ULTRASONIC ADAPTIVE BEAMFORMING METHOD AND ITS APPLICATION FOR TRANSCRANIAL IMAGING

Abstract

A method for ultrasonic imaging a volume behind a barrier includes placing an array of elements of a probe adjacent the barrier. The local thickness and the sound speed in the barrier are estimated. A plurality of sub-apertures are defined in the array of elements, each sub-aperture defined by a different sub-set of adjacent ones of the array of elements. Ultrasonic waveforms are collected from each of the sub-apertures by focusing each sub-aperture approximately at the inner surface of the barrier. The ultrasonic waveforms through each of the sub-aperture elements are combined to generate a combined A-mode waveform. A profile of an inner surface of the barrier adjacent the probe is extracted based upon the combined A-mode waveforms. A correction is calculated based upon the profile of the inner surface. The volume behind the barrier is scanned using beamforming based upon the correction to generate an image of the volume.

Description

BACKGROUND

Noninvasive ultrasonic imaging of static foreign objects through the human skull or any other highly scattering multilayered solid structure (acoustical barrier) is difficult. The acoustical barrier possesses different acoustic impedance than the interfacing media, so implementation of conventional phased array imaging techniques results in significant image quality degradation. Where the intact human skull is the barrier, the foreign objects could be pieces of bone fragments, pieces of shrapnel, air bubbles, blood clots with high reflectivity, etc., which have higher/lower acoustic impedance than the surrounding brain tissue. The objects can be lodged at different depths, ranging from immediately under the skull up to several centimeters deep in the brain tissue. The thickness of the skull varies unevenly around the skull, which further causes image distortion.

SUMMARY

The present method resolves accurate location of static objects and structures hidden under the human skull or behind any other highly scattering multilayered solid structure (acoustical barrier). The acoustical barrier should possess different acoustic impedance than the interfacing media and thickness greater than or comparable to the wavelength of transmitted acoustical field. The images are obtained noninvasively through the skull/barrier by a single, small-footprint, handheld matrix or linear ultrasonic phased array probe.

The sound speed in the skull bone or barrier material and the profile of the skull/barrier segment under the probe are measured as follows:

• • 1. If allowed by the driving electronics and by the probe, switch to a higher-frequency excitation signal to improve the in-depth resolution of the skull profile.
  • 2. Simultaneously estimate the local thickness and the sound speed in the skull bone or barrier using a known algorithm, such as an iterative focusing algorithm or similar.
  • 3. Using a small sub-aperture comprised of several adjacent elements of the matrix/linear probe, approximately focused at depth of about one skull/barrier thickness under the probe's surface (on the inner surface of the skull/barrier), collect an A-mode waveform containing the reflections from the front and inner surfaces of the skull/barrier. Such waveform is obtained via standard beamforming and summation of RF signals over the sub-aperture elements, and then optionally enhanced by averaging over multiple acquisitions.
  • 4. Collect a sequence of said acoustic A-mode waveforms over the grid of points on the inner surface of the skull/barrier by moving the said small aperture over the probe in a raster scanning (or any other desired) manner. The points in said grid may be distributed unevenly, e.g. sparser where the surface is smooth and denser elsewhere.
  • 5. Extract the profile of the skull/barrier layer under the matrix probe by a cross-correlation technique or any other time delay estimation algorithm.
  • 6. Smooth and oversample the obtained skull/barrier profile by means of interpolation or any other appropriate algorithm.
  • 7. If higher frequency was used at the step 1, switch back to the lower-frequency excitation signal to optimize the trade-off between skull penetration and spatial resolution.

Alternatively, the skull/barrier profile and acoustical properties can be obtained not with the same phased array probe as the one used for beamforming and imaging purposes. For example, these properties can be obtained from a separate acoustic measurement with a different (e.g. higher-frequency) probe, from X-Ray Computed Tomography, Magnetic Resonance Imaging etc.

As another alternative, the local thickness and the sound speed in the skull bone or barrier can be estimated separately (i.e. not simultaneously), using available tools.

By accurately tracing refracted acoustic rays between individual array elements and the given intended focal point through the bone/barrier and the media behind it, the system calculates time delays for the elements of transmit and receive apertures that would allow to achieve accurate single-point focusing in both transmission (active) and reception (passive) modes.

The system and method implement beamforming algorithms both in transmission and reception modes using timing sequences calculated above that are dynamically adapted to the measured local properties and geometry of the skull/barrier to correct for the phase aberration and refraction effects. The algorithms implement spatial apodization to produce optimized directivity pattern of the acoustic field, i.e. enhanced main lobe and suppressed side lobes, in both transmission and reflection modes.

