{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T13:26:56Z","timestamp":1762954016062,"version":"3.41.2"},"reference-count":43,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,10,22]],"date-time":"2024-10-22T00:00:00Z","timestamp":1729555200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000196","name":"Canada Foundation for Innovation","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000196","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000024","name":"Canadian Institutes of Health Research","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000024","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Neurosci."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>White matter hyperintensities (WMHs) are frequently observed on magnetic resonance (MR) images in older adults, commonly appearing as areas of high signal intensity on fluid-attenuated inversion recovery (FLAIR) MR scans. Elevated WMH volumes are associated with a greater risk of dementia and stroke, even after accounting for vascular risk factors. Manual segmentation, while considered the ground truth, is both labor-intensive and time-consuming, limiting the generation of annotated WMH datasets. Un-annotated data are relatively available; however, the requirement of annotated data poses a challenge for developing supervised machine learning models.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>To address this challenge, we implemented a multi-stage semi-supervised learning (M3SL) approach that first uses un-annotated data segmented by traditional processing methods (\u201cbronze\u201d and \u201csilver\u201d quality data) and then uses a smaller number of \u201cgold\u201d-standard annotations for model refinement. The M3SL approach enabled fine-tuning of the model weights with the gold-standard annotations. This approach was integrated into the training of a U-Net model for WMH segmentation. We used data from three scanner vendors (over more than five scanners) and from both cognitively normal (CN) adult and patients cohorts [with mild cognitive impairment and Alzheimer's disease (AD)].<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>An analysis of WMH segmentation performance across both scanner and clinical stage (CN, MCI, AD) factors was conducted. We compared our results to both conventional and transfer-learning deep learning methods and observed better generalization with M3SL across different datasets. We evaluated several metrics (<jats:italic>F<\/jats:italic>-measure, <jats:italic>IoU<\/jats:italic>, and Hausdorff distance) and found significant improvements with our method compared to conventional (<jats:italic>p<\/jats:italic> &amp;lt; 0.001) and transfer-learning (<jats:italic>p<\/jats:italic> &amp;lt; 0.001).<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion<\/jats:title><jats:p>These findings suggest that automated, non-machine learning, tools have a role in a multi-stage learning framework and can reduce the impact of limited annotated data and, thus, enhance model performance.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fncom.2024.1487877","type":"journal-article","created":{"date-parts":[[2024,10,22]],"date-time":"2024-10-22T04:51:11Z","timestamp":1729572671000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["Multi-stage semi-supervised learning enhances white matter hyperintensity segmentation"],"prefix":"10.3389","volume":"18","author":[{"given":"Kau\u00ea T. N.","family":"Duarte","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Abhijot S.","family":"Sidhu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Murilo C.","family":"Barros","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"David G.","family":"Gobbi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Cheryl R.","family":"McCreary","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Feryal","family":"Saad","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Richard","family":"Camicioli","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Eric E.","family":"Smith","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mariana P.","family":"Bento","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Richard","family":"Frayne","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2024,10,22]]},"reference":[{"key":"B1","first-page":"457","article-title":"\u201cMulti-task attention-based semi-supervised