Technology for imaging the heart has advanced dramatically in recent years. In particular, real-time three-dimensional ultrasound (RT3D or 4D) has captured the imagination of cardiologists with its ability to obtain complete three-dimensional images of the heart over an entire cardiac cycle in just a few seconds of imaging. The complex cardiac wall motion and temporal information contained in these four-dimensional (3D + time) data sequences has the potential to greatly enhance clinical diagnoses of the heart. However, most cardiac examination centers still depend on 2D echocardiography, and the temporal information in the 4D cardiac imaging is often overlooked. The goal of this proposal is to utilize the wealth of information contained in 4D ultrasound. In particular, this proposal will focus on a common problem for which 4D cardiac imaging is ideally suited: measuring strain and cardiac torsion directly from 4D ultrasound.
The aims of this proposal are: (1) Quantify errors of displacement and strain from optical flow on 4D ultrasound with respect to results obtained by sonomicrometry implanted in dog hearts;(2) Test the hypothesis that optical flow on 4D ultrasound can have clinically equivalent results on healthy volunteers (normal hearts) and diseased hearts compared to similar measures of strain obtained by MRI based methods;(3) Test the hypothesis that optical flow on 4D ultrasound has specific advantages in terms of performance over existing 2D based methods of strain estimation. The design of our method is based on optical flow to track myocardial motion in 4D ultrasound. Directional displacements and strains and cardiac torsion can be automatically derived from the 3D + time motion field estimated.
In aims 1 and 2, specific displacements, strains, and torsion will be measured and compared to """"""""gold standards"""""""" in quantitative evaluation studies.
In aim 3, we will extensively compare the performance of 4D methods and 2D methods in estimating strain measures across a wide range of clinical echocardiography patient data sets. Once completed, the questions of why and when to best use 4D rather then 2D measures will be answered quantitatively.

Public Health Relevance

The significance of the proposed work is that it will provide a novel and effective 3D strain and torsion measuring tool, allowing clinicians to routinely measure wall motion quantitatively and in real-time. In addition this research will provide cardiologists with improved screening tests for diabetic cardiomyopathy (and dyssynchrony) to help them determine those patients that may require more invasive and costly procedures such as heart catheterization from patients that can be safely spared such tests.

National Institute of Health (NIH)
National Heart, Lung, and Blood Institute (NHLBI)
Research Project (R01)
Project #
Application #
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Buxton, Denis B
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Columbia University (N.Y.)
Biomedical Engineering
Schools of Engineering
New York
United States
Zip Code
Uzunba?, Mustafa Gökhan; Chen, Chao; Metaxas, Dimitris (2014) Optree: a learning-based adaptive watershed algorithm for neuron segmentation. Med Image Comput Comput Assist Interv 17:97-105
Yu, Yang; Zhang, Shaoting; Li, Kang et al. (2014) Deformable models with sparsity constraints for cardiac motion analysis. Med Image Anal 18:927-37
Uzunba?, Mustafa Gökhan; Chen, Chao; Zhang, Shaoting et al. (2013) Collaborative multi organ segmentation by integrating deformable and graphical models. Med Image Comput Comput Assist Interv 16:157-64
Lorsakul, Auranuch; Gamarnik, Viktor; Duan, Qi et al. (2012) Impact of temporal resolution on LV myocardial regional strain assessment with real-time 3D ultrasound. Conf Proc IEEE Eng Med Biol Soc 2012:4075-8
Lorsakul, Auranuch; Duan, Qi; Po, Ming Jack et al. (2011) Parameterization of real-time 3D speckle tracking framework for cardiac strain assessment. Conf Proc IEEE Eng Med Biol Soc 2011:2654-7
Po, Ming Jack; Lorsakul, Auranuch; Duan, Qi et al. (2010) In-vivo clinical validation of cardiac deformation and strain measurements from 4D ultrasound. Conf Proc IEEE Eng Med Biol Soc 2010:41-4
Duan, Qi; Angelini, Elsa D; Laine, Andrew F (2010) Real-time segmentation by Active Geometric Functions. Comput Methods Programs Biomed 98:223-30
Duan, Qi; Angelini, Elsa D; Herz, Susan L et al. (2009) Region-based endocardium tracking on real-time three-dimensional ultrasound. Ultrasound Med Biol 35:256-65
Duan, Qi; Parker, Katherine M; Lorsakul, Auranuch et al. (2009) QUANTITATIVE VALIDATION OF OPTICAL FLOW BASED MYOCARDIAL STRAIN MEASURES USING SONOMICROMETRY. Proc IEEE Int Symp Biomed Imaging 2009:454-457