3D US would provide quantitative information in a timely manner when monitoring IVH and PHVD in pre\u2010term neonates. To this end, we developed a fast and fully automated segmentation method to segment neonatal cerebral lateral ventricles from 3D US images using deep learning.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>Our method consists of a 3D U\u2010Net ensemble model composed of three U\u2010Net variants, each highlighting various aspects of the segmentation task such as the shape and boundary of the ventricles. The ensemble is made of a U\u2010Net++, attention U\u2010Net, and U\u2010Net with a deep learning\u2010based shape prior combined using a mean voting strategy. We used a dataset consisting of 190 3D US images, which was separated into two subsets, one set of 87 images contained both ventricles, and one set of 103 images contained only one ventricle (caused by limited field\u2010of\u2010view during acquisition). We conducted fivefold cross\u2010validation to evaluate the performance of the models on a larger amount of test data; 165 test images of which 75 have two ventricles (two\u2010ventricle images) and 90 have one ventricle (one\u2010ventricle images). We compared these results to each stand\u2010alone model and to previous works including, 2D multiplane U\u2010Net and 2D SegNet models.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Using fivefold cross\u2010validation, the ensemble method reported a Dice similarity coefficient (DSC) of 0.720 \u00b1 0.074, absolute volumetric difference (VD) of 3.7 \u00b1 4.1 cm<jats:sup>3<\/jats:sup>, and a mean absolute surface distance (MAD) of 1.14 \u00b1 0.41\u00a0mm on 75 two\u2010ventricle test images. Using 90 test images with a single ventricle, the model after cross\u2010validation reported DSC, VD, and MAD values of 0.806 \u00b1 0.111, 3.5 \u00b1 2.9 cm<jats:sup>3<\/jats:sup>, and 1.37 \u00b1 1.70\u00a0mm, respectively. Compared to alternatives, the proposed ensemble yielded a higher accuracy in segmentation on both test data sets. Our method required approximately 5 s to segment one image and was substantially faster than the state\u2010of\u2010the\u2010art conventional methods.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>Compared to the state\u2010of\u2010the\u2010art non\u2010deep learning methods, our method based on deep learning was more efficient in segmenting neonatal cerebral lateral ventricles from 3D US images with comparable or better DSC, VD, and MAD performance. Our dataset was the largest to date (190 images) for this segmentation problem and the first to segment images that show only one lateral cerebral ventricle.<\/jats:p><\/jats:sec>","DOI":"10.1002\/mp.15432","type":"journal-article","created":{"date-parts":[[2021,12,27]],"date-time":"2021-12-27T15:42:53Z","timestamp":1640619773000},"page":"1034-1046","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":22,"title":["Automated 3D U\u2010net based segmentation of neonatal cerebral ventricles from 3D ultrasound images"],"prefix":"10.1002","volume":"49","author":[{"given":"Zachary","family":"Szentimrey","sequence":"first","affiliation":[{"name":"School of Engineering University of Guelph  Guelph Ontario Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sandrine","family":"de Ribaupierre","sequence":"additional","affiliation":[{"name":"Department of Clinical Neurological Sciences Schulich School of Medicine and Dentistry University of Western Ontario  London Ontario Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aaron","family":"Fenster","sequence":"additional","affiliation":[{"name":"Robarts Research Institute University of Western Ontario  London Ontario Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eranga","family":"Ukwatta","sequence":"additional","affiliation":[{"name":"School of Engineering University of Guelph  Guelph Ontario Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2022,1,12]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/58\/21\/7513"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1067\/mpd.2001.111822"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.1117\/1.JMI.5.2.026001"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1542\/peds.2007-0423"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.1136\/archdischild-2012-302323"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.pediatrneurol.2019.09.007"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ultrasmedbio.2004.07.017"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1136\/fn.82.3.F218"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1117\/1.JMI.3.4.046003"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1111\/j.1651-2227.2010.01830.x"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ultrasmedbio.2015.11.010"},{"issue":"3","key":"e_1_2_9_13_1","first-page":"559","article-title":"Lateral