{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T11:14:24Z","timestamp":1778238864875,"version":"3.51.4"},"reference-count":37,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T00:00:00Z","timestamp":1690934400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000038","name":"Natural Sciences and Engineering Research Council of Canada","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000038","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100013873","name":"Government of Ontario","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100013873","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000320","name":"Alzheimer's Society","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000320","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neuroinform."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>Acquisition and pre-processing pipelines for diffusion-weighted imaging (DWI) volumes are resource- and time-consuming. Generating synthetic DWI scalar maps from commonly acquired brain MRI sequences such as fluid-attenuated inversion recovery (FLAIR) could be useful for supplementing datasets. In this work we design and compare GAN-based image translation models for generating DWI scalar maps from FLAIR MRI for the first time.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>We evaluate a pix2pix model, two modified CycleGANs using paired and unpaired data, and a convolutional autoencoder in synthesizing DWI fractional anisotropy (FA) and mean diffusivity (MD) from whole FLAIR volumes. In total, 420 FLAIR and DWI volumes (11,957 images) from multi-center dementia and vascular disease cohorts were used for training\/testing. Generated images were evaluated using two groups of metrics: (1) human perception metrics including peak signal-to-noise ratio (PSNR) and structural similarity (SSIM), (2) structural metrics including a newly proposed histogram similarity (Hist-KL) metric and mean squared error (MSE).<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Pix2pix demonstrated the best performance both quantitatively and qualitatively with mean PSNR, SSIM, and MSE metrics of 23.41 dB, 0.8, 0.004, respectively for MD generation, and 24.05 dB, 0.78, 0.004, respectively for FA generation. The new histogram similarity metric demonstrated sensitivity to differences in fine details between generated and real images with mean pix2pix MD and FA Hist-KL metrics of 11.73 and 3.74, respectively. Detailed analysis of clinically relevant regions of white matter (WM) and gray matter (GM) in the pix2pix images also showed strong significant (<jats:italic>p<\/jats:italic> &amp;lt; 0.001) correlations between real and synthetic FA values in both tissue types (<jats:italic>R<\/jats:italic> = 0.714 for GM, <jats:italic>R<\/jats:italic> = 0.877 for WM).<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion\/conclusion<\/jats:title><jats:p>Our results show that pix2pix\u2019s FA and MD models had significantly better structural similarity of tissue structures and fine details than other models, including WM tracts and CSF spaces, between real and generated images. Regional analysis of synthetic volumes showed that synthetic DWI images can not only be used to supplement clinical datasets, but demonstrates potential utility in bypassing or correcting registration in data pre-processing.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fninf.2023.1197330","type":"journal-article","created":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T11:39:33Z","timestamp":1690976373000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":10,"title":["Synthesis of diffusion-weighted MRI scalar maps from FLAIR volumes using generative adversarial networks"],"prefix":"10.3389","volume":"17","author":[{"given":"Karissa","family":"Chan","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Pejman Jabehdar","family":"Maralani","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Alan R.","family":"Moody","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"April","family":"Khademi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2023,8,2]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1016\/j.nicl.2022.102955","article-title":"FLAIR MRI biomarkers of the normal appearing brain matter are related to cognition.","volume":"34","author":"Bahsoun","year":"2022","journal-title":"NeuroImage"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1016\/j.nicl.2023.103385","article-title":"Alzheimer\u2019s and vascular disease classification using regional texture biomarkers in FLAIR MRI.","volume":"38","author":"Chan","year":"2023","journal-title":"Neuroimage Clin."