Ring ultrasound-mammogram device

Abstract

We provide a standing device for breast ultrasound based on a large half-ring array to significantly improve the resolution and contrast of images. It is capable of elevational (outward from the torso) motorized scanning for 3D acquisition. It will be similar in a sense to a mammogram but instead of a flat plate the transducer array is an arc to avoid painful compression. This would provide a comfortable posture and generate ultrasound images in a manner similar to the gold standard mammogram but with enhanced resolution due to the shape. Additionally, the system is capable of imaging modes commonly found on clinical ultrasound machine such as Doppler, contrast or elastography.

Description

[0001] Ring Ultrasound-mammogram device by

[0002] Katherine W . Ferrara

[0003] Eunyeong Park Josquin L . Foiret

[0004] FIELD OF THE INVENTION

[0005] This invention relates to ultrasound imaging for medical applications .

[0006] BACKGROUND

[0007] Ultrasound imaging is a useful medical imaging modality that doesn ' t suf fer from the disadvantages of medical imaging approaches that rely on ioni zing radiation . However, some common applications of ultrasound imaging, such as breast imaging, suf fer from various disadvantages , such as painful compression, slow results , water coupling and / or prone position of the patient . Accordingly, it would be an advance in the art to provide improved ultrasound imaging suitable for applications such as breast imaging .

[0008] SUMMARY

[0009] We have developed a standing device for breast ultrasound based on a large hal f-ring array to signi ficantly improve the resolution and contrast of images . It is capable of elevational ( outward from the torso ) motori zed scanning for 3D acquisition . It will be similar in a sense to a mammogram but instead of a flat plate , the transducer array is an arc to avoid painful compression . This would provide a comfortable posture and generate ultrasound images in a manner similar to the gold standard mammogram but with enhanced resolution due to the shape . Additionally, the system is capable of imaging modes commonly found on clinical ultrasound machine such as Doppler, contrast or elastography . The system can be designed to image the entire breast within 3 seconds .

[0010] Signi ficant advantages are provided . The system provides a desirable combination of fast results , no ioni zing radiation, improved resolution, and / or imaging of both anatomy and functional changes .

[0011] In particular this work uses mechanical scanning of a large ID array of sensors to provide very good breast imaging results . Key features of some embodiments are element spacing of hal f-wavelength ( or less ) and scanning an active window through the ID array where beam steering and / or beam shaping is done by the elements in the active window . Another important point is the use of a speed-of- sound matching material between the sensor array and the tissue being imaged .

[0012] In a preferred embodiment we have a ring of ~ 180 degrees ( so that the breast can slip down into it when standing or sitting) and the 1 / 2 lambda transducer element spacing allows us to steer the beam and interrogate the entire breast cross section without moving it . In contrast , most conventional designs use a full 360 degree transducer, which disadvantageously requires the patient to be prone and use water coupling .

[0013] Further alternatives and variations include the following : 1 ) a transducer that can reconstruct the entire breast cross section in receive mode scanning with X / 2 element spacing without mechanical repositioning .

[0014] 2 ) The form and transducer provide for upright scanning of an entire breast .

[0015] 3 ) The form is created with a set of discrete si zes and the appropriate si ze selected for each patient to ensure that the entire breast is imaged with the arc array of transducers .

[0016] 4 ) An apparatus where the entire upright breast volume is scanned for B mode imaging in less than 10 seconds .

[0017] 5 ) An apparatus that produces video rate cross sectional images interleaved with functional three dimensional imaging of the vasculature and tissue mechanical properties .

[0018] 6 ) An apparatus where an entire 3D scan of echo data is recorded in real time and used to sequentially reconstruct the speed of sound within the breast cross section using a 2 speed-of-sound algorithm for the form and the tissue and where the tissue speed of sound is varied during reconstruction to match each component of fat , glandular tissue , cyst tissue , and tumor .

[0019] 7 ) An apparatus where small lesions / di f f erential tissue components are identi fied through the reconstructions of the changes in the received image as a function of the reconstructed speed of sound using image scaling and subtraction of the varied speed of sound images in the recorded scans of echo data .

