Using ultrasound anisotropic model for organ analysis

Abstract

A method is described for generating high-resolution images of transversely isotropic mechanical properties and attenuation properties of tissues associated with measured compressional waves. A pressure wave is transmitted to a region containing fibers that are oriented with transverse isotropy. Waves are received that propagate after the pressure wave has impinged upon the fibers. An initial velocity and attenuation model is created that takes into account the transverse isotropy of the fibers. Full Waveform Inversion (FWI) is used to iteratively minimize a difference between the received data and the initial velocity and attenuation model to create an image of the region that takes into account the transverse isotropy of the fibers. The image has a higher resolution than if created using the initial velocity and attenuation model.

Description

USING ULTRASOUND ANISOTROPIC MODEL FOR ORGAN ANALYSIS FIELD OF THE INVENTION

[0001] The present invention relates to using ultrasound for organ analysis. BACKGROUND OF THE INVENTION

[0002] The internal structures of muscles, ligaments, and tendons in the human body have a preferred orientation and arrangement based upon the organ’s specific functionality. Muscles are made of fibers called fascicles (50-500 μm in diameter) which are surrounded by a connective tissue layer called perimysium. The fascicle arrangement varies from one muscle type to another depending on the force it generates. Parallel muscles have the fascicles arranged in parallel to each other, circular muscles are arranged in a ring-like formation, and fascicles in a pennate muscle are arranged in an oblique manner to the central tendon. Similar to muscles, tendons, and ligaments, fibrillar structure is also governed by a preferred orientation. The orientation of the fibrillar structure depends on the stress to which the tendon or the ligament is subject. For example, in the patellar tendon, the alignment is predominantly along the long axis of the tendon.

[0003] The internal structure arrangement of muscles, tendons, and ligaments results in what is known as mechanical anisotropy. Mechanical anisotropy is the quality of exhibiting different stress-strain relationships when measuring the properties along different axes in different directions. This can be observed by analyzing 3D ultrasound data of propagating acoustic and elastic waves through the body.

[0004] A good example of many soft, fibrillar, tissues that are located in a very tight proximity that can benefit from the invention is the knee. The human knee is made of several ligaments and tendons connecting the femur, the patella, the fibula, and the tibia. The quadriceps tendon connects the quadriceps muscles to the patella, and below, an additional ligament connects the patella to the tibia. Lateral collateral ligament connects the femur to the fibula on one side and to the tibia on the other. The posterior cruciate ligament (PCL) and the anterior cruciate ligament (ACL) are in the middle of the knee connecting the femur to the tibia through the articular cartilage and the lateral meniscus. Ligament injuries are very common in knee joint trauma. The most common injury is a stretched or torn ACL during a sudden twist motion of the knee. Another common injuryis a PCL injury from a sudden and direct impact. As a result of such injuries, the structural alignment of the tendons and ligaments may change.

[0005] Effect on Imaging

[0006] Standard medical imaging modalities do not make use of the anisotropy of tissues. For example, medical ultrasound measures propagation speed in the direction of the scanning beam, not in the direction of fibers. MRI measures water and fat presence through relaxation times, but does not consider tissue anisotropic quantities.

[0007] In many medical cases, there is a need to determine whether a specific muscle, tendon, or ligament is damaged. In many cases, the tissues in question are located in very close proximity to each other making the distinction between them very difficult. The present invention measures anisotropy to solve this problem and differentiates between anisotropic and isotropic materials.

[0008] Ligaments are built from collagen fibers (10-50 μm in diameter) usually arranged in parallel bundles, and are very strong and flexible. Due to this arrangement, ligaments exhibit mechanical anisotropy: different stress-strain relationships along different axes in different directions. Stress applied on the ligament in a direction parallel to the fiber arrangement will result in observed strain magnitude and direction different from the observed strain response to stress applied in the perpendicular direction. Therefore, analysis of ultrasound waves propagating along different directions, given emitter and sensor locations on the body, enables the estimation of anisotropic properties. An ultrasound system with a wide range angle transducer setup can be used to emit and measure transmitted and refracted waves simultaneously in many directions, providing the base data needed to evaluate anisotropy.

[0009] Fibers arranged in a parallel manner with a given orientation will induce transverse isotropy (TI) in which the stress-strain relationship is rotationally invariant with respect to a single axis of symmetry. (It is noted that the invention is applicable for any general anisotropy.) A pressure wave propagating normal to fiber alignment will travel at a given velocity that is not identical to the velocity of a wave propagating in the direction parallel to the fibers. The result is a change in the propagation direction and the waveform. To estimate the anisotropic properties of muscles, ligaments, tendons, andother organs, the present invention uses travel-time tomography and a Full Waveform Inversion (FWI) algorithm.

[0010] Hard tissue such as bones may have cracks and fractures. The fractures can be small enough to be below ultrasound wavelength resolution. In such a case the cracks cannot be imaged by measuring the ultrasound waves propagating through the bone. Instead, directionally aligned fracture sets will induce additional planes of symmetry to the existing background isotropic or transversely isotropic medium. The invention uses a more general anisotropic model to accurately characterize the fractured hard tissue area and other bone pathologies (for example: contusions, malignancies etc.).

