Jochen F. Krücker

Rapid elastic image registration for 3-D ultrasound

A Subvolume-based algorithm for elastic Ultrasound REgistration (SURE) was developed and evaluated. Designed primarily to improve spatial resolution in three-dimensional compound imaging, the algorithm registers individual image volumes nonlinearly before combination into compound volumes. SURE works in one or two stages, optionally using MIAMI Fuse/spl copy/ software first to determine a global affine registration before iteratively dividing the volume into subvolumes and computing local rigid registrations in the second stage. Connectivity of the entire volume is ensured by global interpolation using thin-plate splines after each iteration. The performance of SURE was quantified in 20 synthetically deformed in vivo ultrasound volumes, and in two phantom scans, one of which was distorted at acquisition by placing an aberrating layer in the sound path. The aberrating layer was designed to induce beam aberrations reported for the female breast. Synthetic deformations of 1.5-2.5 mm were reduced by over 85% when SURE was applied to register the distorted image volumes with the original ones. Registration times were below 5 min on a 500-MHz CPU for an average data set size of 13MB. In the aberrated phantom scans, SURE reduced the average deformation between the two volumes from 1.01 to 0.30mm. This was a statistically significant (P=0.01) improvement over rigid and affine registration transformations, which produced reductions to 0.59 and 0.50 mm, respectively.

3D spatial compounding of ultrasound images using image-based nonrigid registration.

Medical ultrasound images are often distorted enough to significantly limit resolution during compounding (i.e., summation of images from multiple views). A new, volumetric image registration technique has been used successfully to enable high spatial resolution in three-dimensional (3D) spatial compounding of ultrasound images. Volumetric ultrasound data were acquired by scanning a linear matrix array probe in the elevational direction in a focal lesion phantom and in a breast in vivo. To obtain partly uncorrelated views, the volume of interest was scanned at five different transducer tilt angles separated by 4 degrees to 6 degrees. Pairs of separate views were registered by an automatic procedure based on a mutual information metric, using global full affine and thin-plate spline warping transformations. Registration accuracy was analyzed automatically in the phantom data, and manually in vivo, yielding average registration errors of 0.31 mm and 0.65 mm, respectively. In the vicinity of the warping control points, registrations obtained with warping transformations were significantly more accurate than full affine registrations. Compounded images displayed the expected reduction in speckle noise and increase in contrast-to-noise ratio (CNR), as well as better delineation of connective tissues and reduced shadowing. Compounding also revealed some apparent low contrast lobulations that were not visible in the single-sweep images. Given expected algorithmic and hardware enhancements, nonrigid, image-based registration shows great promise for reducing tissue motion and refraction artifacts in 3D spatial compounding.

Sound speed estimation using automatic ultrasound image registration

A mismatch between the sound speed assumed for beamforming and scan conversion and the true sound speed in the tissue to be imaged can lead to significant defocusing and some geometric distortions in ultrasound images. A method is presented for estimating the average sound speed based on detection of these distortions using automatic registration of overlapping, electronically steered images. An acrylamide gel phantom containing vaporized dodecafluoropentane droplets as point targets was constructed to evaluate the technique. Good agreement (rms deviation <0.4%) was found between the sound speeds measured in the phantom using a reference pulse-echo technique and the image-based sound speed estimates. A significant improvement in accuracy (rms deviation <0.1%) was achieved by including the simulated sound field of the probe rather than assuming straight acoustic beams and propagation according to ray acoustics.