[0020] 8 ) We can also perform elastography compounding with the apparatus . The idea is to estimate the elastic modulus of the tissue by imaging the propagation of shear waves after an acoustic radiation force push and to combine the estimations from multiple views (similar to the Doppler images of this work) . Doing (single view) elastography by imaging shear wave propagation is known but compounding elastography from multiple views is novel.

[0021] BRIEF DESCRIPTION OF THE DRAWINGS

[0022] FIGs. 1A-B show an exemplary embodiment of the invention .

[0023] FIGs. 1C-D show another exemplary embodiment of the invention .

[0024] FIGs. 2A-B are images showing the effect of 1- wavelength pitch vs. 1 / 2-wavelength pitch.

[0025] FIGs. 3 is a legend for the results of FIGs. 4A-F and 5A-F.

[0026] FIGs. 4A-F show point spread function (PSF) analysis results for +15 degrees of beam steering.

[0027] FIGs. 5A-F show point spread function (PSF) analysis results for +45 degrees of beam steering.

[0028] FIGs. 6A-D show human calf imaging results.

[0029] FIGs. 7A-B show the effect of speed of sound matching.

[0030] FIGs. 8A-D show an example of speed of sound imaging.

[0031] FIGs. 9A-I show a first example of Doppler flow imaging .

[0032] FIGs. 10A-I show a second example of Doppler flow imaging .

[0033] FIG. 11 shows an example of beam forming using a subarray of an arc transducer array.

[0034] FIG. 12 schematically shows elastography imaging. FIGs. 13A-B schematically show combining elastography imaging from several different beam directions.

[0035] DETAILED DESCRIPTION

[0036] FIGs. 1A-B show an exemplary embodiment of the invention. This exemplary apparatus includes an ultrasound scanner including a 1-D array 106 of sensor elements (e.g., 106a, 106b, ...) . The 1-D array of sensor elements is disposed in an arc having a span between 145 degrees and 315 degrees about an axis of the arc. In the coordinates shown on FIGs. 1A-B, the axis of the arc is the y axis. The 1-D array of sensor elements is configured to capture a complete in-plane 2D cross section image of a target tissue 102 without mechanical motion. Here the in-plane 2D cross section image is in a plane perpendicular to the axis of the arc (i.e., a plane perpendicular to the y axis and on or parallel to the x-z plane) . Let X be an operating acoustic wavelength of the ultrasound scanner in the target tissue. The element spacing of the sensor elements is X / 2 or less.

[0037] The 1-D array of sensor elements is mechanically scannable along the axis of the arc (y axis) , as shown in the side view of FIG. IB, such that 3D imaging is provided by a combination of the in-plane 2D cross section images from two or more positions of the 1-D array of sensor elements. Here the side view of FIG. IB is along the dashed line of FIG. 1A.

[0038] The apparatus also includes a form 104 disposed to fill space between the ultrasound scanner and the target tissue, where the form is configured to have a speed of sound match to the target tissue. The main purposes of form 104 are maintaining a good contact through the scan for different si zes of target tissue , achieving high image quality (matching speed-of-sound, pushing near- field) , and hygiene . As a further point , form 104 can be disposed on a housing of the ultrasound scanner so that the patient is not contacted by any moving parts .

[0039] Optionally, elements of the of the 1-D array of sensor elements are connected 1 : 1 to per-element readout circuits ( e . g . , 108a, 108b, . . . ) in readout electronics 108 , so that transmission and reception on any subaperture or any combination of subapertures at any time is facilitated . Since array 106 is a 1-D array, providing per-element readout electronics is much easier than it is for large 2-D arrays of acoustic sensor elements .

[0040] Readout electronics 108 can include an ultrasound imaging processor configured to electrically scan an active window through the 1-D array of sensor elements ( e . g . , as shown on FIG . 12 ) . Beam steering and / or beam shaping can be provided by sensor elements within the active window . The ultrasound imaging processor can be configured to reconstruct the complete in-plane 2D cross section image in receive mode scanning of the target tissue without mechanical translation .