ventricle segmentation of 3D pre\u2010term neonates US using convex optimization","volume":"16","author":"Qiu W","year":"2013","journal-title":"Med Image Comput Comput Interv"},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/ISBI.2014.6868096"},{"key":"e_1_2_9_15_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ultrasmedbio.2014.09.019"},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2016.06.038"},{"issue":"567","key":"e_1_2_9_17_1","first-page":"1","article-title":"Automatic segmentation of ventricular volume by 3D ultrasonography in post haemorrhagic ventricular dilatation among preterm infants","volume":"11","author":"Gontard LC","year":"2021","journal-title":"Sci Rep"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2016.2644615"},{"key":"e_1_2_9_19_1","doi-asserted-by":"crossref","unstructured":"MartinM SciollaB SdikaM WangX QuetinP DelachartreP.Automatic segmentation of the cerebral ventricle in neonates using deep learning with 3D reconstructed freehand ultrasound imaging.IEEE International Ultrasonics Symposium.2018:1\u20134.","DOI":"10.1109\/ULTSYM.2018.8580214"},{"key":"e_1_2_9_20_1","first-page":"1","article-title":"Automatic deep learning\u2010based segmentation of neonatal cerebral ventricles from 3D ultrasound images","author":"Szentimrey Z","year":"2021","journal-title":"Proceedings of SPIE 11600, Medical Imaging 2021 Biomedical Application in Molecular, Structural and Functional Imaging"},{"key":"e_1_2_9_21_1","first-page":"1","article-title":"Automated registration and stitching of multiple 3D ultrasound images for monitoring neonatal intraventricular hemorrhage","author":"Harris A","year":"2018","journal-title":"SPIE 10580, Medical Imaging 2018 Ultrasonics Imaging and Tomography"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1002\/widm.1249"},{"key":"e_1_2_9_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2917620"},{"key":"e_1_2_9_24_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_9_25_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46723-8_49"},{"key":"e_1_2_9_26_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-00889-5_1"},{"key":"e_1_2_9_27_1","first-page":"110","article-title":"Comparison of UNet architectures for segmentation of the left ventricle endocardial border on two\u2010dimensional ultrasound images","author":"Zyuzin V","year":"2019","journal-title":"ral Symposium on Biomedical Engineering, Radioelectronics and Information Technology"},{"key":"e_1_2_9_28_1","first-page":"1","article-title":"Attention U\u2010Net: learning where to look for the pancreas","author":"Oktay O","year":"2018","journal-title":"1st Conference on Medical Imaging with Deep Learning"},{"key":"e_1_2_9_29_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-66182-7_24"},{"key":"e_1_2_9_30_1","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.123"},{"key":"e_1_2_9_31_1","first-page":"448","article-title":"Batch normalization: accelerating deep network training by reducing internal covariate shift","volume":"37","author":"Ioffe S","year":"2015","journal-title":"Proceedings of 32nd International Conference on Machine Learning PMLR"},{"key":"e_1_2_9_32_1","first-page":"265","article-title":"TensorFlow: a system for large\u2010scale machine learning","author":"Abadi M","year":"2016","journal-title":"Proceeding of 12th USENIX Symposium on Operating Systems and Design Implementation (OSDI \u201916)"},{"key":"e_1_2_9_33_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.mri.2012.05.001"},{"key":"e_1_2_9_34_1","doi-asserted-by":"publisher","DOI":"10.1016\/S1076-6332(03)00671-8"},{"key":"e_1_2_9_35_1","doi-asserted-by":"publisher","DOI":"10.2307\/3001968"},{"key":"e_1_2_9_36_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0022-3476(05)80701-2"},{"key":"e_1_2_9_37_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-24574-4_11"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.15432","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.15432","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.15432","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,9,28]],"date-time":"2023-09-28T15:37:21Z","timestamp":1695915441000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.15432"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,1,12]]},"references-count":36,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2022,2]]}},"alternative-id":["10.1002\/mp.15432"],"URL":"https:\/\/doi.org\/10.1002\/mp.15432","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,1,12]]},"assertion":[{"value":"2021-05-22","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2021-12-09","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-01-12","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}