Automatic deep learning-based myocardial infarction segmentation from delayed enhancement MRI

被引:22
作者
Chen, Zhihao [1 ]
Lalande, Alain [2 ,3 ]
Salomon, Michel [1 ]
Decourselle, Thomas [4 ]
Pommier, Thibaut [6 ]
Qayyum, Abdul [2 ]
Shi, Jixi [1 ,5 ]
Perrot, Gilles [1 ]
Couturier, Raphael [1 ]
机构
[1] Univ Bourgogne Franche Comte, FEMTO ST Inst, UMR6174, CNRS, Belfort, France
[2] Univ Bourgogne Franche Comte, ImViA Lab, EA7535, Dijon, France
[3] Univ Hosp Dijon, Dept Med Imaging, Dijon, France
[4] CASIS Co, Quetigny, France
[5] Univ Normandie, ESIGELEC, EA4353, IRSEEM, Rouen, France
[6] Univ Hosp Dijon, Dept Cardiol, Dijon, France
关键词
CNN; Semantic segmentation; DE-MRI; Myocardial infarction; Adaptive framework; LEFT-VENTRICLE; NETWORKS;
D O I
10.1016/j.compmedimag.2021.102014
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Delayed Enhancement cardiac MRI (DE-MRI) has become indispensable for the diagnosis of myocardial diseases. However, to quantify the disease severity, doctors need time to manually annotate the scar and myocardium. To address this issue, in this paper we propose an automatic myocardial infarction segmentation approach on the left ventricle from short-axis DE-MRI based on Convolutional Neural Networks (CNN). The objective is to segment myocardial infarction on short-axis DE-MRI images of the left ventricle acquired 10 min after the in-jection of a gadolinium-based contrast agent. The segmentation of the infarction area is realized in two stages: a first CNN model finds the contour of myocardium and a second CNN model segments the infarction. Compared to the manual intra-observer and inter-observer variations for the segmentation of myocardial infarction, and to the automatic segmentation with Gaussian Mixture Model, our proposal achieves satisfying segmentation results on our dataset of 904 DE-MRI slices.
引用
收藏
页数:9
相关论文
共 32 条
[1]   Accurate and objective infarct sizing by contrast-enhanced magnetic resonance imaging in a canine myocardial infarction model [J].
Amado, LC ;
Gerber, BL ;
Gupta, SN ;
Szarf, G ;
Schock, R ;
Nasir, K ;
Kraitchman, DL ;
Lima, JAC .
JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY, 2004, 44 (12) :2383-2389
[2]  
[Anonymous], 2016, ARXIV161205360
[3]   A combined deep-learning and deformable-model approach to fully automatic segmentation of the left ventricle in cardiac MRI [J].
Avendi, M. R. ;
Kheradvar, Arash ;
Jafarkhani, Hamid .
MEDICAL IMAGE ANALYSIS, 2016, 30 :108-119
[4]  
Bahdanau D., 2015, ICLR
[5]   COLOR EQUALIZATION METHOD AND ITS APPLICATION TO COLOR IMAGE-PROCESSING [J].
BOCKSTEIN, IM .
JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION, 1986, 3 (05) :735-737
[6]   Deep Learning for Cardiac Image Segmentation: A Review [J].
Chen, Chen ;
Qin, Chen ;
Qiu, Huaqi ;
Tarroni, Giacomo ;
Duan, Jinming ;
Bai, Wenjia ;
Rueckert, Daniel .
FRONTIERS IN CARDIOVASCULAR MEDICINE, 2020, 7
[7]   DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs [J].
Chen, Liang-Chieh ;
Papandreou, George ;
Kokkinos, Iasonas ;
Murphy, Kevin ;
Yuille, Alan L. .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2018, 40 (04) :834-848
[8]   Myocardial Infarction Quantification from Late Gadolinium Enhancement MRI Using Top-Hat Transforms and Neural Networks [J].
de la Rosa, Ezequiel ;
Sidibe, Desire ;
Decourselle, Thomas ;
Leclercq, Thibault ;
Cochet, Alexandre ;
Lalande, Alain .
ALGORITHMS, 2021, 14 (08)
[9]   Automated Cardiac MR Scar Quantification in Hypertrophic Cardiomyopathy Using Deep Convolutional Neural Networks [J].
Fahmy, Ahmed S. ;
Rausch, Johannes ;
Neisius, Ulf ;
Chan, Raymond H. ;
Maron, Martin S. ;
Appelbaum, Evan ;
Menze, Bjoern ;
Nezafat, Reza .
JACC-CARDIOVASCULAR IMAGING, 2018, 11 (12) :1917-1918
[10]  
Goodfellow I, 2016, ADAPT COMPUT MACH LE, P1