TY - JOUR
T1 - Development and testing of a deep learning-based strategy for scar segmentation on CMR-LGE images
AU - Moccia, Sara
AU - Banali, Riccardo
AU - Martini, Chiara
AU - Muscogiuri, Giuseppe
AU - Pontone, Gianluca
AU - Pepi, Mauro
AU - Caiani, Enrico Gianluca
PY - 2018/1/1
Y1 - 2018/1/1
N2 - Objective: The aim of this paper is to investigate the use of fully convolutional neural networks (FCNNs) to segment scar tissue in the left ventricle from cardiac magnetic resonance with late gadolinium enhancement (CMR-LGE) images. Methods: A successful FCNN in the literature (the ENet) was modified and trained to provide scar-tissue segmentation. Two segmentation protocols (Protocol 1 and Protocol 2) were investigated, the latter limiting the scar-segmentation search area to the left ventricular myocardial tissue region. CMR-LGE from 30 patients with ischemic-heart disease were retrospectively analyzed, for a total of 250 images, presenting high variability in terms of scar dimension and location. Segmentation results were assessed against manual scar-tissue tracing using one-patient-out cross validation. Results: Protocol 2 outperformed Protocol 1 significantly (p value < 0.05), with median sensitivity and Dice similarity coefficient equal to 88.07% [inter-quartile range (IQR) 18.84%] and 71.25% (IQR 31.82%), respectively. Discussion: Both segmentation protocols were able to detect scar tissues in the CMR-LGE images but higher performance was achieved when limiting the search area to the myocardial region. The findings of this paper represent an encouraging starting point for the use of FCNNs for the segmentation of nonviable scar tissue from CMR-LGE images.
AB - Objective: The aim of this paper is to investigate the use of fully convolutional neural networks (FCNNs) to segment scar tissue in the left ventricle from cardiac magnetic resonance with late gadolinium enhancement (CMR-LGE) images. Methods: A successful FCNN in the literature (the ENet) was modified and trained to provide scar-tissue segmentation. Two segmentation protocols (Protocol 1 and Protocol 2) were investigated, the latter limiting the scar-segmentation search area to the left ventricular myocardial tissue region. CMR-LGE from 30 patients with ischemic-heart disease were retrospectively analyzed, for a total of 250 images, presenting high variability in terms of scar dimension and location. Segmentation results were assessed against manual scar-tissue tracing using one-patient-out cross validation. Results: Protocol 2 outperformed Protocol 1 significantly (p value < 0.05), with median sensitivity and Dice similarity coefficient equal to 88.07% [inter-quartile range (IQR) 18.84%] and 71.25% (IQR 31.82%), respectively. Discussion: Both segmentation protocols were able to detect scar tissues in the CMR-LGE images but higher performance was achieved when limiting the search area to the myocardial region. The findings of this paper represent an encouraging starting point for the use of FCNNs for the segmentation of nonviable scar tissue from CMR-LGE images.
KW - CMR-LGE images
KW - Deep learning
KW - Fully-convolutional neural networks
KW - Scar segmentation
UR - http://www.scopus.com/inward/record.url?scp=85057046615&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85057046615&partnerID=8YFLogxK
U2 - 10.1007/s10334-018-0718-4
DO - 10.1007/s10334-018-0718-4
M3 - Article
AN - SCOPUS:85057046615
JO - Magnetic Resonance Materials in Physics, Biology, and Medicine
JF - Magnetic Resonance Materials in Physics, Biology, and Medicine
SN - 1352-8661
ER -