{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T06:04:33Z","timestamp":1790661873804,"version":"4.1.0"},"reference-count":31,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Radiol."],"abstract":"<jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>To develop and evaluate a unified framework for multiple sclerosis (MS) white matter lesion segmentation that operates robustly under clinically prevalent missing sequence magnetic resonance imaging conditions.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We propose a multi-scale Gaussian product-of-experts framework (MG-PoE) integrated into nnU-Net, where sequence-specific inputs are independently processed through a shared encoder and parameterized as Gaussian experts before combination through precision-weighted product-of-experts (PoE) fusion at each encoder scale. The unified fusion rule handles different sequence configurations without requiring subset-specific architectures. The framework was systematically evaluated under different sequence configurations on diverse datasets (MSSEG2016, MS3SEG, and MSLesSeg24).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>On the five-sequence MSSEG2016, the proposed framework achieved the best overall performance compared to benchmark methods across all full-sequence and partially observed multisequence settings, with a mean Dice Similarity Coefficient (DSC) of 0.7029, a mean 95th percentile Hausdorff distance (HD95) of 22.4030, and a mean Composite Score (CS) of 0.7235 across the 26 sequence configurations. On MS3SEG, MG-PoE achieved the highest mean DSC and CS across the full sequence and missing-sequence settings, with significant improvements over competing methods, while HD95 performance was generally comparable. In zero-shot evaluation on MSLesSeg24, the method achieved the best overall DSC and HD95 in the full sequence setting (DSC, 0.5566; HD95, 27.3729) and the best mean DSC, HD95 and CS across the missing-sequence settings (DSC, 0.5096; HD95, 30.5054; CS, 0.5811).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>The MG-PoE framework provides a unified approach to missing-sequence MS lesion segmentation, achieving favorable performance on independent datasets and providing initial evidence of cross-dataset robustness.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fradi.2026.1868811","type":"journal-article","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:48:08Z","timestamp":1790660888000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Multi-scale Gaussian product-of-experts for missing sequence multiple sclerosis lesion segmentation"],"prefix":"10.3389","volume":"6","author":[{"given":"Seung Hoon","family":"Lee","sequence":"first","affiliation":[{"name":"Department of Imaging and Interventional Radiology, The Chinese University of Hong Kong","place":["Shatin, Hong Kong SAR, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lun Matthew","family":"Wong","sequence":"additional","affiliation":[{"name":"Department of Imaging and Interventional Radiology, The Chinese University of Hong Kong","place":["Shatin, Hong Kong SAR, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tiffany Y.","family":"So","sequence":"additional","affiliation":[{"name":"Department of Imaging and Interventional Radiology, The Chinese University of Hong Kong","place":["Shatin, Hong Kong SAR, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,29]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1502","DOI":"10.1016\/s0140-6736(08)61620-7","article-title":"Multiple sclerosis","volume":"372","author":"Compston","year":"2008","journal-title":"Lancet"},{"key":"B2","doi-asserted-by":"publisher","first-page":"169","DOI":"10.1056\/NEJMra1401483","article-title":"Multiple sclerosis","volume":"378","author":"Reich","year":"2018","journal-title":"N Engl J Med"},{"key":"B3","doi-asserted-by":"publisher","first-page":"659","DOI":"10.1148\/radiol.11101362","article-title":"MR Imaging of multiple sclerosis","volume":"259","author":"Filippi","year":"2011","journal-title":"Radiology"},{"key":"B4","doi-asserted-by":"publisher","first-page":"653","DOI":"10.1016\/S1474-4422(21)00095-8","article-title":"2021 MAGNIMS\u2013CMSC\u2013NAIMS consensus recommendations on the use of MRI in patients with multiple sclerosis","volume":"20","author":"Wattjes","year":"2021","journal-title":"Lancet Neurol"},{"key":"B5","doi-asserted-by":"publisher","first-page":"2917","DOI":"10.1007\/s00415-020-09930-0","article-title":"Diagnosis and management of multiple sclerosis: MRI in clinical practice","volume":"267","author":"Tomassini","year":"2020","journal-title":"J Neurol"},{"key":"B6","doi-asserted-by":"publisher","first-page":"975862","DOI":"10.3389\/fnins.2022.975862","article-title":"New multiple sclerosis lesion segmentation and detection using pre-activation U-net","volume":"16","author":"Ashtari","year":"2022","journal-title":"Front Neurosci"},{"key":"B7","doi-asserted-by":"publisher","first-page":"1188336","DOI":"10.3389\/fnins.2023.1188336","article-title":"3D U-Net for automated detection of multiple sclerosis lesions: utility of transfer learning from other pathologies","volume":"17","author":"Wahlig","year":"2023","journal-title":"Front Neurosci"},{"key":"B8","first-page":"389","article-title":"MIMoSA: an automated method for intermodal segmentation analysis of multiple sclerosis brain lesions","volume-title":"J Neuroimaging","author":"Valcarcel","year":"2018"},{"key":"B9","first-page":"447","article-title":"Robust multimodal brain tumor segmentation via feature disentanglement and gated