The system and method combine the results of adaptive active and passive focusing at a prescribed number of points distributed in the imaged volume to generate an image of said volume with objects and structures showing at their correct locations.

The system and method may represent the results as interactive 3D maps in one of the common data storage formats (e.g. DICOM), which can be output to an image viewer.

The tracing of the refracted acoustic rays between individual array elements and the intended focal point through the bone/barrier and the media behind it may be accomplished by solving the Helmholtz wave propagation equation numerically in each medium with appropriate boundary conditions. For this, the algorithm searches through the boundary points and picks the one at which all constraints, including the Fermat's principle and boundary conditions, are satisfied. Therefore the ultrasound transmitted from the array element under consideration to the desired focal point in the brain will pass through this point on the skull-brain interface. For optimum results, more points may be interpolated in between the original grid points on highly undulating areas of the inner boundary. On the other hand, when the thickness profile has a smooth variation, some boundary points may be skipped. The above process may then be repeated for all the active array elements to find all the refracted paths.

The active and passive beamforming algorithms are not limited to those used in the preferred embodiment, namely, those using particular spatial and temporal apodization and frequency modulation to minimize the side lobes.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows an imaging system according to one embodiment of the present invention.

FIG. 2 shows an arrangement of the elements and their scanning path used for the profile extraction via the matrix probe.

FIG. 3 is a flowchart of the one embodiment for skull profile extraction.

FIG. 4 is a flowchart of an example single-point adaptive focusing method.

FIG. 5 shows example contour plots of the boundary points which satisfy the first two boundary conditions when seeking for the refracted path between one particular element of the matrix array and one particular intended focal point in a single-point adaptive focusing scenario.

FIG. 6 shows a configurational imaging setup for the adaptive imaging method described herein.

FIG. 7 shows a sample 3D sonogram showing high-impedance static objects through intact human skull.

DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT

An imaging system 10 according to one embodiment of the present invention is shown schematically in FIG. 1. The imaging system 10 generally includes an ultrasonic phased array probe 12, such as a single, small-footprint, handheld matrix or linear ultrasonic phased array probe. The probe 12 is connected to a phased array controller 11 capable of performing on-the-flight adaptive time delay correction and other functions required by the algorithm. The phased array controller 11 includes management software running on a controller 13 (with processor, memory, etc), which governs the system's operation in both transmission (TX) and reception (RX) modes. In transmission, the adjustable parameters include the frequency, amplitude, and shape of the excitation signal, the arrangement of elements in the transmit aperture, and the spatial apodization over the transmit aperture. The active TX beamformer 14 applies both standard and adapted-to-the-barrier focal rules to the elements of the transmit aperture of the probe 12 to correct for the barrier-induced phase aberration of transmitted signals. In reception, the received echo signals pass through RX control 15, including an optional multiplexer, pre-amplifier, AD converter receiving the amplified signal, and demodulator. The digitized echo signals are modulated and saved to memory. The passive RX beamformer 16 then adjusts the individual delays of the echo signals for all elements of the receive aperture of the probe 12 to perform passive focusing and to correct for phase aberration due to the barrier. Signal and image processing routines are applied to the beamformed signals by signal/image processing unit 17 to enhance image quality, mainly in terms of resolution and contrast, prior to displaying the final sonogram. The phased array controller is in turn connected to a computer having at least one processor communicating with storage (such as electronic, magnetic, optical or other storage). Typical user interface hardware may be provided, such as a display 20, user input devices, etc. The computer carries high-level instructions to the phased array controller, receives the data from it, runs data post processing, visualization, and manipulation software, as well as controls the storage of the raw and processed data.

The probe 12 is shown adjacent the head of a patient 30 for imaging the patient's brain 32. Between the probe 12 and the brain 32 is skin 34 and skull 36. The system 10 illustrated in FIG. 1 can be used, for example, to detect, locate and/or image foreign objects 38 through the skull 36 by combining the transmission and reception ultrasonic modes, as explained below.

