Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningfagfællebedømt

Standard

Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty. / van Engelen, Arna; Niessen, Wiro J.; Klein, Stefan; Groen, Harald C.; Verhagen, Hence J. M.; Wentzel, Jolanda J.; van der Lugt, Aad; de Bruijne, Marleen.

I: PLoS ONE, Bind 9, Nr. 4, e94840, 2014.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningfagfællebedømt

Harvard

van Engelen, A, Niessen, WJ, Klein, S, Groen, HC, Verhagen, HJM, Wentzel, JJ, van der Lugt, A & de Bruijne, M 2014, 'Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty', PLoS ONE, bind 9, nr. 4, e94840. https://doi.org/10.1371/journal.pone.0094840

APA

van Engelen, A., Niessen, W. J., Klein, S., Groen, H. C., Verhagen, H. J. M., Wentzel, J. J., ... de Bruijne, M. (2014). Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty. PLoS ONE, 9(4), [e94840]. https://doi.org/10.1371/journal.pone.0094840

Vancouver

van Engelen A, Niessen WJ, Klein S, Groen HC, Verhagen HJM, Wentzel JJ o.a. Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty. PLoS ONE. 2014;9(4). e94840. https://doi.org/10.1371/journal.pone.0094840

Author

van Engelen, Arna ; Niessen, Wiro J. ; Klein, Stefan ; Groen, Harald C. ; Verhagen, Hence J. M. ; Wentzel, Jolanda J. ; van der Lugt, Aad ; de Bruijne, Marleen. / Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty. I: PLoS ONE. 2014 ; Bind 9, Nr. 4.

Bibtex

@article{ca06d6fd206b4b4a997f88463bc93aa5,
title = "Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty",
abstract = "Atherosclerotic plaque composition can indicate plaque vulnerability. We segment atherosclerotic plaque components from the carotid artery on a combination of in vivo MRI and CT-angiography (CTA) data using supervised voxelwise classification. In contrast to previous studies the ground truth for training is directly obtained from 3D registration with histology for fibrous and lipid-rich necrotic tissue, and with [Formula: see text]CT for calcification. This registration does, however, not provide accurate voxelwise correspondence. We therefore evaluate three approaches that incorporate uncertainty in the ground truth used for training: I) soft labels are created by Gaussian blurring of the original binary histology segmentations to reduce weights at the boundaries between components, and are weighted by the estimated registration accuracy of the histology and in vivo imaging data (measured by overlap), II) samples are weighted by the local contour distance of the lumen and outer wall between histology and in vivo data, and III) 10{\%} of each class is rejected by Gaussian outlier rejection. Classification was evaluated on the relative volumes ({\%} of tissue type in the vessel wall) for calcified, fibrous and lipid-rich necrotic tissue, using linear discriminant (LDC) and support vector machine (SVM) classification. In addition, the combination of MRI and CTA data was compared to using only one imaging modality. Best results were obtained by LDC and outlier rejection: the volume error per vessel was 0.9[Formula: see text]1.0{\%} for calcification, 12.7[Formula: see text]7.6{\%} for fibrous and 12.1[Formula: see text]8.1{\%} for necrotic tissue, with Spearman rank correlation coefficients of 0.91 (calcification), 0.80 (fibrous) and 0.81 (necrotic). While segmentation using only MRI features yielded low accuracy for calcification, and segmentation using only CTA features yielded low accuracy for necrotic tissue, the combination of features from MRI and CTA gave good results for all studied components.",
author = "{van Engelen}, Arna and Niessen, {Wiro J.} and Stefan Klein and Groen, {Harald C.} and Verhagen, {Hence J. M.} and Wentzel, {Jolanda J.} and {van der Lugt}, Aad and {de Bruijne}, Marleen",
year = "2014",
doi = "10.1371/journal.pone.0094840",
language = "English",
volume = "9",
journal = "P L o S One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "4",

}

RIS

TY - JOUR

T1 - Atherosclerotic plaque component segmentation in combined carotid MRI and CTA data incorporating class label uncertainty

AU - van Engelen, Arna

AU - Niessen, Wiro J.

