M.M. Voormolen

Comparison of fundamental, second harmonic, and superharmonic imaging: a simulation study

In medical ultrasound, fundamental imaging (FI) uses the reflected echoes from the same spectral band as that of the emitted pulse. The transmission frequency determines the trade-off between penetration depth and spatial resolution. Tissue harmonic imaging (THI) employs the second harmonic of the emitted frequency band to construct images. Recently, superharmonic imaging (SHI) has been introduced, which uses the third to the fifth (super) harmonics. The harmonic level is determined by two competing phenomena: nonlinear propagation and frequency dependent attenuation. Thus, the transmission frequency yielding the optimal trade-off between the spatial resolution and the penetration depth differs for THI and SHI. This paper quantitatively compares the concepts of fundamental, second harmonic, and superharmonic echocardiography at their optimal transmission frequencies. Forward propagation is modeled using a 3D-KZK implementation and the iterative nonlinear contrast source (INCS) method. Backpropagation is assumed to be linear. Results show that the fundamental lateral beamwidth is the narrowest at focus, while the superharmonic one is narrower outside the focus. The lateral superharmonic roll-off exceeds the fundamental and second harmonic roll-off. Also, the axial resolution of SHI exceeds that of FI and THI. The far-field pulse-echo superharmonic pressure is lower than that of the fundamental and second harmonic. SHI appears suited for echocardiography and is expected to improve its image quality at the cost of a slight reduction in depth-of-field.

Harmonic 3-d echocardiography with a fast-rotating ultrasound transducer

Although the advantages of three-dimensional (3-D) echocardiography have been acknowledged, its application for routine diagnosis is still very limited. This is mainly due to the relatively long acquisition time. Only recently has this problem been addressed with the introduction of new real-time 3-D echo systems. This paper describes the design, characteristics, and capabilities of an alternative concept for rapid 3-D echocardiographic recordings. The presented fast-rotating ultrasound (FRU) transducer is based on a 64-element phased array that rotates with a maximum speed of 8 Hz (480 rpm). The large bandwidth of the FRU-transducer makes it highly suitable for tissue and contrast harmonic imaging. The transducer presents itself as a conventional phased-array transducer; therefore, it is easily implemented on existing 2-D echo systems, without additional interfacing. The capabilities of the FRU-transducer are illustrated with in-vitro volume measurements, harmonic imaging in combination with a contrast agent, and a preliminary clinical study

A new transducer for 3D harmonic imaging

We developed a fast rotating ultrasound transducer for 3D-echocardiography containing a broadband linear array. In the harmonic mode the tissue to clutter ratio was measured as function of frequency. This resulted in an optimal harmonic transmit frequency of 2.0 MHz. In-vitro evaluation, using the optimal frequency, showed an average volume error, for 5 reconstructed agar phantoms, of approximately 1%. Further, a clinical case study proved the feasibility of dynamic left ventricular volume measurements for the transducer.

Coherent compounding in doppler imaging

Coherent compounding can provide high frame rates and wide regions of interest for imaging of blood flow. However, motion will cause out-of-phase summation, potentially causing image degradation. In this work the impact of blood motion on SNR and the accuracy of Doppler velocity estimates are investigated. A simplified model for the compounded Doppler signal is proposed. The model is used to show that coherent compounding acts as a low-pass filter on the coherent compounding Doppler signal, resulting in negatively biased velocity estimates. Simulations and flow phantom experiments are used to quantify the bias and Doppler SNR for different velocities and beam-to-flow (BTF) angles. It is shown that the bias in the mean velocity increases with increasing beam-to-flow angle and/or blood velocity, whereas the SNR decreases; losses up to 4 dB were observed in the investigated scenarios. Further, a 2-D motion correction scheme is proposed based on multi-angle vector Doppler velocity estimates. For a velocity of 1.1 vNyq and a BTF angle of 75°, the bias was reduced from 30% to less than 4% in simulations. The motion correction scheme was also applied to flow phantom and in vivo recordings, in both cases resulting in a substantially reduced mean velocity bias and an SNR less dependent on blood velocity and direction.

P2A-6 Automatic Segmentation of the Left Ventricle in 3D Echocardiography Using Active Appearance Models

Assessment of left ventricular (LV) functional parameters, such as LV volume, ejection fraction and stroke volume, from real-time 3D echocardiography (RT3DE) is labor intensive and subjective, because in current analyses it requires input from the user. Automating these procedures will save valuable time in the analysis and will remove interobserver variability. We propose a fully automatic segmentation approach for the left ventricle in real-time 3D echocardiography, based on active appearance models (AAMs), using ultrasound specific grey value normalization. We evaluated shape and texture model generalization. Also, automatic segmentation has been preliminarily evaluated on transthoracic, apical acquisitions of 54 patients, acquired with the fast rotating ultrasound (FRU-) transducer (18 patients) and with the Philips Sonos 7500 (36 patients). The evaluations were done in a leave-N-out manner (with N=5). We evaluated point-to-surface (P2S) distances for the segmented endocardial contours to the expert manual contours. The generalization of the shape model was good, but texture model generalization was moderate, hampering the AAM matching We found preliminary segmentation errors (P2S) of 3.9plusmn 1.6 mm (N=54) for detection using AAM matching These results indicate that fully automatic segmentation of the LV in RT3DE using AAMs is feasible.