The system 10 performs a noninvasive ultrasonic adaptive beamforming method for imaging of static foreign objects 38 through the human skull 36 or any other highly scattering multilayered solid structure (acoustical barrier). The acoustical barrier should possess different acoustic impedance than the interfacing media. In presence of such barrier, implementation of conventional phased array imaging techniques results in significant image quality degradation. In case of the intact human skull 36 as the barrier, the foreign objects 38 could be pieces of bone fragments, pieces of shrapnel, air bubbles, blood clots with high reflectivity, etc., which have higher/lower acoustic impedance than the surrounding brain tissue 32. The objects 38 can be lodged at different depths, ranging from immediately under the skull 36 up to several centimeters deep in the brain tissue 32. The method uses an ultrasonic phased array 12 contacting the skull 36 either directly or through an intermediate layer (e.g. human head skin 34). An ultrasonic coupling agent can be used to optimize acoustic contact between the array and the skull 36 (or skin 34).

The adaptive beamforming method is based on the knowledge of the geometry and acoustical properties of the barrier (e.g. skull 36) segment in contact with the phased array 12. In one embodiment, the characteristics of the barrier (mainly longitudinal sound speed and thickness profile) are determined from the preliminary reflection-mode acoustic measurements. Such measurements are conducted prior to the main beamforming steps and may be done according to the techniques described in A. Wydra, E. Malyarenko, et al., “Development of a practical ultrasonic approach for simultaneous measurement of the thickness and the sound speed in human skull bones: a laboratory phantom study,” Phys. Med. Biol., vol. 58, no. 4, pp. 1083-102, February 2013, which is hereby incorporated by reference in its entirety. The measurements may be implemented via the same phased array 12 that is further used in the imaging process. Alternatively, these characteristics can be also obtained from other measurement modalities, including X-Ray CT images, independent ultrasonic measurements, literature data, or any other suitable means.

The adaptive beamforming method includes two separate modes, transmission (Tx) and reception (Rx), the merging of which results in obtaining a non-distorted image of the targeted objects 38. Each of the two modes includes a set of consecutive single-point adaptive focusing steps. Transmission adaptive focusing has an active nature while reception adaptive focusing is performed passively. The term adaptive focusing means adaptation of the phased array 12 timing pattern to the properties of the underlying barrier segment in order to refocus/redirect the acoustic beam to intended focal coordinates. To achieve accurate focusing, the new timing pattern should account for acoustic impedance mismatch at the barrier surfaces as well as attenuation in the propagation media. The corrected timing patterns are calculated at both transmission and reception modes and applied to the elements of the array together with proper spatial apodization rules. The final stage includes post-processing of the received signals and image reconstruction of the targeted objects showing their correct positions. Below, the above steps are explained in detail.

1. Extraction of the Skull/Barrier Profile

FIG. 2 illustrates the matrix of elements 44 of the probe 12. A higher or lower number of elements 44 could be used. FIG. 3 is a flowchart of the one embodiment for skull profile extraction.

Referring to FIGS. 2 and 3, in one embodiment, to extract the skull 36 profile via the same imaging matrix probe 12 prior to the main beamforming process, first the longitudinal sound speed in the skull layer 36 is measured, such as by using a simultaneous thickness/speed measurement technique in step 50. In step 52, a small sub-aperture 46 of the matrix probe elements 44 at a time is arranged to focus at about the skull layer's 36 thickness, e.g. ˜1 cm distance from the probe's surface. In FIG. 2, the sub-aperture 46 comprises nine elements 44, but the number of elements could be higher or lower. Starting from one corner of the array, this small aperture 46 is moved in single element steps to the opposite corner of the matrix probe 12 in a raster scanning manner (or other pattern). The line in FIG. 2 shows one possible scanning path of the center element 44 of the nine-element aperture 46. At each step, multiple acquisitions of the RF signals are recorded, averaged and fed through standard nine-element beamforming and summation processes in step 54. The sub aperture 46 is then incremented along the path by one element in step 56. The profile of the skull layer in front of the matrix probe is then extracted by a cross-correlation algorithm in step 58.

2. Single-Point Adaptive Focusing

The skull profile extraction algorithm comprised of steps 50-58 outputs the inner surface profile needed by the adaptive focusing algorithm. As the matrix probe 12 is placed in contact with the skull layer 36, the skull profile extraction algorithm is executed in its entirety in step 60. Step 62 is a check whether this execution went without errors and produced a valid skull profile or an error occurred. An error means the algorithm has failed to detect a valid skull profile due to poor contact between the probe and the skull, due to extremely rough inner surface of the skull, extreme porosity at this location, or for other reasons. In case of an error, step 64 directs moving the probe to a different location and repeating from step 60 until a valid skull profile is detected and check at step 64 switches the execution to step 66.