[0011] Using transverse isotropy in Full Waveform Inversion (FWI) could give an advantage for tumor diagnosis by enabling more accurate modeling of tissue stiffness and directional wave propagation. Malignant tumors often exhibit transversely isotropic mechanical behavior due to aligned fibrotic tissue or collagen structures in the tumor microenvironment. Modeling this transverse isotropy would allow FWI to account for direction-dependent variations in wave speed and attenuation, resulting in more precise reconstructions of tissue properties. This could improve accuracy of tissue characterization and improve the detection of tumor margins and potentially support the differentiation between malignant and benign lesions, which often differ in both their structural organization and degree of anisotropy.

[0012] The amplitude of a propagating acoustic / elastic wave is attenuated due to effects such as scattering, absorption, and the existence of abrupt interfaces between two mediums [2]. Elastic medium interfaces cause the wave’s amplitude to decrease as some of the energy is transmitted through while the rest is reflected. Absorption occurs when part of the propagation energy is lost due to heat. Both scattering and absorption cause energy at the higher frequencies (shorter wavelengths) within the wave’s bandwidth to decrease. The observed effective changes to the measured waves’ amplitude and frequency bandwidth can be used as input to the tomography [3] and the FWI [4] to solve for an attenuation model.

[0013] Imaging of shear anisotropy in muscles was investigated using shear wave elastography [5,6]. Elastography machines measure the anisotropic shear modulus of soft tissues by inducing a distortion and measuring the resulting shear wave speed. The pointof induced distortion is moved so the relevant organ is scanned to produce an image. The articles [5,6] showed that it is possible to generalize the mechanism to measure shear anisotropy. SUMMARY

[0014] The invention seeks to derive the compressional wave anisotropy using a novel variation of ultrasound tomography and full waveform inversion based on ultrasound waves measured by a multi-transceiver system. The new method solves for the compressional anisotropy components for the entire scanned body part.

[0015] Travel-time tomography is a process in which propagating waves’ travel-time from emitter to sensor is used to estimate the medium’s acoustic / elastic velocities. It has been widely applied in earth seismology to study the earth’s interior and in exploration seismology for imaging subsurface oil and gas-bearing rock formations [7]. The method is based on modeling the propagating wave travel-time from any given emitter to any given receiver. The travel-time is calculated using a high-frequency approximation of the elastodynamic wave equation, the ray equation and the eikonal equation. The modeled time of travel is compared to the measured time of propagation and a misfit is calculated. An optimization process is used to update the acoustic / elastic velocity model until the discrepancy between the modeled and observed is minimized. Due to its nature, travel- time tomography results are smooth with low spatial resolution. Travel-time tomography can also be used to estimate an anisotropic model using anisotropic ray equation and eikonal equation. Anisotropic travel-time tomography solves for the TI medium’s axis of symmetry, the medium’s velocity along the symmetry axis, and additional mechanical anisotropic properties [8].

[0016] Full waveform inversion (FWI) is another algorithm that has been applied widely in geophysics on 2D / 3D seismic data to image underground structures for oil and gas exploration, engineering, and environmental purposes [9]. The method is based on fully simulating acoustic / elastic waves propagating in acoustic / elastic media. The modeled waveforms are then compared to the measured waveforms and a misfit is calculated. To complete the process, optimization algorithms are used to update the acoustic / elastic model to minimize the discrepancy between measured and modeled waveforms. The final result is the model that best describes the observed data. FWI can reach higher spatialresolution than standard ultrasound images. Furthermore, the FWI can also be used with an anisotropic wave equation to solve for the independent elastic properties of anisotropic transverse isotropy (TI) medium and its symmetry axis

[0010] .

[0017] FWI stands for the method FWI and its modifications such as: adaptive waveform inversion (AWI), wavefield-reconstruction Inversion (WRI).

[0018] FWI is computationally expensive as it relies on simulating the full wave equation. The complexity and the number of model parameters that the FWI solves for depend on the type of medium in which the ultrasound waves propagate. It is much more challenging to get both a fast and accurate solution to a medium that is characterized by more parameters. An anisotropic elastic medium can be characterized by five independent elastic constants, isotropic elastic medium requires two elastic parameters and isotropic acoustic can be characterized with one parameter. To obtain a meaningful solution cost-effectively, it is best to approximate the medium and linearize the problem. This means using a cascade workflow starting from a cost-effective acoustic FWI and proceeding to a more computationally expensive elastic FWI. At the later stages of the workflow, the FWI can be carried out on a focused localized area based on the region of interest.

[0019] The inverted anisotropic model from FWI can be used for a medical interpretation of injured tendons, ligaments, or muscles. The orientation and the anisotropy magnitude of the model used to characterize these organs can differentiate between them and between healthy and unhealthy regions.

