Automatic Cardiac Disease Assessment on cine-MRI via Time-Series Segmentation and Domain Specific Features

被引:175
作者
Isensee, Fabian [1 ]
Jaeger, Paul F. [1 ]
Full, Peter M. [2 ,3 ]
Wolf, Ivo [3 ]
Engelhardt, Sandy [2 ,3 ]
Maier-Hein, Klaus H. [1 ]
机构
[1] German Canc Res Ctr, Med Image Comp, Heidelberg, Germany
[2] German Canc Res Ctr, Div Comp Assisted Med Intervent, Heidelberg, Germany
[3] Mannheim Univ Appl Sci, Dept Comp Sci, Mannheim, Germany
来源
STATISTICAL ATLASES AND COMPUTATIONAL MODELS OF THE HEART: ACDC AND MMWHS CHALLENGES | 2018年 / 10663卷
关键词
Automated cardiac diagnosis challenge; Cardiac magnetic resonance imaging; Disease prediction; Deep learning; CNN;
D O I
10.1007/978-3-319-75541-0_13
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
Cardiac magnetic resonance imaging improves on diagnosis of cardiovascular diseases by providing images at high spatiotemporal resolution. Manual evaluation of these time-series, however, is expensive and prone to biased and non-reproducible outcomes. In this paper, we present a method that addresses named limitations by integrating segmentation and disease classification into a fully automatic processing pipeline. We use an ensemble of UNet inspired architectures for segmentation of cardiac structures such as the left and right ventricular cavity (LVC, RVC) and the left ventricular myocardium (LVM) on each time instance of the cardiac cycle. For the classification task, information is extracted from the segmented time-series in form of comprehensive features handcrafted to reflect diagnostic clinical procedures. Based on these features we train an ensemble of heavily regularized multilayer perceptrons (MLP) and a random forest classifier to predict the pathologic target class. We evaluated our method on the ACDC dataset (4 pathology groups, 1 healthy group) and achieve dice scores of 0.945 (LVC), 0.908 (RVC) and 0.905 (LVM) in a cross-validation over the training set (100 cases) and 0.950 (LVC), 0.923 (RVC) and 0.911 (LVM) on the test set (50 cases). We report a classification accuracy of 94% on a training set cross-validation and 92% on the test set. Our results underpin the potential of machine learning methods for accurate, fast and reproducible segmentation and computer-assisted diagnosis (CAD).
引用
收藏
页码:120 / 129
页数:10
相关论文
共 11 条
[1]  
[Anonymous], 2017, ARXIV170205747
[2]  
[Anonymous], 2017, ARXIV170508943
[3]  
[Anonymous], 2017, ARXIV170508302
[4]  
[Anonymous], 2017, ARXIV170103056
[5]  
Cicek Ozgun, 2016, Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016. 19th International Conference. Proceedings: LNCS 9901, P424, DOI 10.1007/978-3-319-46723-8_49
[6]   Computer-aided diagnosis in medical imaging: Historical review, current status and future potential [J].
Doi, Kunio .
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, 2007, 31 (4-5) :198-211
[7]  
JAY NC, 2000, J AM COLL CARDIOL, V35, P569, DOI DOI 10.1016/S0735-1097(99)00630-0
[8]   Left ventricular shape variation in asymptomatic populations: the multi-ethnic study of atherosclerosis [J].
Medrano-Gracia, Pau ;
Cowan, Brett R. ;
Ambale-Venkatesh, Bharath ;
Bluemke, David A. ;
Eng, John ;
Finn, John Paul ;
Fonseca, Carissa G. ;
Lima, Joao A. C. ;
Suinesiaputra, Avan ;
Young, Alistair A. .
JOURNAL OF CARDIOVASCULAR MAGNETIC RESONANCE, 2014, 16
[9]   U-Net: Convolutional Networks for Biomedical Image Segmentation [J].
Ronneberger, Olaf ;
Fischer, Philipp ;
Brox, Thomas .
MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION, PT III, 2015, 9351 :234-241
[10]  
Tran P.V., 2016, ARXIV