[0041] The apparatus can be configured for whole-breast imaging via upright scanning . Preferably a whole-breast scan for B-mode imaging can be performed in 10 seconds or less .

[0042] The apparatus can further include a functional imager 120 of the target tissue 102 . In such cases , it is preferred that the apparatus is configured to perform interleaving of ultrasound imaging and functional imaging at a rate of 30 Hz or more . Speed of sound image reconstruction can be used to differentiate tissue types within the target tissue, such as: fat tissue, glandular tissue, cyst tissue, and tumor tissue. The speed of sound image reconstruction can be performed by determining changes in a received image as a function of speed of sound using image scaling and subtraction of varied speed of sound images in scans of echo data .

[0043] FIGs. 1C-D show some further details of an exemplary embodiment for breast imaging. Here 104 is the speed of sound matching form as described above. For breast imaging, it is expected that this form would come in several different sizes to accommodate patient variation, and that switching between forms of different sizes would be a simple operation for a user of the system to perform before a scan. 110 is a pillow (e.g., a conformal acoustic absorbing material) intended to reduce reverberation from the top of the breast. 112 is a height adjustment control, 114 is a linear actuator for the scanning (shown schematically) , and 116 is a counter-balance.

[0044] We have developed a standing device for breast ultrasound based on a large half-ring array (106 on FIGs. 1A-B) to significantly improve the resolution and contrast of images. It is capable of elevational (outward from the torso, z axis on FIG. IB) motorized scanning for 3D acquisition. It will be similar in a sense to a mammogram but instead of a flat plate has an arc without painful compression. This would provide a comfortable posture and generate ultrasound images in a manner similar to the gold standard mammogram but with enhanced resolution due to the shape. Additionally, the system is capable of imaging modes commonly found on clinical ultrasound machine such as Doppler, contrast or elastography . The system is designed to image the entire breast rapidly, preferably within 3 seconds .

[0045] Key features of the sensor array are element spacing of half-wavelength (or less) and electrically scanning an active window through the ID array where beam steering and / or shaping is done by the elements in the active window. We do not know of another array or approach like this for breast scanning. All of the full circles we see use lambda spacing or a small array and mechanically moving this array. The ideal image metrics would be achieved in the semi- tomographic configuration where the arrays describe a halfring and the resolution is equal or smaller than a half wavelength of the transmit signal. The half-wavelength pitch enables a large degree of beam steering without generating grating lobes, unlike what a one-wavelength pitch would generate, achieving uniform resolution not only at the physical center of the array but all over the region of interest. It can be operated and process up to 1024 independent channels in real-time to maximize acquisition speed and minimize motion artifact. Real-time imaging has been realized by distributed GPU computing. It is able to create an image of large fields of view, currently about 200 mm in diameter. FIGs. 2A-B show a comparison of one- wavelength pitch vs. 1 / 2 wavelength pitch. In these images, it is apparent the 1 / 2 wavelength pitch has fewer imaging artifacts between the points of the image.

[0046] The pattern of insonation has been optimized to yield a minimized point spread function throughout the field of view using either plane or diverging waves. Two examples of the improved PSF across the field of view are shown on FIGs. 4A-F and FIGs. 5A-F. FIG. 3 is a legend for these results, where the thick black lines show the array elements, and the dashed rectangle encloses the image points quantified in the tables of FIGs. 4A-F and FIGs. 5A-F. These results use an assembly of 8 separate apertures to realize a large half-ring array; however a full ring is also feasible .

[0047] FIGs. 4A-F show the results for 37 plane waves, ±15° of beam steering. The middle case (FIG. 4B) shows that the 0.5-lambda pitch combined with 1 aperture transmission / 3 reception and transmission summed across arrays yields a summed PSF (quantifying the PSF over the whole field-of- view) of 5175 microns, where a single aperture (FIG. 4A) is 92% worse and 1 lambda spacing is 26% worse (FIG. 4G) . Here FIGs. 4D-F show corresponding off-energy results, showing generally lower off-energy for 0.5-lambda pitch combined with 1 aperture transmission / 3 reception (FIG. 4E) than for a single aperture (FIG. 4D) or 1 lambda spacing (FIG. 4F) .