},{"key":"B3","doi-asserted-by":"publisher","first-page":"64747","DOI":"10.1109\/ACCESS.2021.3075608","article-title":"Synthesis of 3D MRI brain images with shape and texture generative adversarial deep neural networks.","volume":"9","author":"Chong","year":"2021","journal-title":"IEEE Access"},{"key":"B4","doi-asserted-by":"publisher","first-page":"313","DOI":"10.1148\/radiol.2021203786","article-title":"Generative adversarial networks to synthesize missing T1 and FLAIR MRI sequences for use in a multisequence brain tumor segmentation model.","volume":"299","author":"Conte","year":"2021","journal-title":"Radiology"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1016\/j.ynirp.2021.100006","article-title":"Intracranial volume segmentation for neurodegenerative populations using multicentre FLAIR MRI.","volume":"1","author":"DiGregorio","year":"2021","journal-title":"Neuroimage"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1856","DOI":"10.2174\/1381612826666201125110710","article-title":"Generative adversarial networks in medical image processing.","volume":"27","author":"Gong","year":"2021","journal-title":"Curr Pharm Des"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-20205-7_40","article-title":"Generating diffusion MRI scalar maps from T1 weighted images using generative adversarial networks","author":"Gu","year":"2019","journal-title":"Image analysis, lecture notes in computer science"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1706.08500","article-title":"GANs trained by a two time-scale update rule converge to a local nash equilibrium.","author":"Heusel","year":"2018","journal-title":"arXiv"},{"key":"B9","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.mri.2021.06.001","article-title":"Realistic generation of diffusion-weighted magnetic resonance brain images with deep generative models.","volume":"81","author":"Hirte","year":"2021","journal-title":"Magn. Reson. Imag."},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR.2010.579","author":"Hor\u00e9","year":"2010","journal-title":"Image quality metrics: PSNR vs. SSIM."},{"key":"B11","doi-asserted-by":"crossref","first-page":"800","DOI":"10.1049\/el:20080522","article-title":"Scope of validity of PSNR in image\/video quality assessment.","volume":"44","author":"Huynh-Thu","year":"2008","journal-title":"Electron. Lett."},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1611.07004","article-title":"Image-to-Image Translation with Conditional Adversarial Networks.","author":"Isola","year":"2018","journal-title":"arXiv"},{"key":"B13","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1007\/s10278-021-00556-w","article-title":"Systematic review of generative adversarial networks (GANs) for medical image classification and segmentation.","volume":"35","author":"Jeong","year":"2022","journal-title":"J. Digit. Imaging"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2020.101938","article-title":"GANs for medical image analysis.","volume":"109","author":"Kazeminia","year":"2020","journal-title":"Artif. Intell. Med."},{"key":"B15","doi-asserted-by":"publisher","first-page":"159","DOI":"10.18383\/j.tom.2018.00042","article-title":"Generative adversarial networks for the creation of realistic artificial brain magnetic resonance images.","volume":"4","author":"Kazuhiro","year":"2018","journal-title":"Tomography"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.3389\/fpubh.2020.00164","article-title":"Generative adversarial networks and its applications in biomedical informatics.","volume":"8","author":"Lan","year":"2020","journal-title":"Front. Public Health"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2009.14075","article-title":"Backpropagating through frechet inception distance.","author":"Mathiasen","year":"2021","journal-title":"arXiv"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.210315","article-title":"RadImageNet: An open radiologic deep learning research dataset for effective transfer learning.","volume":"4","author":"Mei","year":"2022","journal-title":"Radiology"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1802.05957","article-title":"Spectral normalization for generative adversarial networks.","author":"Miyato","year":"2018","journal-title":"arXiv"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.3389\/fninf.2018.00085","article-title":"National neuroinformatics framework for Canadian Consortium on Neurodegeneration in Aging (CCNA).","volume":"12","author":"Mohaddes","year":"2018","journal-title":"Frontiers in Neuroinformatics"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.13530","article-title":"Deep learning-based convolutional neural network for intramodality brain MRI synthesis.","volume":"23","author":"Osman","year":"2022","journal-title":"J. Appl. Clin. Med. Phys."