To implement the outlined beamforming method, a ray-tracing computer model has been developed to simulate acoustical beam patterns in the skull-brain propagation environment followed by calculation of the adapted timing patterns. This model incorporates the effects of phase aberration and attenuation by the acoustic barrier and by the rest of the propagation media (e.g. by the human skull and the brain tissue). Based on the measured variable thickness of the skull segment under the phased array and the acoustic impedance mismatch between skull and brain tissue, the refracted acoustic ray paths connecting active array elements with each intended focal area are numerically calculated through the computer model. For this, first the extracted skull profile is discretized with a mesh-grid of boundary points in step 66. The grid size is automatically chosen based on the thickness gradient of the extracted skull profile. In step 68, normal and gradient vectors are calculated over the inner boundary of the extracted skull layer at each grid point. The normal and gradient vectors fields are then output to the main body of the ray-tracing code.

To find physically possible refracted paths from each active array element to a desired focal point in step 72, the Helmholtz wave propagation equation:

$\frac{1}{r^2}\frac{\partial }{\partial r}\left(r^2\frac{\partial \Phi }{\partial x}\right)+\frac{1}{r^2\sin \theta }\frac{\partial }{\partial \theta }\left(\sin \theta \frac{\partial \Phi }{\partial \theta }\right)+\frac{1}{r^2{\sin }^2\theta }\frac{{\partial }^2\Phi }{\partial {\varphi }^2}+k^2\Phi \left(r,\theta ,\varphi \text{:}\omega \right)=0$

is solved numerically in each medium with appropriate boundary conditions in step 70. For this, the algorithm searches through the boundary points and picks the one at which all constraints are satisfied. In the proposed algorithm, there are three major constraints to be satisfied:

• • 1. The normal vector at each grid point should be perpendicular to the norm of the communal plane of incident and refracted beams;
  • 2. The intracranial sound pressure must be equal to the normal component of the stress just within the skull layer; the tangential stress in the skull must vanish, and the normal component of the velocity must be continuous;
  • 3. Fermat's principle.

In the computer model, first contour plots of the skull-brain boundary points which satisfy the first two boundary conditions are calculated. Example of such contour plots are shown in FIG. 5. The solid curves are contour-plots of the boundary points at which the first constraint is satisfied; and the dotted lines are contour-plots of the boundary points at which the first constraint is satisfied. The intersection points of the two plots satisfy both constraints. These points are identified by small black circles in FIG. 5. Among these selected points, the one that will correspond to the shortest refracted path between the selected array element and the intended focal point is selected based on the Fermat's principle. This point is shown by a small cross in FIG. 5. Therefore the ultrasound transmitted from the array element under consideration to the desired focal point in the brain will pass through this point on the skull-brain interface.

For optimum results, more points are interpolated in between the original grid points on highly undulating areas of the inner boundary; on the other hand, when the thickness profile has a smooth variation, some boundary points may be skipped. The above process is then repeated for all the active elements on the array to find all the refracted paths.

With the knowledge of these refracted paths for each single-point focusing scenario, in step 74 the model calculates the corrected time delay pattern for the array elements through

$\begin{array}{cc}{\tau }_{\mathrm{nm}}=\frac{1}{c_0}\left[\begin{array}{c}\sqrt{\begin{array}{c}{R}_F^2+\frac{d^2}{4}\left[{\left(N-1\right)}^2+{\left(M-1\right)}^2\right]+\\ \left[\left(N-1\right)\cos \phi +\left(M-1\right)\sin \phi \right]{\mathrm{dR}}_F\cos \Psi \end{array}-}\\ \sqrt{R_F^2+\left(n^2+m^2\right)d^2-2\left(n\cos \phi +m\sin \phi \right){\mathrm{dR}}_F\cos \Psi }\end{array}\right]& \left(2\right)\end{array}$

where the index n denotes the element number along x-axis for −(N−1)/2≦n≦(N−1)/2, N is the total number of elements in width of the array. The index m denotes the element number along the y-axis for −(M−1)/2≦m≦(M−1)/2, M is the total number of elements in length of the array. c₀ is the compression wave sound speed in propagation medium and d is the array's pitch. The adapted time delay pattern is sent to the beamformer in step 76.

3. Transmission and Reception Modes Adaptive Beamforming and Final Image Generation

The above single-point adaptive focusing method is used to create adapted timing patterns (focal rules) for the phased array elements in transmission mode. When a particular new focal rule is applied, the ultrasound beam produced by the transmitting aperture is directed towards the intended focal point behind the barrier. To optimally probe the entire target volume behind the barrier, multiple focal rules can be successively created to steer the acoustic beam at different directions and to focus at multiple points along each direction.

