Scanning-mode 2D acoustic radiation force impulse (s2D-ARFI) imaging based on GPU acceleration

Acoustic radiation force impulse (ARFI) technique is a quantitative method for tissue stiffness assessment. It has been proved to be less operator dependent than the quasi-static elastography, and has more simple hardware architecture than the supersonic shearwave imaging (SSI) technique, which make it easier to be miniaturized for some special clinical applications. However, unlike the SSI, ARFI cannot provide real-time 2D images of tissue stiffness distribution mainly due to its data-intensive and time-consuming algorithms. In this study, the algorithms of ARFI were modified and improved to fit for the parallel computation on graphics processing unit (GPU), and the quasi-real-time scanning-mode 2D ARFI images (s2D-ARFI) were implemented on a self-developed compact system. High ratio of the time consumptions between the algorithms using CPU and using GPU has been verified, and it was also proved that there was no distinct difference between the stiffness images obtained by these two methods. The s2D-ARFI provides us an additional choice for quantitatively imaging the tissue stiffness, and has a potential to be miniaturized and used in the emergency treatments in field first-aid and the donor evaluation for organ transplantation.