Real time vector Doppler for tissue motion

Abstract

Tissue Doppler Imaging (TDI) can assess tissue motion in vascular and cardiac imaging. However, a major drawback of these measurements is that the motion estimation is limited to the component along the ultrasound beam axis. Cardiac and vessel wall motion studies have shown that complex three-dimensional motions can be observed, and that there is a clinical need to fully assess the three components of the vector motion. This work describes how TDI can be extended by acquiring a real time two-component velocity vector via a dual beam vector Doppler technique. A vector Doppler velocity estimator using a small interbeam angle can suffer from both bias and large variance. This estimator is also strongly dependent on the settings of the echographic system. To reduce the large bias and variance, most vector velocity techniques use a very large ensemble length (EL) (>20), which does not allow real time implementation. We propose a new processing technique, which reduces the bias and the standard deviation of the vector velocity estimate. The new method assumes that the vector velocity angle varies slowly over the cardiac cycle. The angle can then be estimated using a large time window. The performance of this new technique has been tested experimentally using a tissue mimicking rotating phantom. It is shown that the factors influencing the results are the EL, the precision of the TDI estimates and the time window. The results indicate that the variance and bias of velocity magnitude and orientation estimates decrease with increasing EL, increasing precision of the TDI estimates and increasing time window. Using an EL of 9, 8 bits for the velocity estimate, and an observation time of one second, a 5-degree bias of the angle estimate is observed, with a variance below 7 degree averaged over all angles. A 10% bias of the velocity magnitude is observed, with a variance of 1%. In conclusion, TDI can be improved with vector Doppler providing two-dimensional tissue motion estimation, enabling more accurate biomechanical tissue property assessment.