Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images

White matter hyperintensities (WMHs) have been associated with various cerebrovascular and neurodegenerative diseases. Reliable quantification of WMHs is essential for understanding their clinical impact in normal and pathological populations. Automated segmentation of WMHs is highly challenging due...

पूर्ण विवरण

ग्रंथसूची विवरण
मुख्य लेखकों: Sundaresan, V, Zamboni, G, Rothwell, PM, Jenkinson, M, Griffanti, L
स्वरूप: Journal article
भाषा:English
प्रकाशित: Elsevier 2021
_version_ 1826260067045867520
author Sundaresan, V
Zamboni, G
Rothwell, PM
Jenkinson, M
Griffanti, L
author_facet Sundaresan, V
Zamboni, G
Rothwell, PM
Jenkinson, M
Griffanti, L
author_sort Sundaresan, V
collection OXFORD
description White matter hyperintensities (WMHs) have been associated with various cerebrovascular and neurodegenerative diseases. Reliable quantification of WMHs is essential for understanding their clinical impact in normal and pathological populations. Automated segmentation of WMHs is highly challenging due to heterogeneity in WMH characteristics between deep and periventricular white matter, presence of artefacts and differences in the pathology and demographics of populations. In this work, we propose an ensemble triplanar network that combines the predictions from three different planes of brain MR images to provide an accurate WMH segmentation. In the loss functions the network uses anatomical information regarding WMH spatial distribution in loss functions, to improve the efficiency of segmentation and to overcome the contrast variations between deep and periventricular WMHs. We evaluated our method on 5 datasets, of which 3 are part of a publicly available dataset (training data for MICCAI WMH Segmentation Challenge 2017 - MWSC 2017) consisting of subjects from three different cohorts, and we also submitted our method to MWSC 2017 to be evaluated on the unseen test datasets. On evaluating our method separately in deep and periventricular regions, we observed robust and comparable performance in both regions. Our method performed better than most of the existing methods, including FSL BIANCA, and on par with the top ranking deep learning methods of MWSC 2017.
first_indexed 2024-03-06T18:59:44Z
format Journal article
id oxford-uuid:1319c8f5-8f19-40f4-a5ac-4628fec3895c
institution University of Oxford
language English
last_indexed 2024-03-06T18:59:44Z
publishDate 2021
publisher Elsevier
record_format dspace
spelling oxford-uuid:1319c8f5-8f19-40f4-a5ac-4628fec3895c2022-03-26T10:11:55ZTriplanar ensemble U-Net model for white matter hyperintensities segmentation on MR imagesJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:1319c8f5-8f19-40f4-a5ac-4628fec3895cEnglishSymplectic ElementsElsevier2021Sundaresan, VZamboni, GRothwell, PMJenkinson, MGriffanti, LWhite matter hyperintensities (WMHs) have been associated with various cerebrovascular and neurodegenerative diseases. Reliable quantification of WMHs is essential for understanding their clinical impact in normal and pathological populations. Automated segmentation of WMHs is highly challenging due to heterogeneity in WMH characteristics between deep and periventricular white matter, presence of artefacts and differences in the pathology and demographics of populations. In this work, we propose an ensemble triplanar network that combines the predictions from three different planes of brain MR images to provide an accurate WMH segmentation. In the loss functions the network uses anatomical information regarding WMH spatial distribution in loss functions, to improve the efficiency of segmentation and to overcome the contrast variations between deep and periventricular WMHs. We evaluated our method on 5 datasets, of which 3 are part of a publicly available dataset (training data for MICCAI WMH Segmentation Challenge 2017 - MWSC 2017) consisting of subjects from three different cohorts, and we also submitted our method to MWSC 2017 to be evaluated on the unseen test datasets. On evaluating our method separately in deep and periventricular regions, we observed robust and comparable performance in both regions. Our method performed better than most of the existing methods, including FSL BIANCA, and on par with the top ranking deep learning methods of MWSC 2017.
spellingShingle Sundaresan, V
Zamboni, G
Rothwell, PM
Jenkinson, M
Griffanti, L
Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title_full Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title_fullStr Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title_full_unstemmed Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title_short Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images
title_sort triplanar ensemble u net model for white matter hyperintensities segmentation on mr images
work_keys_str_mv AT sundaresanv triplanarensembleunetmodelforwhitematterhyperintensitiessegmentationonmrimages
AT zambonig triplanarensembleunetmodelforwhitematterhyperintensitiessegmentationonmrimages
AT rothwellpm triplanarensembleunetmodelforwhitematterhyperintensitiessegmentationonmrimages
AT jenkinsonm triplanarensembleunetmodelforwhitematterhyperintensitiessegmentationonmrimages
AT griffantil triplanarensembleunetmodelforwhitematterhyperintensitiessegmentationonmrimages