learning for medical image segmentation,\u201d","volume-title":"Medical Image Computing and Computer Assisted Intervention-MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13-17, 2019, Proceedings, Part III 22","author":"Chen","year":"2019"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.3390\/biom11081102","article-title":"Role of white matter hyperintensities and related risk factors in vascular cognitive impairment: a review","author":"Chen","year":"2021","journal-title":"Biomolecules"},{"key":"B3","doi-asserted-by":"crossref","first-page":"248","DOI":"10.1109\/CVPR.2009.5206848","article-title":"\u201cImageNet: a large-scale hierarchical image database,\u201d","volume-title":"2009 IEEE Conference on Computer Vision and Pattern Recognition","author":"Deng","year":"2009"},{"key":"B4","doi-asserted-by":"publisher","first-page":"90","DOI":"10.1016\/j.patrec.2023.07.014","article-title":"Segmenting white matter hyperintensities in brain magnetic resonance images using convolution neural networks","volume":"175","author":"Duarte","year":"2023","journal-title":"Pattern Recognit. Lett"},{"key":"B5","doi-asserted-by":"crossref","DOI":"10.1109\/SIBGRAPI55357.2022.9991752","article-title":"\u201cSegmenting white matter hyperintensity in Alzheimer's disease using U-Net CNNs,\u201d","volume-title":"2022 35th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)","author":"Duarte","year":"2022"},{"key":"B6","article-title":"\u201cChanges in 3D radiomic texture descriptors in Alzheimer's disease stages,\u201d","volume-title":"18th International Symposium on Medical Information Processing and Analysis, Vol. 12567","author":"Duarte","year":"2023"},{"key":"B7","first-page":"553","article-title":"\u201cPredicting the early stages of the Alzheimer's disease via combined brain multi-projections and small datasets,\u201d","volume-title":"Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2019)","author":"Duarte","year":"2019"},{"key":"B8","doi-asserted-by":"crossref","first-page":"180","DOI":"10.1109\/SIBGRAPI51738.2020.00032","article-title":"\u201cDetecting Alzheimer's disease based on structural region analysis using a 3d shape descriptor,\u201d","volume-title":"2020 33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)","author":"Duarte","year":"2020"},{"key":"B9","doi-asserted-by":"publisher","first-page":"602","DOI":"10.1016\/S1474-4422(23)00131-X","article-title":"Neuroimaging standards for research into small vessel disease\u2014advances since 2013","volume":"22","author":"Duering","year":"2023","journal-title":"Lancet Neurol"},{"key":"B10","doi-asserted-by":"publisher","first-page":"341","DOI":"10.1016\/S0896-6273(02)00569-X","article-title":"Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain","volume":"33","author":"Fischl","year":"2002","journal-title":"Neuron"},{"key":"B11","first-page":"E128","article-title":"Cerebra-WML: a rapid workflow for quantification of white matter hyperintensities","volume":"40","author":"Gobbi","year":"2012","journal-title":"Can. Stroke Congress"},{"key":"B12","doi-asserted-by":"publisher","first-page":"918","DOI":"10.1016\/j.nicl.2017.12.022","article-title":"White matter hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks","volume":"17","author":"Guerrero","year":"2018","journal-title":"NeuroImage: Clin"},{"key":"B13","doi-asserted-by":"publisher","first-page":"123052","DOI":"10.1016\/j.eswa.2023.123052","article-title":"Deep semi-supervised learning for medical image segmentation: a review","volume":"245","author":"Han","year":"2024","journal-title":"Expert Syst. Appl"},{"key":"B14","doi-asserted-by":"publisher","first-page":"1344","DOI":"10.1002\/hbm.26109","article-title":"Semisupervised white matter hyperintensities segmentation on MRI","volume":"44","author":"Huang","year":"2023","journal-title":"Hum. Brain Mapp"},{"key":"B15","doi-asserted-by":"publisher","first-page":"685","DOI":"10.1002\/jmri.21049","article-title":"The Alzheimer's disease neuroimaging initiative (ADNI): MRI methods","volume":"27","author":"Jack","year":"2008","journal-title":"J. Magn. Reson. Imaging"},{"key":"B16","doi-asserted-by":"publisher","first-page":"539","DOI":"10.1016\/j.neuroimage.2018.03.050","article-title":"Ubo detector \u2013 a cluster-based, fully automated pipeline for extracting white matter