Spatial compounding for 2D strain estimation in the mouse heart: A pilot study

Abstract

Abstract-Estimating cardiac strain in the mouse in the lateral direction using speckle tracking with adapted clinical equipment was shown to be challenging due to the fast heart rate and the large speckle size relative to the wall thickness. Compounding axial motion estimates acquired from different insonation angles can potentially improve lateral strain estimates. Therefore, the aim of this study was to test the feasibility of this methodology in the murine heart based on simulated data sets.A 3D kinematic model of a murine left ventricle was simulated and filled randomly with scatterers. Ultrasound short-axis images (10mm x 6mm) were obtained by assuming a linear array transducer. Beam steering was simulated at 3 different angles (22. Axial motion was estimated in each data set by 1D cross-correlation. A dynamic programming approach was integrated in the motion estimation algorithm to avoid discontinuities. Axial components were combined to reconstruct the in-plane motion vector. The 2D displacement fields were subsequently accumulated over the whole cycle. The procedure was repeated for 10 different distributions of scatterers to acquire 10 different RF data sets (5 for parameter tuning and 5 for comparing the methods).Radial and circumferential RMS strain errors calculated from the accumulated motion fields were compared with those obtained with 2D speckle tracking. Spatial compounding yielded significantly better radial (RMSE: 0.0737 ± 0.0078 vs. 0.112 ± 0.0094) as well as circumferential strain (RMSE: 0.102 ± 0.0097 vs. 0.281 ± 0.054).

2D myocardial strain in the mouse through spatial compounding: In-vivo feasibility study

“…Therefore, we applied the compounding approach on simulated short axis data sets of a murine left ventricle in a previous study [6]. It was shown that the compounding method significantly outperformed the standard 2D speckle tracking algorithm.…”