fusion","author":"Chen","year":"2019"},{"key":"B10","first-page":"107","article-title":"mmFormer: Multimodal Medical Transformer for Incomplete Multimodal Learning of Brain Tumor Segmentation","author":"Zhang","year":"2022"},{"key":"B11","first-page":"3975","article-title":"RFNet: region-aware fusion network for incomplete multi-modal brain tumor segmentation","author":"Ding","year":"2021"},{"key":"B12","doi-asserted-by":"publisher","first-page":"203","DOI":"10.1038\/s41592-020-01008-z","article-title":"nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation","volume":"18","author":"Isensee","year":"2021","journal-title":"Nat Methods"},{"key":"B13","doi-asserted-by":"publisher","first-page":"2075","DOI":"10.1007\/s10278-023-00860-7","article-title":"Brain tumor segmentation for multi-modal MRI with missing information","volume":"36","author":"Feng","year":"2023","journal-title":"J Digit Imaging"},{"key":"B14","doi-asserted-by":"publisher","first-page":"227","DOI":"10.1016\/S0378-3758(00)00115-4","article-title":"Improving predictive inference under covariate shift by weighting the log-likelihood function","volume":"90","author":"Shimodaira","year":"2000","journal-title":"J Stat Plan Inference"},{"key":"B15","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.1109\/TBME.2021.3117407","article-title":"Domain adaptation for medical image analysis: a survey","volume":"69","author":"Guan","year":"2022","journal-title":"IEEE Trans Biomed Eng"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1692","DOI":"10.1109\/TPAMI.2015.2461544","article-title":"Moddrop: adaptive multi-modal gesture recognition","volume":"38","author":"Neverova","year":"2016","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"B17","first-page":"444","article-title":"Moddrop++: a dynamic filter network with intra-subject co-training for multiple sclerosis lesion segmentation with missing modalities","author":"Liu","year":"2022"},{"key":"B18","first-page":"469","article-title":"Hemis: hetero-modal image segmentation","author":"Havaei","year":"2016"},{"key":"B19","first-page":"34","article-title":"A task-conditional mixture-of-experts model for missing modality segmentation","author":"Novosad","year":"2024"},{"key":"B20","first-page":"1","article-title":"Products of experts","author":"Hinton","year":"1999"},{"key":"B21","first-page":"5580","article-title":"Multimodal generative models for scalable weakly-supervised learning","volume":"31","author":"Wu","year":"2018","journal-title":"Adv Neural Inf Process Syst"},{"key":"B22","article-title":"Auto-encoding variational Bayes","author":"Kingma","year":"2014"},{"key":"B23","first-page":"234","article-title":"U-net: convolutional networks for biomedical image segmentation","author":"Ronneberger","year":"2015"},{"key":"B24","doi-asserted-by":"publisher","first-page":"118589","DOI":"10.1016\/j.neuroimage.2021.118589","article-title":"Multiple sclerosis lesions segmentation from multiple experts: the MICCAI 2016 challenge dataset","volume":"244","author":"Commowick","year":"2021","journal-title":"Neuroimage"},{"key":"B25","doi-asserted-by":"publisher","first-page":"867","DOI":"10.1038\/s41597-026-07184-5","article-title":"A multiple sclerosis MRI dataset with tri-mask annotations for lesion segmentation","volume":"13","author":"Bawil","year":"2026","journal-title":"Sci Data"},{"key":"B26","doi-asserted-by":"publisher","first-page":"920","DOI":"10.1038\/s41597-025-05250-y","article-title":"MSLesseg: baseline and benchmarking of a new multiple sclerosis lesion segmentation dataset","volume":"12","author":"Guarnera","year":"2025","journal-title":"Sci Data"},{"key":"B27","first-page":"8026","article-title":"Pytorch: an imperative style, high-performance deep learning library","author":"Paszke","year":"2019"},{"key":"B28","doi-asserted-by":"publisher","first-page":"29","DOI":"10.1186\/s12880-015-0068-x","article-title":"Metrics for evaluating 3D medical image segmentation: analysis, selection, and tool","volume":"15","author":"Taha","year":"2015","journal-title":"BMC Med Imaging"},{"key":"B29","doi-asserted-by":"publisher","first-page":"2680","DOI":"10.1109\/jbhi.2022.3151741","article-title":"Multiple sclerosis lesion analysis in brain magnetic resonance images: techniques and clinical applications","volume":"26","author":"Ma","year":"2022","journal-title":"IEEE J Biomed Health Inform"},{"key":"B30","doi-asserted-by":"publisher","first-page":"77","DOI":"10.1016\/j.neuroimage.2016.12.064","article-title":"Longitudinal multiple sclerosis lesion segmentation: resource and challenge","volume":"148","author":"Carass","year":"2017","journal-title":"NeuroImage"},{"key":"B31","doi-asserted-by":"publisher","first-page":"102854","DOI":"10.1016\/j.nicl.2021.102854","article-title":"ALL-Net: anatomical information lesion-wise loss function integrated into neural network for multiple sclerosis lesion segmentation","volume":"32","author":"Zhang","year":"2021","journal-title":"NeuroImage: Clin"}],"container-title":["Frontiers in Radiology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fradi.2026.1868811\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:48:10Z","timestamp":1790660890000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fradi.2026.1868811\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,29]]},"references-count":31,"alternative-id":["10.3389\/fradi.2026.1868811"],"URL":"https:\/\/doi.org\/10.3389\/fradi.2026.1868811","relation":{},"ISSN":["2673-8740"],"issn-type":[{"value":"2673-8740","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,29]]},"article-number":"1868811"}}