Morten Høgholm Pedersen

Clinical evaluation of chirp-coded excitation in medical ultrasound

Despite the enormous development in medical ultrasound (US) imaging over the last decades, penetration depth with satisfying image quality is often a problem in clinical practice. Coded excitation, used for years in radar techniques to increase signal-to-noise ratio (SNR), has recently been introduced in medical US scanning. In the present study, coded excitation using frequency-modulated US signals is implemented and evaluated in vivo.A total of nine male volunteers were scanned in three different abdominal locations, using both conventional pulsed and coded excitation. A modified scanner (B-K Medical model 3535) with transmitter and receiver boards developed in our group and a mechanical 4 MHz transducer were used. The system acquired coded and conventional US image frames interleaved, yielding identical acquisitions with the two techniques. Cine-loop sequences were evaluated by three experienced sonographers estimating penetration depth and scoring image quality of both conventional and coded imaging. The results showed a significant (p < 0.001) increase in penetration depth around 2 cm. Image quality was significantly (p < 0.001) better using codes at full usable depth and slightly, but also significantly (p < 0.05), better above depths, where the effect of coded excitation was noticeable to the sonographers. We conclude that the higher SNR offered by coded excitation gives improved image quality and provides increased penetration in medical US imaging. This increased SNR can, alternatively, be used to allow imaging at higher frequencies and thereby increase spatial resolution without any loss of penetration.

Clinical use and evaluation of coded excitation in B-mode images

Use of long encoded waveforms can be advantageous in ultrasound imaging, as long as the pulse compression mechanism ensures low range sidelobes and preserves both axial resolution and contrast. A coded excitation/compression scheme was previously presented by our group, which is based on a predistorted FM excitation and a mismatched compression filter designed for medical ultrasonic applications. The attenuation effect, analyzed in this paper using the ambiguity function and simulations, dictated the choice of the coded waveform. In this study clinical images, images of wire phantoms, and digital video demonstrate the applicability, clinical relevance, and improvement attained with the proposed scheme. A commercial scanner (B-K Medical 3535) was modified and interfaced to a software configurable transmitter board and to a sampling system with a 2 GB memory storage. The experimental system was programmed to allow alternating excitation on every second frame. That offers the possibility of direct comparison of the same set of image pairs; one with pulsed and one with encoded excitation. Abdominal clinical images from healthy volunteers were acquired and statistically analyzed by means of the auto-covariance matrix of the image data. The resolution laterally is retained, axially is improved, while there is a clear increase in penetration. Phantom images using the proposed FM-based scheme as well as complementary Golay codes were also acquired, in order to quantitatively evaluate the characteristics of the compressed output and its stability to attenuation. Images of a wire phantom in water show that the range sidelobes resulting from pulse compression are below the acoustic noise, which was at 50 dB for the pulsed images. For images acquired from an attenuation phantom, the proposed compression scheme was robust to frequency shifts resulting from attenuation. The range resolution is improved 12% by the coded scheme compared to a 2-cycle pulse excitation. For the maximum acquisition depth of 15 cm, where the coded excitations are primarily intended, the improvement in SNR was more than 10 dB, while the resolution was retained.

Preliminary in-vivo evaluation of convex array synthetic aperture imaging

Synthetic transmit aperture (STA) imaging has previously been investigated and compared to traditional imaging techniques in simulations and phantom studies. However, a full in-vivo study evaluating its clinical potential has yet to be conducted. This paper presents a preliminary in-vivo study of STA imaging in comparison to conventional imaging. The purpose is to evaluate whether STA imaging is feasible in-vivo, and whether the image quality obtained is comparable to traditional scanned imaging in terms of penetration depth, spatial resolution, contrast resolution, and artifacts. Acquisition was done using our RASMUS research scanner and a 5.5 MHz convex array transducer. STA imaging applies spherical wave emulation using multi-element subapertures and a 20 µs linear FM signal as excitation pulse. For conventional imaging a 64 element aperture was used in transmit and receive with a 1.5 cycle sinusoid excitation pulse. Conventional and STA images were acquired interleaved yielding ensuring exact same anatomical location. Image sequences were recorded in real-time, and processing was done offline. Male volunteers were scanned abdominally, and resulting images were compared by medical doctors using randomized blinded presentation. Penetration and image quality were scored and evaluated statistically. Results show that in-vivo imaging using STA imaging is feasible with improved image quality compared to conventional imaging.