Suspicious Breast Lesions: Assessment of 3D Doppler US Indexes for Classification in a Test Population and Fourfold Cross-Validation Scheme

PURPOSE To assess the diagnostic performance of various Doppler ultrasonographic (US) vascularity measures in conjunction with grayscale (GS) criteria in differentiating benign from malignant breast masses, by using histologic findings as the reference standard. MATERIALS AND METHODS Institutional Review Board and HIPAA standards were followed. Seventy-eight women (average age, 49 years; range, 26-70 years) scheduled for breast biopsy were included. Thirty-eight patient scans were partially analyzed and published previously, and 40 additional scans were used as a test set to evaluate previously determined classification indexes. In each patient, a series of color Doppler images was acquired and reconstructed into a volume encompassing a suspicious mass, identified by a radiologist-defined ellipsoid, in which six Doppler vascularity measures were calculated. Radiologist GS ratings and patient age were also recorded. Multivariable discrimination indexes derived from the learning set were applied blindly to the test set. Overall performance was also confirmed by using a fourfold cross-validation scheme on the entire population. RESULTS By using all cases (46 benign, 32 malignant), the area under the receiver operating characteristic curve (A(z)) values confirmed results of previous analyses: Speed-weighted pixel density (SWPD) performed the best as a diagnostic index, although statistical significance (P = .01) was demonstrated only with respect to the normalized power-weighted pixel density. In both learning and test sets, the three-variable index (SWPD-age-GS) displayed significantly better diagnostic performance (A(z) = 0.97) than did any single index or the one two-variable index (age-GS) that could be obtained without the data from the Doppler scan. Results of the cross validation confirmed the trends in the two data sets. CONCLUSION Quantitative Doppler US vascularity measurements considerably contribute to malignant breast tissue identification beyond subjective GS evaluation alone. The SWPD-age-GS index has high performance (A(z) = 0.97), regardless of incidental performance variations in its single variable components.

Fusion of digital mammography with breast ultrasound: a phantom study

The objective of this work was to acquire co-registered digital tomosynthesis mammograms and 3-D breast ultrasound images of breast phantoms. A prototype mammography compression paddle was built for this application and installed on an x-ray tomosynthesis prototype system (GE). Following x-ray exposure, an automated two-dimensional ultrasound probe mover assembly is precisely positioned above the compression plate, and an attached high-frequency ultrasound transducer is scanned over the acoustically coupled phantom or localized region of interest within the phantom through computerized control. The co-ordinate system of one of the two data sets is then transformed into that of the other, and matching regions of interest on either image set can be simultaneously viewed on the x-ray and ultrasound images thus enhancing qualitative visualization, localization and characterization of regions of interest. The potentials of structured noise reduction, cyst versus solid mass differentiation and full 3-D visualization of multi-modality registered data sets in a single automated combined examination are realized for the first time. Elements of system design and required image correction algorithms will be described and phantom studies with this prototype, automated system on an anthropomorphic breast phantom will be presented.

Rigid piston approximation for computing the transfer function and angular response of a fiber-optic hydrophone

The transfer function of a fiber-optic hydrophone (FOH) is computed for various fiber core radii. The hydrophone is modeled as a rigid disk, with plane waves impinging at normal or oblique incidence. The total sound field is written as the sum of the incident field and the field diffracted from the hydrophone. The diffracted field is approximated by the field generated by a vibrating planar piston in an infinite rigid baffle. For normal incidence and a pointlike fiber core, an analytical solution is presented. For finite fiber core radii, and for oblique incidence, the transfer functions are computed numerically. The calculated transfer functions exhibit an oscillatory frequency dependency that is most pronounced for small fiber cores. The solution for a core radius of 2.5 microm can be very well approximated by the analytical solution for a pointlike core at frequencies of up to 30 MHz. The results for normal incidence can be directly employed to deconvolute ultrasonic pressure signals measured with an FOH. From the transfer functions for oblique incidence, the angular response of the hydrophone is calculated. The angular response obtained here differs significantly from the model commonly used for piezoelectric hydrophones. The effective hydrophone radius derived from the angular response shows a strong frequency dependency. For low frequencies, it is found to be larger than the outer fiber radius, whereas it generally lies between the outer radius and the fiber core radius for frequencies above 10 MHz.

