Torbjørn Hergum

3-D Speckle Tracking for Assessment of Regional Left Ventricular Function

Speckle tracking in 2-D ultrasound images has become an established tool for assessment of left ventricular function. The recent development of ultrasound systems with capability to acquire real-time full volume data of the left ventricle makes it possible to perform speckle tracking in three dimensions, and thereby track the real motion of the myocardium. This paper presents a method for assessing local strain and rotation from 3-D speckle tracking in apical full-volume datasets. The method has been tested on simulated ultrasound data based on a computer model of the left ventricle, and on patients with myocardial infarction. When applied on simulated ultrasound data, the method showed good agreement with strain and rotation traces calculated from the reference motion, and the method was able to capture segmental differences in the deformation pattern, although the magnitudes of strains were systematically lower than the reference strains. When applied on patients, the method demonstrated reduced strain in the infarcted areas. Bulls-eye plots of regional strains showed good correspondence with wall motion scoring based on 2-D apical images, although the dyskinetic and hypokinetic regions were not apparent in all strain components.

Parallel beamforming using synthetic transmit beams

Parallel beamforming is frequently used to increase the acquisition rate of medical ultrasound imaging. However, such imaging systems will not be spatially shift invariant due to significant variation across adjacent beams. This paper investigates a few methods of parallel beamforming that aims at eliminating this flaw and restoring the shift invariance property. The beam-to-beam variations occur because the transmit and receive beams are not aligned. The underlying idea of the main method presented here is to generate additional synthetic transmit beams (STB) through interpolation of the received, unfocused signal at each array element prior to beamforming. Now each of the parallel receive beams can be aligned perfectly with a transmit beam - synthetic or real - thus eliminating the distortion caused by misalignment. The proposed method was compared to the other compensation methods through a simulation study based on the ultrasound simulation software Field II. The results have been verified with in vitro experiments. The simulations were done with parameters similar to a standard cardiac examination with two parallel receive beams and a transmit-line spacing corresponding to the Rayleigh criterion, wavelength times f-number (ulambdamiddotf#). From the results presented, it is clear that straightforward parallel beamforming reduces the spatial shift invariance property of an ultrasound imaging system. The proposed method of using synthetic transmit beams seems to restore this important property, enabling higher acquisition rates without loss of image quality

Fast ultrasound imaging simulation in K-space

Most available ultrasound imaging simulation methods are based on the spatial impulse response approach. The execution speed of such a simulation is of the order of days for one heart-sized frame using desktop computers. For some applications, the accuracy of such rigorous simulation approaches is not necessary. This work outlines a much faster 3-D ultrasound imaging simulation approach that can be applied to tasks like simulating 3-D ultrasound images for speckle-tracking. The increased speed of the proposed simulation method is based primarily on the approximation that the point spread function is set to be spatially invariant, which is a reasonably good approximation when using polar coordinates for simulating images from phased arrays with constant aperture. Ultrasound images are found as the convolution of the PSF and an object of sparsely distributed scatterers. The scatterers are passed through an anti-aliasing filter before insertion into a regular beam-space grid to reduce the bandwidth and significantly reduce the amount of data. A comparison with the well-established simulation software package field II has been made. A simulation of a cyst image using the same input object was found to be in the order of 7000 times slower than the presented method. Following these considerations, the proposed simulation method can be a rapid and valuable tool for working with 3-D ultrasound imaging and in particular 3-D speckle-tracking.

Quantification of mitral regurgitation using high pulse repetition frequency three-dimensional color Doppler.

BACKGROUND The aim of this study was to validate a novel method of determining vena contracta area (VCA) and quantifying mitral regurgitation using multibeam high-pulse repetition frequency (HPRF) color Doppler. METHODS The Doppler signal was isolated from the regurgitant jet, and VCA was found by summing the Doppler power from multiple beams within the vena contracta region, where calibration was done with a reference beam. In 27 patients, regurgitant volume was calculated as the product of VCA and the velocity-time integral of the regurgitant jet, measured by continuous-wave Doppler, and compared with regurgitant volume measured by magnetic resonance imaging (MRI). RESULTS Spearmans rank correlation and the 95% limits of agreement between regurgitant volume measured by MRI and by multibeam HPRF color Doppler were r(s) = 0.82 and -3.0 +/- 26.2 mL, respectively. CONCLUSION For moderate to severe mitral regurgitation, there was good agreement between MRI and multibeam HPRF color Doppler. Agreement was lower in mild regurgitation.

The effect of including myocardial anisotropy in simulated ultrasound images of the heart

We have examined the effect of incorporating tissue anisotropy in simulated ultrasound images of the heart. In simulation studies, the cardiac muscle (myocardium) is usually modeled as a cloud of uncorrelated point scatterers. Although this approach successfully generates a realistic speckle pattern, it fails to reproduce any effects of image anisotropy seen in real ultrasound images. We hypothesize that some of this effect is caused by the varying orientation of anisotropic myocardial structures relative to the ultrasonic beam and that this can be taken into account in simulations by imposing an angle dependent correlation of the scatterer points. Ultrasound images of a porcine heart were obtained in vitro, and the dominating fiber directions were estimated from the insonification angles that gave rise to the highest backscatter intensities. A cylindrical sample of the myocardium was then modeled as a grid of point scatterers correlated in the principal directions of the muscle fibers, as determined experimentally. Ultrasound images of the model were simulated by using a fast k-space based convolution approach, and the results were compared with the in vitro recordings. The simulated images successfully reproduced the insonification dependent through-wall distribution of backscatter intensities in the myocardial sample, as well as a realistic speckle pattern.