Improving 3D active appearance model segmentation of the left ventricle with Jacobian tuning

Automated image processing techniques may prove invaluable in the examination of real-time three-dimensional echocardiograms, by providing quantitative and objective measurements of functional parameters such as left ventricular (LV) volume and ejection fraction. In this study, we investigate the use of active appearance models (AAMs) for automatic detection of left ventricular endocardial contours. AAMs are especially useful in segmenting ultrasound images, due to their ability to model the typical LV appearance. However, since only a limited number of images is available for training, the model may be incapable of capturing the large variability in ultrasound image appearance. This may cause standard AAM matching procedures to fail if the model and image are significantly different. Recently, a Jacobian-tuning method for AAM matching was proposed, which allowed the models training matrix to adapt to the new, unseen image. This may potentially result in a more robust matching. To compare both matching methods, AAMs were built with end-diastolic images from 54 patients. Larger capture ranges and higher accuracy were obtained when the new method was used. In conclusion, this method has great potential for segmentation in echocardiograms and will improve the assessment of LV functional parameters.

Improved spatiotemporal voxel space interpolation for 3D echocardiography with irregular sampling and multibeat fusion

We developed a novel multi-beat image fusion technique using a special spatiotemporal interpolation for sparse, irregularly sampled data (ISI). It is applied to irregularly distributed 3D cardiac ultrasound data acquired with the fast rotating ultrasound (FRU) transducer developed in our laboratory, a phased array rotating mechanically at very high speed (240-480rpm). High-quality 2D images are acquired at ~100 frames/s over 5-10 seconds. ISI was compared quantitatively to spatiotemporal nearest neighbor interpolation (STNI) on synthetic data and compared qualitatively to classical trilinear voxel interpolation on 10 in-vivo cardiac image sets. ISI showed considerably lower absolute distance errors than STNI. For in-vivo images, ISI voxel sets showed reduced motion artifacts, suppression of noise and interpolation artifacts and better delineation of endocardium. In conclusion, ISI improves the quality of 3D+T images acquired with a fast rotating transducer in simulated and in-vivo data.

Quantitative harmonic 3D echocardiography with a fast rotating ultrasound transducer

Left ventricular (LV) volume and function measurement is the most common clinical referral question in the echocardiography laboratory. A fast, practical and accurate method would facilitate access to this important diagnostic information. We developed a fast rotating phased array transducer for 3D harmonic imaging of the heart, which makes fast and accurate LV volume and function measurement feasible. In this manuscript the implementation and validation of the data processing is described.

Superharmonic imaging based on chirps

In medical ultrasound harmonic images of biological tissue are commonly obtained by analyzing the reflected echoes from the 2nd harmonic band. A new modality dubbed super-harmonic imaging (SHI) targets a combination of the 3rd–5th harmonics. SHI is expected to yield enhanced spatial resolution and thus to increase the quality of echographic images. On the other hand, those images obtained using short imaging pulses are susceptible to so-called multiple axial reflection artifacts, stemming from the troughs in between harmonics in the frequency domain. The recently proposed dual-pulse frequency compounding method suppresses these artifacts but reduces the frame rate by a factor of 2. In this work we research the feasibility of employing a chirp protocol to perform SHI without compromising the frame rate. The chirp protocol was implemented using an interleaved phased array transducer (44 elements tuned at 1 MHz, 44 elements at 3.7 MHz) in combination with a fully programmable ultrasound system. The transducer was mounted in the side of a water-filled tank. Linear chirps with a center frequency of 1 MHz and a bandwidth of 40% were used as excitation pulses. Radio frequency traces were recorded at the focal plane along the lateral axis using a hydrophone, filtered over the superharmonic band and convolved with a decoding signal to obtain point spread functions (PSFs). The decoding signal was acquired by simulating the emitted beam using the KZK method for a rectangular aperture. The decoded superharmonic chirp had an SNR of 35–40 dB. Comparing to a the 3rd harmonic produced by a 2.5 cycle 1 MHz Gaussian apodized sine burst transmission the lateral beam width of the superharmonic chirp signal is 0.8 and 0.9 times that of the 3rd harmonic at the −6 dB and −20 dB levels respectively. Regarding the axial beam width, the superharmonic chirp signal has 0.9 and 0.8 times the axial beam width of the 3rd harmonic at the −6 dB and −20 dB levels respectively. The superharmonic chirp PSF is virtually free from imaging artifacts. Based on the SNR measurements the chirp protocol yields a sufficient dynamic range. The PSF has increased spatial resolution in comparison with the 3rd harmonic. The first in-vitro images show promise, but the decoding pulse requires improvement.

On the accuracy of coherent compounding Doppler imaging

The influence of blood motion on the accuracy of coherent compounding Doppler imaging is investigated. A simplified model for the compounded Doppler signal is proposed. The model is used to show that coherent compounding acts as a low pass filter, inferring a negative shift of the Doppler spectrum and a bias in mean frequency estimates. Simulations and in vitro experiments were used to quantify the bias and signal-to-noise ratio (SNR) for different velocities and beam-to-flow angles. Further, a multi-angle vector Doppler approach is proposed for 2D correction of blood motion during coherent compounding, which was verified in simulations and in vitro to improve SNR and substantially reduce the mean velocity bias.