Session: 06-11-02: Biotechnology and General Applications
Paper Number: 148669
148669 - A Computational Model to Process Dense Mr Images and Compute Cardiac Strains
Cardiac strains describe the local state of deformation of the cardiac tissue throughout the cardiac cycle and therefore may aid the detection of cardiac diseases, their progression, their severity, and their response to therapy. However, cardiac strains are still not integrated in clinical practice, which mostly still relies on global indices of cardiac function such as ejection fraction. One of the major obstacles preventing routine use of local cardiac strains in the clinic is the lack of robustness in the computed strain values. Indeed, the high variability in strain values may prevent their use as a reliable marker to characterize cardiac health.
In this work, we present a method to process Displacement Encoding with Stimulated Echoes (DENSE) magnetic resonance (MR) images and more robustly compute cardiac strains. DENSE provides direct measures of voxel-wise – not only surface as, for example, cine MRI – displacements and therefore is an ideal starting point to compute accurate cardiac strains. Our method aims to process the acquired images using a computational model with both smoothing and denoising elements. Smoothing is achieved by penalizing large variations and abrupt changes in the components of the displacement field. The denoising element consists in penalizing the part of the displacement field leading to volumetric deformation of the cardiac tissue. The penalization of volume deformation leverages the fact that the myocardial deformation is nearly incompressible during the cardiac cycle and therefore filters out the displacement noise associated with non-isochoric tissue deformations. The proposed method is tested using pre-clinical, midventricular DENSE MR images acquired in healthy swine subjects (all experiments were conducted under UCLA IACUC protocol ARC #2015-124). The DENSE images in plane resolution was 2.5 x 2.5 mm2 while the slice thickness was equal to 8 mm. The acquired temporal resolution was 30 ms with viewsharing reconstruction at 15 ms. Other key sequence parameters are: TE/TR=1.04/15, ke=0.08 cycles/mm, Navg=3, spiral interleaves =10.
Cardiac strains are computed in the radial, circumferential, and longitudinal directions and are evaluated in terms of absolute values and presence (for radial and circumferential strains) and absence (for longitudinal strains) of a transmural gradient. Furthermore, strain values are compared to those computed using radial basis functions (a traditional scheme to interpolate displacement data) and the DENSE analysis software [1,2], a freely available framework to process DENSE MRI data.
Finally, we will discuss the current limitations of the denoising and interpolation components and future steps to develop this processing pipeline and its inclusion into motion guided schemes for semi-automatic processing of DENSE images to compute cardiac strains.
Acknowledgements
This material is based upon work supported by the National Science Foundation under Grant No. 2237391. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
References
[1] Spottiswoode, B.S., Zhong, X., Hess, A.T., Kramer, C.M., Meintjes, E.M., Mayosi, B.M., and Epstein, F.H. (2007). Tracking myocardial motion from cine DENSE images using spatiotemporal phase unwrapping and temporal fitting. IEEE Transactions on Medical Imaging, 26(1), 15–30. http://doi.org/10.1109/TMI.2006.884215
[2] Gilliam, A.D., Suever, J.D., and contributors (2021). DENSEanalysis. Retrieved from https://github.com/denseanalysis/denseanalysis
Presenting Author: Luigi Perotti University of Central Florida
Presenting Author Biography: Luigi Perotti received his Laurea (B.S./M.S.) degree in Civil Engineering from Politecnico di Milano, Italy, in 2004. Subsequently he continued his studies in Mechanical Engineering at Caltech where he received his M.S. in 2006 and his Ph.D. in 2011. After defending his PhD thesis, he was a postdoctoral scholar at Caltech from the end of 2010 to the end of 2011. At the end of 2011, he joined the Mechanical and Aerospace Engineering department at UCLA as a postdoctoral scholar where he pursued his research interests in biomechanics. In 2014 he joined the Radiological
Sciences department at UCLA, first as a postdoctoral scholar and later (2016) as a project scientist. Dr. Perotti joined the Mechanical and Aerospace Engineering department at the University of Central Florida in 2019 where he leads the Computational Biomechanics Lab.
Authors:
Uditha Weerasinghage University of Central FloridaKevin Moulin Boston Children's Hospital
Luigi Perotti University of Central Florida
A Computational Model to Process Dense Mr Images and Compute Cardiac Strains
Paper Type
Technical Presentation