hyperintensities","volume":"174","author":"Jiang","year":"2018","journal-title":"Neuroimage"},{"key":"B17","doi-asserted-by":"publisher","first-page":"510","DOI":"10.1007\/978-3-030-32248-9_57","article-title":"\u201cA soft staple algorithm combined with anatomical knowledge,\u201d","author":"Kats","year":"2019","journal-title":"Medical Image Computing and Computer Assisted Intervention-MICCAI 2019"},{"key":"B18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s12880-022-00793-7","article-title":"Transfer learning for medical image classification: a literature review","volume":"22","author":"Kim","year":"2022","journal-title":"BMC Med. Imaging"},{"key":"B19","doi-asserted-by":"publisher","first-page":"79","DOI":"10.1016\/j.bbe.2021.11.004","article-title":"Transfer learning techniques for medical image analysis: a review","volume":"42","author":"Kora","year":"2022","journal-title":"Biocybern. Biomed. Eng"},{"key":"B20","doi-asserted-by":"publisher","first-page":"302","DOI":"10.1111\/j.1747-4949.2010.00579.x","article-title":"Quantomo: Validation of a computer-assisted methodology for the volumetric analysis of intracerebral haemorrhage","volume":"6","author":"Kosior","year":"2011","journal-title":"Int. J. Stroke"},{"key":"B21","doi-asserted-by":"publisher","first-page":"2556","DOI":"10.1109\/TMI.2019.2905770","article-title":"Standardized assessment of automatic segmentation of white matter hyperintensities and results of the WMH segmentation challenge","volume":"38","author":"Kuijf","year":"2019","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B22","doi-asserted-by":"publisher","first-page":"110825","DOI":"10.1016\/j.brainresbull.2023.110825","article-title":"Automatic segmentation of white matter hyperintensities in t2-flair with aqua: a comparative validation study against conventional methods","volume":"205","author":"Lee","year":"2023","journal-title":"Brain Res. Bull"},{"key":"B23","doi-asserted-by":"publisher","first-page":"872","DOI":"10.1016\/j.scib.2024.01.034","article-title":"Deepwmh: A deep learning tool for accurate white matter hyperintensity segmentation without requiring manual annotations for training","volume":"69","author":"Liu","year":"2024","journal-title":"Sci. Bull"},{"key":"B24","doi-asserted-by":"publisher","first-page":"e038120","DOI":"10.1136\/bmjopen-2020-038120","article-title":"Calgary normative study: design of a prospective longitudinal study to characterise potential quantitative MR biomarkers of neurodegeneration over the adult lifespan","volume":"10","author":"McCreary","year":"2020","journal-title":"BMJ Open"},{"key":"B25","doi-asserted-by":"publisher","first-page":"865920","DOI":"10.3389\/fneur.2022.865920","article-title":"Research progress on MRI for white matter hyperintensity of presumed vascular origin and cognitive impairment","volume":"13","author":"Meng","year":"2022","journal-title":"Front. Neurol"},{"key":"B26","doi-asserted-by":"publisher","first-page":"107904","DOI":"10.1016\/j.cmpb.2023.107904","article-title":"Deep learning-based grading of white matter hyperintensities enables identification of potential markers in multi-sequence MRI data","volume":"243","author":"Mu","year":"2024","journal-title":"Comput. Methods Programs Biomed"},{"key":"B27","doi-asserted-by":"publisher","first-page":"1659","DOI":"10.1212\/01.wnl.0000435291.49598.54","article-title":"Neurovascular decoupling is associated with severity of cerebral amyloid angiopathy","volume":"81","author":"Peca","year":"2013","journal-title":"Neurology"},{"key":"B28","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13195-019-0521-0","article-title":"Independent effects of white matter hyperintensities on cognitive, neuropsychiatric, and functional decline: a longitudinal investigation using the national Alzheimer's coordinating center uniform data set","volume":"11","author":"Puzo","year":"2019","journal-title":"Alzheimers Res. Ther"},{"key":"B29","doi-asserted-by":"publisher","first-page":"720","DOI":"10.3390\/brainsci11060720","article-title":"Semi-supervised learning in medical MRI segmentation: brain tissue with white matter hyperintensity segmentation using flair MRI","volume":"11","author":"Rieu","year":"2021","journal-title":"Brain Sci"},{"key":"B30","doi-asserted-by":"publisher","first-page":"234","DOI":"10.1007\/978-3-319-24574-4_28","article-title":"\u201cU-Net: convolutional networks for biomedical image