DocumentCode :
2352573
Title :
P1A-5 Analysis of 3D Motion Effects in Myocardial Elastography
Author :
Konofagou, Elisa E. ; Lee, Wei-Ning
Author_Institution :
Dept. of Biomed. Eng., Columbia Univ., New York, NY
fYear :
2006
fDate :
2-6 Oct. 2006
Firstpage :
1217
Lastpage :
1220
Abstract :
In this paper, we investigate the elevational beamwidth and the extent of 3D decorrelation on 3D motion estimation of multiple short-axis slices under a previously established theoretical framework, which simulated two sequences, including passive filling and active contraction, in normal and ischemic canine finite-element (FE) left ventricular models. A 3D convolutional image formation model was developed to generate successive RF signals in 3D. Incremental lateral and axial motion components were estimated from lateral-axial plane using 1D cross-correlation and recorrelation techniques in a 2D search with a 1D matching kernel of 3 mm and 80% overlap. Similarly, the elevational displacement was estimated from elevational-axial plane using the same strategy. The cumulative 3D motion was obtained by accumulating the incremental motion from end-diastole to end-systole, while the cumulative lateral and axial strains were obtained using a least-squares strain estimator on the cumulative in-plane motion. Radial and circumferential strains, independent of the orientation of the transducer, were also computed from the in-plane finite strains. The elastographic estimation error was estimated by calculating the relative difference between the FEA and elastographic results. When the elevational beamwidth increased from 2 to 4 mm, the mean absolute differences for cumulative in-plane motion and systolic strains in both the normal and the ischemic cases, decreased. However, the larger the elevational beamwidth was, the larger elevational displacement estimation error was obtained. The findings showed that the elevational beamwidth of 2 mm was the optimal value for both in-plane deformation and out-of-plane motion estimation. In summary, even at the presence of physiologic elevational motion, the in-plane elastographic estimates remained at good agreement with the FE solutions. Error analyses indicated thus a substantial improvement in the performance of the recorrelation techni- que at various short-axis slices of the 3D left ventricle and at elevational beamwidths well within the clinical equipment range
Keywords :
biomechanics; biomedical ultrasonics; cardiology; elasticity; finite element analysis; least squares approximations; 1D matching kernel; 2 mm; 3D convolutional image formation model; 3D decorrelation; 3D motion effect analysis; 4 mm; active ventricular contraction; axial motion components; circumferential strains; cross correlation techniques; cumulative 3D motion; elastographic estimation error; elevational beamwidth; elevational displacement; end-diastole; end-systole; finite-element models; in-plane deformation motion estimation; ischemic canine FE left ventricular models; lateral motion components; least squares strain estimator; myocardial elastography; normal canine FE left ventricular models; out-of-plane motion estimation; passive ventricular filling; radial strains; recorrelation techniques; systolic strains; transducer orientation; Capacitive sensors; Convolution; Decorrelation; Estimation error; Filling; Finite element methods; Motion analysis; Motion estimation; Myocardium; Signal generators;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Ultrasonics Symposium, 2006. IEEE
Conference_Location :
Vancouver, BC
ISSN :
1051-0117
Print_ISBN :
1-4244-0201-8
Electronic_ISBN :
1051-0117
Type :
conf
DOI :
10.1109/ULTSYM.2006.310
Filename :
4152166
Link To Document :
بازگشت