},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1109\/ICIP.2015.7351345","article-title":"Limitations of the SSIM quality metric in the context of diagnostic imaging","author":"Pambrun","year":"2015","journal-title":"Paper Presented at the 2015 IEEE International Conference on Image Processing (ICIP)"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1117\/12.2293515","article-title":"Learning implicit brain MRI manifolds with deep learning","author":"Plassard","year":"2018","journal-title":"Medical Imaging 2018: Image Processing. Presented at the Image Processing"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2021.103283","article-title":"Image translation for medical image generation: Ischemic stroke lesion segmentation.","volume":"72","author":"Platscher","year":"2022","journal-title":"Biomed. Signal Process. Control"},{"key":"B25","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1016\/j.mri.2019.05.001","article-title":"Pathology-preserving intensity standardization framework for multi-institutional FLAIR MRI datasets.","volume":"62","author":"Reiche","year":"2019","journal-title":"Magn. Reson. Imaging"},{"key":"B26","doi-asserted-by":"publisher","first-page":"3023","DOI":"10.1002\/jemt.23861","article-title":"Deep convolutional generative adversarial network for Alzheimer\u2019s disease classification using positron emission tomography (PET) and synthetic data augmentation.","volume":"84","author":"Sajjad","year":"2021","journal-title":"Microsc. Res. Techn."},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.105556","article-title":"From CNNs to GANs for cross-modality medical image estimation.","volume":"146","author":"Shokraei Fard","year":"2022","journal-title":"Comput. Biol. Med."},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.3390\/jimaging9030069","article-title":"GANs for medical image synthesis: An empirical study.","volume":"9","author":"Skandarani","year":"2023","journal-title":"J. Imaging"},{"key":"B29","doi-asserted-by":"publisher","first-page":"297","DOI":"10.1016\/j.cjca.2012.09.017","article-title":"Atherosclerosis imaging and the Canadian Atherosclerosis Imaging Network.","volume":"29","author":"Tardif","year":"2013","journal-title":"Can. J. Cardiol."},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2020.116889","article-title":"TractoFlow: A robust, efficient and reproducible diffusion MRI pipeline leveraging Nextflow & Singularity.","volume":"218","author":"Theaud","year":"2020","journal-title":"NeuroImage"},{"key":"B31","doi-asserted-by":"publisher","first-page":"600","DOI":"10.1109\/TIP.2003.819861","article-title":"Image quality assessment: From error visibility to structural similarity.","volume":"13","author":"Wang","year":"2004","journal-title":"IEEE Trans. Image Process."},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1806.07777","article-title":"Generative adversarial networks for image-to-image translation on multi-contrast MR images \u2013 A Comparison of CycleGAN and UNIT.","author":"Welander","year":"2018","journal-title":"arXiv"},{"key":"B33","author":"Winkler","year":"2009","journal-title":"FLAIR Templates."},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1801.06940","article-title":"MRI cross-modality neuroimage-to-neuroimage translation.","author":"Yang","year":"2018","journal-title":"arXiv"},{"key":"B35","doi-asserted-by":"publisher","first-page":"17","DOI":"10.1109\/JBHI.2021.3088866","article-title":"Multi-Modal MRI image synthesis via GAN with multi-scale gate mergence.","volume":"26","author":"Zhan","year":"2022","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics12040816","article-title":"Multi-Contrast MRI image synthesis using switchable cycle-consistent generative adversarial networks.","volume":"12","author":"Zhang","year":"2022","journal-title":"Diagnostics"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.244","article-title":"Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks","author":"Zhu","year":"2017","journal-title":"Paper Presented at the 2017 IEEE International Conference on Computer Vision (ICCV)"}],"container-title":["Frontiers in Neuroinformatics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2023.1197330\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T11:39:46Z","timestamp":1690976386000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2023.1197330\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,8,2]]},"references-count":37,"alternative-id":["10.3389\/fninf.2023.1197330"],"URL":"https:\/\/doi.org\/10.3389\/fninf.2023.1197330","relation":{},"ISSN":["1662-5196"],"issn-type":[{"value":"1662-5196","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,8,2]]},"article-number":"1197330"}}