DocumentCode
674102
Title
Dynamic registration of multiple-view-US and MRI for the characterization of hypertrophic cardiomyopathy
Author
Betancur, J. ; Simon, A. ; Schnell, Frederic ; Tavard, François ; Donal, Erwan ; Hernandez, Alfredo I. ; Garreau, M.
Author_Institution
INSERM, Rennes, France
fYear
2013
fDate
22-25 Sept. 2013
Firstpage
237
Lastpage
240
Abstract
The relationship between strain and fibrosis for patients with hypertrophic cardiomyopathy (HCM) remains unclear. This work aims to fuse (i) myocardial fibrosis assessed from late-galodinium-enhanced MRI (LGE) sequence and, (ii) myocardial strain from multiple-view 2D speckle tracking echocardiography (STE). A spatio-temporal alignment is performed between segmented data extracted from cine-MRI short-axis-view and STE apical two- and four-chamber views. Dynamic Time Warping (DTW) and Fourier descriptors are used to handle the non-linear relationship between the cardiac phases of input sequences and the different levels of details provided by the two modalities. An expert evaluated the results visually, leading to a percentage of exploitable cardiac segments of 87%. Finally, MRI-LGE image was superimposed to the corresponding mid-diastolic STE contours to perform the combined tissue and strain analysis. Results on ten patients suggest that myocardial fibrosis is associated with a decrease in longitudinal strain and therefore that STE can be used to detect myocardial fibrosis.
Keywords
Fourier analysis; biological tissues; biomedical MRI; cardiovascular system; diseases; echocardiography; feature extraction; gallium; image segmentation; image sequences; medical image processing; spatiotemporal phenomena; ultrasonic imaging; Fourier descriptors; STE apical four-chamber views; STE apical two-chamber views; cardiac phases; cine-MRI short-axis-view; dynamic registration; dynamic time warping; exploitable cardiac segments; hypertrophic cardiomyopathy characterization; input sequences; late-galodinium-enhanced MRI sequence; longitudinal strain; middiastolic STE contours; multiple-view 2D speckle tracking echocardiography; multiple-view-ultrasonic imaging; myocardial strain; nonlinear relationship; segmented data extraction; spatiotemporal alignment; strain analysis; tissue analysis; Echocardiography; Heart; Image segmentation; Magnetic resonance imaging; Measurement; Myocardium; Strain;
fLanguage
English
Publisher
ieee
Conference_Titel
Computing in Cardiology Conference (CinC), 2013
Conference_Location
Zaragoza
ISSN
2325-8861
Print_ISBN
978-1-4799-0884-4
Type
conf
Filename
6712455
Link To Document