Automatic quantification of the LV function and mass: A deep learning approach for cardiovascular MRI. (February 2019)

• Record Type: Journal Article
• Authors: Curiale, Ariel H.; Colavecchia, Flavio D.; Mato, German
• Is Part Of: Computer methods and programs in biomedicine. Volume 169(2019)
• Page Start: 37
• Page End: 50
• Publication Date: 2019-02
• Subjects: Left ventricle quantification -- Myocardial segmentation -- Convolutional neural network -- Deep learning; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2018.12.002
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: This paper proposes an accurate deep learning approach to assist the automatization of left ventricle function and mass quantification in cardiac MRI. We study three new deep learning architectures specially designed for this task where the generalized Jaccard distance is used as optimization objective function. Also, we integrate the idea of sparsity and depthwise separable convolution into the Unet architecture, as well as, a residual learning strategy level to level. Our results on a databset of 137 patients have demonstrated that the proposed approach gets a suitable accuracy for myocardial segmentation, and a strong correlation with the most relevant functional measures. It is important to note that the errors are comparable to the inter e intra-operator ranges for manual contouring. Abstract: Objective: This paper proposes a novel approach for automatic left ventricle (LV) quantification using convolutional neural networks (CNN). Methods: The general framework consists of one CNN for detecting the LV, and another for tissue classification. Also, three new deep learning architectures were proposed for LV quantification. These new CNNs introduce the ideas of sparsity and depthwise separable convolution into the U-net architecture, as well as, a residual learning strategy level-to-level. To this end, we extend the classical U-net architecture and use the generalized Jaccard distance as optimization objective function. Results: The CNNs were trained and evaluated with 140 patients from two public cardiovascular magnetic resonance datasets (Sunnybrook and Cardiac Atlas Project) by using a 5-fold cross-validation strategy. Our results demonstrate a suitable accuracy for myocardial segmentation ( ∼ 0.9 Dice's coefficient), and a strong correlation with the most relevant physiological measures: 0.99 for end-diastolic and end-systolic volume, 0.97 for the left myocardial mass, 0.95 for the ejection fraction and 0.93 for the stroke volume and cardiac output. Conclusion: Our simulation and clinical evaluation results demonstrate the capability and merits of the proposed CNN to estimate different structural and functional features such as LV mass and EF which are commonly used for both diagnosis and treatment of different pathologies. Significance: This paper suggests a new approach for automatic LV quantification based on deep learning where errors are comparable to the inter- and intra-operator ranges for manual contouring.