A tissue-mimicking ultrasound test object using droplet vaporization to create point targets

Ultrasound test objects containing reference point targets could be useful for evaluating ultrasound systems and phase aberration correction methods. Polyacrylamide gels containing albumin-stabilized droplets (3.6 μm mean diameter) of dodecafiuoropentane (DDFP) are being developed for this purpose. Perturbation by ultrasound causes spontaneous vaporization of the superheated droplets to form gas bubbles, a process termed acoustic droplet vaporization (ADV). The resulting bubbles (20 to 160 μm diameter) are small compared with acoustic wavelengths in diagnostic ultrasound and are theoretically suitable for use as point targets (phase errors <;20° for typical f-numbers). Bubbles distributed throughout the material are convenient for determining the point spread function in an imaging plane or volume. Cooling the gel causes condensation of the DDFP droplets, which may be useful for storage. Studying ADV in such viscoelastic media could provide insight into potential bioeffects from rapid bubble formation.

Sound speed estimation using ultrasound image registration

A method for estimating the average speed-of-sound (SOS) in tissue or other materials using automatic registration of electronically steered ultrasound images is presented. A mismatch between the SOS assumed for beam forming and image reconstruction and the actual SOS in the tissue to be imaged can lead to significant defocusing and geometric distortions of the image. The geometric distortions can be detected by registering multiple, overlapping images steered in different look directions. The average SOS in the field of view can be estimated from the geometric transformation required to register the images. The theoretical basis for SOS estimation using phased array and steered linear array transducers is derived. Simulations show a moderate sensitivity to noise, suggesting that SOS estimations can be accurate enough to improve image quality when used in the image formation process. Examples of SOS estimates in water and tissue-mimicking phantoms show promising results, with errors below 1%.

Evaluation of thin compression paddles for mammographically compatible ultrasound

We are developing a combined digital mammography/3D ultrasound system for breast cancer imaging to better detect and/or characterize breast lesions. Scanning a GE Logiq 9 M12L transducer array over a mammographic compression paddle/plate introduces an attenuating layer with sound speed and impedance different from that of tissue. This reduces signal level and affects beam focusing, Making the choice of a suitable paddle is essential for accurate sonographic detection of lesions. Similar work has been reported, but we present a more complete characterization of image quality through mammographic paddles of varying materials, (e.g., Lexan, Polyurethane, TPX, Mylar) and thicknesses. Quantitative measures such as spatial and contrast resolution, signal strength, and range lobe levels were compared to images without a paddle. In vivo patient studies compared images with standard handheld scans to images with 0.25, 1.0, and 2.5 mm thick paddles to examine restricted access problems, coupling issues, and overall lesion clarity. For mammography, filters were added to account for differences in X-ray transmission properties between the tested paddle and the standard mammography paddle. When lateral beamforming corrections were implemented to partially account for the speed of sound through the paddles, experiments conducted on 25 /spl mu/m line targets with several plastic paddles between 0.25-2.5 mm thick demonstrated image quality measures close to those with no paddle present. In some paddles <1.0 mm thick, a worst-case 5% reduction in linear spatial resolution and a maximum 4 dB signal loss averaged over 4 cm occurred. In those better paddles up to 2.5 mm thick, range lobe levels were consistently 35-40 dB lower than the signal maximum. Areas of restricted access (such as near the chest wall) were minimized by imaging in trapezoidal (virtual convex) format. TPX paddles <2.5 mm were the most ideal for ultrasound and mammogram imaging requirements and, after accounting for signal loss through the paddle, appearance of cysts was comparable to images obtained from handheld, direct contact sweeps.

Registration-based sound speed estimation in phantoms with acoustically vaporized droplets

A mismatch between the sound speed assumed for beam forming and scan conversion and the true sound speed in the field of view can lead to significant defocusing and some geometric distortions in ultrasound images. A method is presented for estimating the average sound speed based on detection of these distortions using automatic registration of overlapping, electronically steered images. An acrylamide gel phantom containing vaporized dodecafluoropentane (DDFP) droplets as point targets was constructed to evaluate the technique. Good agreement (errors <1 %) was found between the sound speeds measured in the phantom using a reference pulse-echo technique and the image-based sound speed estimates. A small improvement in accuracy was achieved by including the simulated sound field of the probe rather than assuming straight acoustic beams and refraction according to Snells law.