For each particular focal rule in the transmission mode it is necessary to collect acoustic echoes from the propagation media. This is done by appropriately configuring a selected group of array elements 44 (receiving aperture) and represents the reception mode. The receiving aperture can be either the same as the transmitting aperture or different. In the reception mode, the explained above single-point adaptive focusing algorithm is employed, in a similar way to the transmission mode, but adapted for the passive dynamic focusing nature of this mode. The beamforming algorithm is used again to calculate and apply corrected timing patterns adapted to the geometry and acoustical properties of the barrier segment in contact with the array. The adapted timing patterns are then applied to the receiving elements on the array during dynamic focusing in reception mode right after each transmission shot. The adapted transmission and reception modes are then combined in the final image reconstruction process, which includes post-processing and filtration of received signals to assign an acoustical intensity value to each pixel and generate a final image (sonogram). An imaging setup is shown in FIG. 6. The final sonogram shows the targeted objects 38 in their correct positions (FIG. 7).

In accordance with the provisions of the patent statutes and jurisprudence, exemplary configurations described above are considered to represent a preferred embodiment of the invention. However, it should be noted that the invention can be practiced otherwise than as specifically illustrated and described without departing from its spirit or scope. For example, it should be noted that although calculation and derivations are explained in case of a 2D probe for a 3D image generation, the disclosed here method can be applied to any other phased array geometry (e.g. 1D, 1.5D, conical, etc). The developed algorithm can be used for adaptive beamforming through any attenuative multilayered solid structure (acoustical barrier) having an acoustic impedance mismatch between the interfacing media. It can also be considered for large and synthetic aperture focusing.

Claims

1. A method for ultrasonic imaging a volume behind a barrier, the method including the steps of: a) placing an array of elements of a probe capable of transmitting and receiving ultrasound adjacent the barrier;b) estimating local thickness and sound speed in the barrier;c) defining a plurality of sub-apertures in the array of elements, each sub-aperture defined by a different sub-set of adjacent ones of the array of elements;d) collecting ultrasonic waveforms from each of the sub-apertures by focusing each sub-aperture approximately at the inner surface of the barrier based on the estimated thickness and estimated sound speed data obtained in said step b), the ultrasonic waveforms containing reflections from the front and inner surfaces of the barrier;e) combining the ultrasonic waveforms through each of the sub-aperture elements to generate a combined A-mode waveform for each of the sub-apertures;f) extracting a profile of an inner surface of the barrier adjacent the probe based upon the combined A-mode waveforms;g) calculating a correction based upon the profile of the inner surface, including calculating the correction for each element of the sub-apertures by tracing refracted acoustic rays between said each array element and an intended focal point through the barrier and the volume behind the barrier, wherein the corrections are in the form of modified time delays to achieve accurate single-point focusing; andh) scanning the volume behind the barrier using beamforming based upon the correction to generate an image of the volume.

2. The method of claim 1 wherein the correction is a corrected timing pattern based upon the profile of the inner surface.

3. The method of claim 1 where the tracing of the refracted acoustic rays between each array element and the intended focal point through the barrier and the volume is accomplished by solving the Helmholtz wave propagation equation numerically in each medium with appropriate boundary conditions.

4. The method of claim 3 wherein the step of solving the Helmholtz wave propagation equation includes searching through boundary points and selecting boundary points at which all constraints are satisfied.

5. The method of claim 4 wherein the constraints include the Fermat's principle and the boundary conditions.

6. The method of claim 5 wherein an ultrasound transmitted from the array element to the desired focal point in the brain will pass through the boundary point at which all constraints are satisfied.

7. The method of claim 6 further including the step of interpolating more points between original grid points on highly undulating areas of the inner surface.

8. The method of claim 1 wherein said step b) is performed using the probe.

9. The method of claim 1 wherein said step h) is performed in both active mode and passive mode, wherein the active mode accurately focuses the acoustic field emitted by the probe and transmitted through the barrier at intended focal locations behind the barrier, and the passive mode accurately focuses the acoustic energy reflected from the media behind the barrier, passing through the barrier, and received by the probe and generates an image.

10. The method of claim 1 further including the step of optimizing the beamforming of said step h) for side lobe suppression.

11. The method of claim 1 wherein the local thickness of the barrier is estimated in said step b) based upon the sound speed of the barrier determined in said step b).