Clinical comparison of pulse and chirp excitation

Coded excitation (CE) using frequency modulated signals (chirps) combined with modified matched filtering has earlier been presented showing promising results in simulations and in-vitro. In this study an experimental ultrasound system is evaluated in a clinical setting, where image sequences are assessed by skilled medical doctors. The effect on penetration depth and image quality were measured. A modified clinical scanner with a 4 MHz single element mechanical transducer, and external transmitter and receiver boards (RASMUS system) were used. The system allowed rapid toggling between chirp and short pulse excitation to simultaneously produce identical image sequences using both techniques. Nine healthy male volunteers were scanned in abdominal locations. All sequences were evaluated by 3 skilled medical doctors, blinded to each other and to the technique used. They assessed the depth (1) in which image quality decreased and (2) in which the image would be insufficient for clinical diagnosis. Furthermore they compared image quality in matching pairs of conventional and CE images. The average increase in penetration depth were almost 2 cm. Side-by-side comparison showed that coded image quality was consistently rated better; significant (p/spl les/0.05) when images were cropped at minimum the depth for good image quality and highly significant (p<0.001) when cropped at maximum depth sufficient for clinical diagnosis. We conclude that coded excitation with linear FM chirps improves penetration and image quality significantly in a clinical setting.

OC095: Synthetic aperture imaging in medical ultrasound

After pituitary suppression, the ovarian vascularization/follicle was lower in PCO-patients. The follicles in polycystic ovaries required much less FSH stimulation to acquire the same level of vascularization than the follicles in normal ovaries. Also hCG induced an increase in the follicular vascularization in both normal and polycystic ovaries. The follicle count correlated with the total vascularized volume in the ovaries throughout the IVF-cycle. Conclusions: Follicles in polycystic ovaries seem to be less vascularized than the follicles in normal ovaries after GnRHtreatment. It is possible that restricted blood supply to the follicles in PCO might be associated with the follicular arrest observed with PCO. We could confirm that follicles in PCO are more sensitive to gonadotrophin stimulation than follicles in normal ovaries.

OC095: Synthetic aperture imaging in medical ultrasound: Oral communication abstracts

After pituitary suppression, the ovarian vascularization/follicle was lower in PCO-patients. The follicles in polycystic ovaries required much less FSH stimulation to acquire the same level of vascularization than the follicles in normal ovaries. Also hCG induced an increase in the follicular vascularization in both normal and polycystic ovaries. The follicle count correlated with the total vascularized volume in the ovaries throughout the IVF-cycle. Conclusions: Follicles in polycystic ovaries seem to be less vascularized than the follicles in normal ovaries after GnRHtreatment. It is possible that restricted blood supply to the follicles in PCO might be associated with the follicular arrest observed with PCO. We could confirm that follicles in PCO are more sensitive to gonadotrophin stimulation than follicles in normal ovaries.

Synthetic Aperture Imaging in Medical Ultrasound

After pituitary suppression, the ovarian vascularization/follicle was lower in PCO-patients. The follicles in polycystic ovaries required much less FSH stimulation to acquire the same level of vascularization than the follicles in normal ovaries. Also hCG induced an increase in the follicular vascularization in both normal and polycystic ovaries. The follicle count correlated with the total vascularized volume in the ovaries throughout the IVF-cycle. Conclusions: Follicles in polycystic ovaries seem to be less vascularized than the follicles in normal ovaries after GnRHtreatment. It is possible that restricted blood supply to the follicles in PCO might be associated with the follicular arrest observed with PCO. We could confirm that follicles in PCO are more sensitive to gonadotrophin stimulation than follicles in normal ovaries.