Parallel beamforming using synthetic transmit beams [biomedical ultrasound imaging]

Ultrasound images generated using conventional parallel beamforming contains stationary stripes due to beam-to-beam variations across adjacent beams. Such an imaging system is thus shift variant. This paper investigates a method which eliminates this flaw and restores the shift invariant property. The beam-to-beam variations occur because the transmit and receive beams are not aligned. The underlying idea is then to generate additional synthetic transmit beams (STB) through interpolation of the received, unfocused signal at each array element prior to beamforming. Now each of the parallel receive beams can be aligned perfectly with a transmit beam-synthetic or real-thus eliminating the distortion caused by misalignment. To investigate the performance of the method a simulation study has been conducted. The simulations were done with parameters similar to a standard cardiac examination. Using the proposed method the variation was reduced to 1.5 dB. The simulations were confirmed by results front in vitro experiments where vertical line artifacts were unnoticeable when using the proposed method.

Comparison of the performance of different tools for fast simulation of ultrasound data.

Simulation of ultrasound data is often performed for developing new ultrasound data processing techniques. The spatial impulse response method (as implemented in FieldII) has typically been used as the gold standard due to its excellent accuracy in the linear domain. When scatterer numbers become significant and when 3D volumetric data sets need to be computed, calculation times can become an issue. In order to solve this problem, two alternative methods have recently been proposed both of which are based on the principle of convolving a set of point scatterers with a point spread function. ldquoFUSKrdquo operates in the frequency domain while ldquoCOLErdquo runs in the spatio-temporal domain. The aim of this study was to compare both methodologies in terms of accuracy and processing speed using FieldII as a reference. Hereto, images were generated from identical scattering distributions using FieldII, COLE and FUSK and were compared in terms of their first and second order statistical properties. Moreover, computation times were recorded and contrasted.

P3A-6 The Effect of Including Fiber Orientation in Simulated 3D Ultrasound Images of the Heart

We have examined the effect of muscle fiber orientation on simulated ultrasound images. In ultrasound simulation studies, the myocardium is usually modeled as a cloud of uncorrelated point scatterers. This approach successfully generates a fully developed speckle pattern, but it does not reproduce any effects of image anisotropy which is seen in vivo. We hypothesize that some of these effects are caused by the varying orientation of myocardium muscle fibers relative to the ultrasonic beam. Images of the septal wall of a pig heart were obtained in vitro, and the dominating fiber direction in each layer was found from the backscattered signal. A segment of the heart muscle was then modeled as a grid of point scatterers correlated in the direction of the muscle fibers, as determined experimentally. Ultrasound images of this model were simulated by using a k-space based convolution approach. The simulated images showed a line of increased intensity where the fiber direction was nearly perpendicular to the beam direction, similar to the experimental results. We conclude that inclusion of the orientation of the muscle fibers is important, in order to generate realistic ultrasound images

Quantification of aortic regurgitation using high-pulse repetition frequency three-dimensional colour Doppler

AIMS The aim of this study was to validate and assess the feasibility of a previously described method using multibeam high-pulse repetition frequency (HPRF) colour Doppler to quantify the vena contracta area (VCA) in aortic regurgitation (AR). METHODS Twenty-nine patients with mild to severe AR were studied. Regurgitant volume and fraction measured by magnetic resonance imaging (MRI) were used as the standard of reference. The VCA was measured automatically by combining the Doppler power from multiple beams with a priori knowledge of the individual beam profiles, to give an absolute measurement of the VCA. The regurgitant volume was calculated as the product of the VCA and the velocity time integral, measured separately by continuous wave Doppler. RESULTS The Spearmans rank correlation between regurgitant volume by MRI and multibeam HPRF colour Doppler was rs = 0.73 (P < 0.01), with 95% limits of agreement of -14.4 ± 29.1 mL. The mean difference between the methods in those with MRI regurgitant volume of ≥30 mL (n = 14) was -7.6 (95% confidence interval -13.9 to -1.2) mL. CONCLUSION There was good agreement between MRI and multibeam HPRF colour Doppler in patients with moderate to severe AR, while agreement for those with mild AR was modest.

Reducing color flow artifacts caused by parallel beamforming

In color flow imaging for medical diagnosis the inherent trade-off between frame rate and image quality may often lead to suboptimal images. Parallel receive beamforming is used to help overcome this problem, but this introduces artifacts in the images. In addition to the parallel beamforming artifacts found in B-mode imaging, we have found that a difference in curvature of transmit and receive beams gives a bias in the Doppler velocity estimates. This bias causes a discontinuity in the velocity estimates in color flow images.