In Vivo Ultrasound Dynamic Coronary Blood Flow Imaging through Adaptive Frame Selection Method

Use of ultrasound for coronary imaging in commercial echocardiography remains challenging because of the small nature of coronary vasculature and the myocardium's intricate motion. Several super-resolution imaging techniques have been applied for coronary imaging; however, most of them only measure the coronary flow during the diastolic phase and have a long data acquisition time. To address these problems, this study proposes an adaptive frame selection approach for coronary vasculature imaging. In this approach, similar frames within cardiac cycles are selected using the sum of absolute difference (SAD) algorithm, and the coronary vasculature blood flow is calculated without using electrocardiographic (ECG) gating data. Experiments were performed in mouse hearts through high-frequency ultrafast ultrasound imaging. After similar frames were selected from several cardiac cycles (one-five cycles), a singular value decomposition (SVD) filter was applied to extract blood flow signals and obtain a dynamic coronary vasculature image, the accuracy of which was confirmed by measuring Doppler sonograms from the left coronary artery (LCA) and arterioles. The conventional method (without SAD), in which only blood flow in the diastolic phase is calculated, was also conducted to enable a comparison in terms of measured vessel size and signal-to-noise ratio (SNR). The SNR for the proposed approach was found to be $20.74~\pm ~1.62$ dB, under the best parameter settings. The proposed approach was successfully verified in the small animal model and has potential for use in human dynamic coronary artery imaging.