High-order singular value decomposition of contrast pulsing sequences for improved clutter suppression

Abstract

Contrast-enhanced ultrasound is a diagnostic tool used to visualize blood flow in the cardiovascular system. The use of ultrasound contrast agent (microbubbles) in combination with contrast pulsing scheme (CPS) improves the sensitivity and specificity of ultrasound flow imaging by enhancing the signal in the blood compartment. The commonly used CPS are pulse inversion (PI), amplitude modulation (AM), and amplitude-modulated pulse inversion (AMPI). Using differences in phase or amplitude of multiple pulses, the linear tissue clutter signal can be suppressed. However, this process can be degraded by motion or non-linear propagation of the ultrasound wave. These effects cause the cancellation of linear clutter signal to be ineffective We propose using higher-order singular value decomposition (HOSVD) with spatial, temporal, and pulsing dimensions as the input to improve clutter suppression under the conditions of motion and non-linear propagation. We performed systematic in-vitro- experiment emulating these conditions, as well as in-vivo cardiac measurements. The results showed that HOSVD increases the clutter suppression of all the 3 CPS compared to the conventional linear processing. The improvement of clutter reduction could be beneficial to various cardiac evaluation like myocardial perfusion or intra ventricular flow assessment.