AU - Klein, Stefan

AU - Groen, Harald C.

AU - Verhagen, Hence J. M.

AU - Wentzel, Jolanda J.

AU - van der Lugt, Aad

AU - de Bruijne, Marleen

PY - 2014

Y1 - 2014

N2 - Atherosclerotic plaque composition can indicate plaque vulnerability. We segment atherosclerotic plaque components from the carotid artery on a combination of in vivo MRI and CT-angiography (CTA) data using supervised voxelwise classification. In contrast to previous studies the ground truth for training is directly obtained from 3D registration with histology for fibrous and lipid-rich necrotic tissue, and with [Formula: see text]CT for calcification. This registration does, however, not provide accurate voxelwise correspondence. We therefore evaluate three approaches that incorporate uncertainty in the ground truth used for training: I) soft labels are created by Gaussian blurring of the original binary histology segmentations to reduce weights at the boundaries between components, and are weighted by the estimated registration accuracy of the histology and in vivo imaging data (measured by overlap), II) samples are weighted by the local contour distance of the lumen and outer wall between histology and in vivo data, and III) 10% of each class is rejected by Gaussian outlier rejection. Classification was evaluated on the relative volumes (% of tissue type in the vessel wall) for calcified, fibrous and lipid-rich necrotic tissue, using linear discriminant (LDC) and support vector machine (SVM) classification. In addition, the combination of MRI and CTA data was compared to using only one imaging modality. Best results were obtained by LDC and outlier rejection: the volume error per vessel was 0.9[Formula: see text]1.0% for calcification, 12.7[Formula: see text]7.6% for fibrous and 12.1[Formula: see text]8.1% for necrotic tissue, with Spearman rank correlation coefficients of 0.91 (calcification), 0.80 (fibrous) and 0.81 (necrotic). While segmentation using only MRI features yielded low accuracy for calcification, and segmentation using only CTA features yielded low accuracy for necrotic tissue, the combination of features from MRI and CTA gave good results for all studied components.

AB - Atherosclerotic plaque composition can indicate plaque vulnerability. We segment atherosclerotic plaque components from the carotid artery on a combination of in vivo MRI and CT-angiography (CTA) data using supervised voxelwise classification. In contrast to previous studies the ground truth for training is directly obtained from 3D registration with histology for fibrous and lipid-rich necrotic tissue, and with [Formula: see text]CT for calcification. This registration does, however, not provide accurate voxelwise correspondence. We therefore evaluate three approaches that incorporate uncertainty in the ground truth used for training: I) soft labels are created by Gaussian blurring of the original binary histology segmentations to reduce weights at the boundaries between components, and are weighted by the estimated registration accuracy of the histology and in vivo imaging data (measured by overlap), II) samples are weighted by the local contour distance of the lumen and outer wall between histology and in vivo data, and III) 10% of each class is rejected by Gaussian outlier rejection. Classification was evaluated on the relative volumes (% of tissue type in the vessel wall) for calcified, fibrous and lipid-rich necrotic tissue, using linear discriminant (LDC) and support vector machine (SVM) classification. In addition, the combination of MRI and CTA data was compared to using only one imaging modality. Best results were obtained by LDC and outlier rejection: the volume error per vessel was 0.9[Formula: see text]1.0% for calcification, 12.7[Formula: see text]7.6% for fibrous and 12.1[Formula: see text]8.1% for necrotic tissue, with Spearman rank correlation coefficients of 0.91 (calcification), 0.80 (fibrous) and 0.81 (necrotic). While segmentation using only MRI features yielded low accuracy for calcification, and segmentation using only CTA features yielded low accuracy for necrotic tissue, the combination of features from MRI and CTA gave good results for all studied components.

U2 - 10.1371/journal.pone.0094840

DO - 10.1371/journal.pone.0094840

M3 - Journal article

C2 - 24762678

VL - 9

JO - P L o S One

JF - P L o S One

SN - 1932-6203

IS - 4

M1 - e94840

ER -

ID: 109044322