In vivo Apical Infarct Localization using Adaptive Bayesian Cardiac Strain Imaging

Cardiac strain imaging (CSI) using ultrasound radiofrequency (RF) data is more sensitive to subtle myocardial motion abnormalities compared to echocardiographic measurements and envelope-based speckle tracking. Yet, CSI is being actively researched to address challenges from out-of-plane scatter motion due to complex 3D cardiac deformation imaged in 2D. Here, we report on a feasibility study to apply adaptive Bayesian CSI for in vivo apical infarct localization using 2D ultrasound RF data from murine models of ischemia-reperfusion (IR) and myocardial infarction (MI). High frequency (
$\mathrm{f}_{\mathrm{c}}=30\text{MHz}$
) ultrasound (US) radio-frequency data collected in parasternal short axis view at the apical level were tracked using an adaptive Bayesian regularization incorporated multi-level block matching algorithm. In vivo longitudinal study was designed with five imaging sessions (pre-surgery (BL) and 1,2,7 and 14 days post-surgery). End-systole circumferential strain images and values were compared among three different mouse models - sham, myocardial infarction, and ischemia-reperfusion. Findings from cardiac strain imaging demonstrated good correlation with the findings from ex vivo histopathological image analysis thus showing the feasibility of the proposed method.