segmentation,\u201d","author":"Ronneberger","year":"2015","journal-title":"Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015"},{"key":"B31","doi-asserted-by":"publisher","first-page":"5930","DOI":"10.3390\/su15075930","article-title":"A study of CNN and transfer learning in medical imaging: advantages, challenges, future scope","volume":"15","author":"Salehi","year":"2023","journal-title":"Sustainability"},{"key":"B32","doi-asserted-by":"publisher","first-page":"3774","DOI":"10.1016\/j.neuroimage.2011.11.032","article-title":"An automated tool for detection of flair-hyperintense white-matter lesions in multiple sclerosis","volume":"59","author":"Schmidt","year":"2012","journal-title":"Neuroimage"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1409.1556","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014","journal-title":"arXiv"},{"key":"B34","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3233\/JAD-210138","article-title":"Cortical thickness and its association with clinical cognitive and neuroimaging markers in cerebral amyloid angiopathy","volume":"81","author":"Subotic","year":"2021","journal-title":"J. Alzheimers Dis"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1310","DOI":"10.1109\/TMI.2010.2046908","article-title":"N4ITK: improved N3 bias correction. IEEE Trans","volume":"29","author":"Tustison","year":"2010","journal-title":"Med. Imaging"},{"key":"B36","doi-asserted-by":"publisher","first-page":"102357","DOI":"10.1016\/j.nicl.2020.102357","article-title":"Comparison and validation of seven white matter hyperintensities segmentation software in elderly patients","volume":"27","author":"Vanderbecq","year":"2020","journal-title":"NeuroImage Clin"},{"key":"B37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1161\/JAHA.114.001140","article-title":"What are white matter hyperintensities made of?","volume":"4","author":"Wardlaw","year":"2015","journal-title":"J. Am. Heart Assoc"},{"key":"B38","doi-asserted-by":"publisher","first-page":"822","DOI":"10.1016\/S1474-4422(13)70124-8","article-title":"Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration","volume":"12","author":"Wardlaw","year":"2013","journal-title":"Lancet Neurol"},{"key":"B39","doi-asserted-by":"publisher","first-page":"903","DOI":"10.1109\/TMI.2004.828354","article-title":"Simultaneous truth and performance level estimation (staple): an algorithm for the validation of image segmentation","volume":"23","author":"Warfield","year":"2004","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B40","doi-asserted-by":"publisher","first-page":"107043","DOI":"10.1016\/j.ymssp.2020.107043","article-title":"A multi-stage semi-supervised learning approach for intelligent fault diagnosis of rolling bearing using data augmentation and metric learning","volume":"146","author":"Yu","year":"2021","journal-title":"Mech. Syst. Signal Process"},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.1007\/s00234-021-02820-w","article-title":"A deep learning algorithm for white matter hyperintensity lesion detection and segmentation","author":"Zhang","year":"2022","journal-title":"Neuroradiology"},{"key":"B42","doi-asserted-by":"publisher","first-page":"122807","DOI":"10.1016\/j.eswa.2023.122807","article-title":"A comparison review of transfer learning and self-supervised learning: definitions, applications, advantages and limitations","volume":"242","author":"Zhao","year":"2024","journal-title":"Expert Syst. Appl"},{"key":"B43","doi-asserted-by":"publisher","first-page":"915009","DOI":"10.3389\/fnagi.2022.915009","article-title":"Automatic segmentation of white matter hyperintensities in routine clinical brain MRI by 2D VB-Net: a large-scale study","volume":"14","author":"Zhu","year":"2022","journal-title":"Front. Aging Neurosci"}],"container-title":["Frontiers in Computational Neuroscience"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2024.1487877\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,22]],"date-time":"2024-10-22T04:51:15Z","timestamp":1729572675000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2024.1487877\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,22]]},"references-count":43,"alternative-id":["10.3389\/fncom.2024.1487877"],"URL":"https:\/\/doi.org\/10.3389\/fncom.2024.1487877","relation":{},"ISSN":["1662-5188"],"issn-type":[{"type":"electronic","value":"1662-5188"}],"subject":[],"published":{"date-parts":[[2024,10,22]]},"article-number":"1487877"}}