[0048] We can do even more by steering the beam further.

[0049] FIGs. 5A-F show the results for 37 plane waves, ±45° of beam steering. The middle case (FIG. 5B) shows that the 0.5- lambda pitch combined with 1 aperture transmission / 3 reception and transmission summed across arrays yields a summed PSF (quantifying the PSF over the whole field- of- view) of 3347 microns, where a single aperture (FIG. 5A) is 187% worse and 1 lambda spacing is 75% worse (FIG. 5C) .

[0050] Here FIGs. 5D-F show corresponding off-energy results, showing generally lower off-energy for 0.5-lambda pitch combined with 1 aperture transmission / 3 reception (FIG. 5E) than for a single aperture (FIG. 5D) or 1 lambda spacing (FIG. 5F) . Here, the sum across locations is further improved and the difference between geometries is even greater .

[0051] Examples of the improvement in field of view on human imaging (calf images) are shown on FIGs. 6A-D. It's clear that 8 arrays (FIG. 6B) provides a better contrast and more information on morphologies and textures of tissues than a single array (FIG. 6A) , and that single-array transmission (FIG. 6C) and three-array transmission (FIG. 6D) are both viable options.

[0052] Another important point is the use of a speed-of-sound matching material (104 above) between the sensor array and the tissue being imaged. Its size and shape is preferably optimized to any particular breast and it is preferably disposable to maintain hygiene. The use of this stand-off form improves the image quality by pushing the near-field out of the ROI (region of interest) and matching speed of sound to tissue which is not achievable with water coupling. FIGs. 7A-B show a comparison of images with (FIG. 7A) and without (FIG. 7B) use of the speed-of-sound matching material. Here it is clear the image of FIG. 7A is better.

[0053] With the acquired data, the image can be repeatedly reconstructed with a dual SOS to eliminate any effect of the standoff form and a varied speed of sound (SOS) within the breast to facilitate the user-based or automated recognition of lesions with a different speed of sound. The images are then subtracted from one another to reveal sub-regions with a different SOS that could indicate disease. A set of images that demonstrates this process is shown on FIGs. 8A-D. Here FIG. 8A is a reference image with a single speed of sound (SOS) . FIGs. 8B-D are dual-SOS images computed as described below. Here the highlighted region is different on FIG. 8D than it is on FIGs. 8B-C, which could be significant.

[0054] Exemplary dual-SOS imaging process:

[0055] 1) Beamforming in a coarse grid (a few wavelengths)

[0056] 2) Set ROI (region of interest) based on the known dimension of the spacer / interf ace form

[0057] 3) Set cO as the known speed of sound in the form 4) Repeatedly perform the beamforming process by adjusting delay with a varied speed of sound cl:

[0058] (dO + dl) / cO dO / cO + dl / cl where dO and dl are the pathlengths where sound wave travels at speed cO and cl, respectively .

[0059] 5) Subtract the image amplitudes from one another to identify regions with a different range of SOS.

[0060] The conformal nature of the system allows imaging breast tissue with minimal compression and is thus adapted for blood flow quantification. The wide aperture design facilitates angle-independent doppler flow mapping of vessels with enhanced sensitivity, imaging fluid velocities as low as 0.5 mm / s . This is shown in the images of FIGs. 9A-I (5 cm / s, 300 Hz) and FIGs. 10A-I (0.05 cm / s, 100 Hz) . Here FIG. 9A shows the flow imaging results using all 8 subarrays of the arc transducer array, and FIGs. 9B-I show flow imaging results using single subarrays (as shown by the white lines in the images) . Similarly, FIG. 10A shows the flow imaging results using all 8 subarrays of the arc transducer array, and FIGs. 10B-I show flow imaging results using single subarrays (as shown by the white lines in the images) .

