Exploring the Potential of 3D Ultrasound Reverse Time Migration in Human Brain Imaging

Abstract

Ultrasound technology has become a major research focus in medical imaging because of its noninvasiveness, cost-effectiveness, and safety. However, due to the physical diffraction limit, conventional ultrasound imaging struggles to satisfy increasing clinical demands for resolution and image quality. Reverse time migration (RTM) can accurately capture small-scale features and complex structures, and this work investigates its application to human brain imaging. First, the finite-difference (FD) solver is easily affected by spatial discretization errors during the simulation of high-frequency ultrasound, leading to spatial dispersion effects. Therefore, we employ the pseudo-spectral (PS) solver, which accurately calculates the spatial derivatives with Fourier transform. Second, to further overcome the cumulative errors of time derivatives, we introduce an improved finite difference scheme to eliminate the time dispersion effect. Under comparable accuracy, 2D single-source simulations show that a 16th-order FD solver, implemented with a reduced time step, requires 18.2601% more computation time than the PS solver operating with a larger time step. Third, an illumination-compensated inverse scattering imaging condition is introduced to suppress backscattering noise and reduce migration artifacts. Simulation results indicate that the 3D RTM algorithm performs excellently in the imaging of brain tissue structures and enabling the accurate reconstruction of the brain's anatomical structure when using a 360-degree “ring” shaped observation system. Experimental tests further demonstrate the significant advantage of RTM over the traditional straight-ray delay-and-sum (DAS) method when using a finite ring array. Quantitative evaluations confirm a notable improvement in image quality: the contrast-to-noise ratio increases from 16.3432 (straight-ray DAS) to 41.39 (cross-correlation RTM) and 43.95 (inverse scattering RTM), while the edge sharpness improves from 0.0742 to 0.1060 and 0.1106, respectively. These results clearly indicate that RTM offers superior capabilities in enhancing image sharpness and overall reconstruction quality under limited aperture conditions.