[0061] An exemplary process to reconstruct flow from all directions using the 1 / 2 ring is the following:

[0062] 1) Reconstruct IQ (In-phase and Quadrature) data ensemble

[0063] 2) Perform SVD (singular value decomposition) processing on each subaperture

[0064] 3) Filter the clutter by choosing appropriate limits based on the correlation of the singular vectors

[0065] 4) Repeat steps l)-3) for each window

[0066] 5) Incoherently combine the Doppler images from each angle. This process is unique. We demonstrate on FIGs. 9A-I and 10A-I that low velocity flow can be reconstructed regardless of direction where the sum is shown on FIGs. 9A and 10A.

[0067] FIG. 11 shows a simulation example of beamforming using a sub array of an arc array.

[0068] In some embodiments of the invention, shear wave elastography is employed. FIG. 12 schematically shows this, where 1202 is the active sub-aperture of arc array 106, 1204 is the acoustic beam (e.g., as shown on FIG. 11) , 1206 is a breast being imaged, 1208 is a suspicious lesion, and 1210 schematically shows acoustic radiation force.

[0069] When a suspicious lesion 1208 is detected on the image data, shear wave elastography can be employed to assess the elastic property of the lesion. An acoustic radiation force 1210 is generated from a subaperture 1202 of the array. The propagation of the generated shear wave is then imaged at high framerates (>500 Hz) to provide an elasticity map. The process can be repeated for several subapertures to cover the entire half ring array and to generate multiple elasticity maps. Finally, a compounded elasticity map combining all the different views can provide a high quality estimation of the elastic modulus of the lesion.

[0070] FIGs. 13A-B schematically show this compounding of elastography images. FIG. 13A is an enlarged view of single-view elastography, where 1302 is the region of interest. FIG. 13B shows several such images 1304 from different subapertures being combined to form the compounded elasticity image 1306 of lesion 1208 with region of interest 1302 ’ .

Claims

CLAIMS1 . Apparatus for ultrasonic tissue imaging, the apparatus comprising : an ultrasound scanner including a 1-D array of sensor elements , wherein the 1-D array of sensor elements is disposed in an arc having a span between 145 degrees and 315 degrees about an axis of the arc ; wherein the 1-D array of sensor elements is configured to capture a complete in-plane 2D cross section image of a target tissue without mechanical motion, wherein the inplane 2D cross section image is in a plane perpendicular to the axis of the arc ; wherein the 1-D array of sensor elements is mechanically scannable along the axis of the arc such that 3D imaging is provided by a combination of the in-plane 2D cross section images from two or more positions of the 1-D array of sensor elements ; and a form disposed to fill space between the ultrasound scanner and the target tissue , wherein the form is configured to have a speed of sound match to the target tissue ; wherein X is an operating acoustic wavelength of the ultrasound scanner in the target tissue , and wherein an element spacing of the sensor elements is X / 2 or less .2 . The apparatus of claim 1 , further comprising : an ultrasound imaging processor configured to electrically scan an active window through the 1-D array of sensor elements , wherein beam steering and / or beam shaping is provided by sensor elements within the active window .

3. The apparatus of claim 2, wherein the ultrasound imaging processor is configured to reconstruct the complete in-plane 2D cross section image in receive mode scanning of the target tissue without mechanical translation.

4. The apparatus of claim 1, wherein the apparatus is configured for whole-breast imaging via upright scanning.

5. The apparatus of claim 4, wherein a whole-breast scan for B-mode imaging is performed in 10 seconds or less.

6. The apparatus of claim 1, further comprising a functional imager of the target tissue, and wherein the apparatus is configured to perform interleaving of ultrasound imaging and functional imaging at a rate of 30 Hz or more.

7. The apparatus of claim 1, wherein speed of sound image reconstruction is used to differentiate tissue types within the target tissue, wherein the tissue types are selected from the group consisting of: fat tissue, glandular tissue, cyst tissue, and tumor tissue.

8. The apparatus of claim 7, wherein the speed of sound image reconstruction is performed by determining changes in a received image as a function of speed of sound using image scaling and subtraction of varied speed of sound images in scans of echo data.

9. The apparatus of claim 1, wherein elements of the 1-D array of sensor elements are connected 1:1 to per-element readout circuits, whereby transmission and reception on any subaperture or any combination of subapertures at any time is facilitated.

Images