[0020] There is provided in accordance with a non-limiting embodiment of the invention a method to generate images including transmitting a pressure wave to a region in a living creature, said region containing fibers that are oriented with transverse isotropy, in which a stress or strain relationship of said fibers is assumed to be rotationally invariant with respect to a single axis of symmetry said fibers, receiving received waves that propagate after said pressure wave has impinged upon said fibers, inputting data from said received waves to a wave propagation simulator to create an initial velocity and attenuation model that takes into account the transverse isotropy of the fibers, and using Full Waveform Inversion (FWI) to iteratively minimize a difference between the received data and theinitial velocity and attenuation model to create an image of said region that takes into account the transverse isotropy of the fibers, said image having a higher resolution than if created using said initial velocity and attenuation model. The image of said region may include an image which displays principal axes orientation and a measurement of anisotropy is discernible from said image which is a ratio between stress or strain along an axis of symmetry and stress or strain along an axis which is orthogonal to said axis of symmetry. The initial velocity and attenuation model may be created by using travel-time tomography. The method may use the FWI to segment near isotropic inclusions in said region. Using the FWI may include first carrying out acoustic FWI and then carrying out isotropic or anisotropic FWI to improve resolution accuracy. Using the FWI may include first carrying out elastic FWI and then carrying out isotropic or anisotropic elastic FWI to improve resolution accuracy. The initial velocity and attenuation model may be derived using a deep neural network. A physical orientation of the patient may be used to induce strain in said region. The region may include soft tissues around limbs or joints, and limb muscles away from the joints may be flexed to increase muscle strain during measurement, or limb muscles away from joints may be extended to reduce muscle strain during measurements, or limb joints may be kept unbent during measurement to maintain specific strain and position of tissue while the limb is extended, or limb joints may be kept bent during measurement to allow different exposure of ligaments and tendons to transceiver locations and to apply additional background strain that will impact the intensity magnitude of the anisotropy and its orientation. BRIEF DESCRIPTION OF DRAWINGS

[0021] The present invention will be understood and appreciated more fully from the following detailed description, taken in conjunction with the drawings in which:

[0022] Figure 1 - An overview of tendon structure

[0012] .

[0023] Figure 2 - A schematic drawing of a transversely isotropic medium.

[0024] Figure 3 - A schematic drawing of a human body with two proposed schematic designs of a multi-transceiver ultrasound measurement system that encircles the body.

[0025] Figure 4 - A high-level diagram of the full process step from acquisition to end image.

[0026] Figure 5 - A high-level diagram of the processing pipeline.DETAILED DESCRIPTION

[0027] To better understand the invention, there are different stages, not all of which are essential to the invention:

[0028] Patient preparation stage: patient positioning before measurement system placement.

[0029] Acquisition stage: ultrasound waves are emitted and measured by a multi- transceiver system that encircles the patient’s region of interest.

[0030] Data analysis and model building: a high-resolution model is built from data. Each voxel (3d pixel) of the model holds the elastic coefficients. Part of the volume (“region of interest”) or all of it can be anisotropic and the rest is isotropic.

[0031] Post-Processing Stage: Derivation of zones with high and low anisotropy magnitude based on anisotropic model property values.

[0032] Without limitation, the invention provides a novel method for data analysis and for building the model. The acquisition may use an ultrasound sound array with wide angle range measurement capabilities. The acquisition works best for 360° angle measurements, but can be carried out with other ranges of angles less than 360°.

[0033] Patient Preparation

[0034] Contracting muscles and muscles under different levels of tension may exhibit different in-situ stress-dependent mechanical properties. While muscles contract, ligaments and tendons act as taut ropes to allow joint stability and to pull the bone into movement. Both ligaments and tendons may stretch a bit but have very little elasticity compared to muscles

[0014] . Placing a joint or a limb in a given position will impact the stress on the local tendons / ligaments and the results strain. Moreover, the general orientation of both may also change. Similar to muscles, there may be in-situ stress- dependent mechanical properties changes and this can also be observed through the analysis of ultrasound waves propagating through. Specific body stances are used for normal ultrasound imaging, and similarly, certain body stances are very effective for imaging using TI anisotropic ultrasound imaging.

[0035] The patient's body stance along with limb and joint positioning is set during the placement of the measurement system around the region of interest. The controlled positioning of the different body parts (limbs, joints, etc.) is predetermined based on the objectives of the analysis and the region of interest. Limbs will be positioned relative to the torso and joints will be kept straight or bent. Apart from the measurement system, additional support instruments may be used to maintain the patient’s posture throughout the data acquisition stage.

[0036] In a specific embodiment, the measuring system checks the stance during the acquisition stage (using fast low-resolution tomography) and reports to the operator whether it is adequate.

[0037] Multiple successive data acquisitions may take place, depending on the given region of interest. Before each measurement, the patient will assume a predetermined body positioning and will maintain the position until all data has been collected. In between each measurement, the system placement will be adjusted on the patient. This will ensure accurate analysis and successful fit-to-purpose measurements of ultrasound wave propagation.

[0038] Acquisition Stage

[0039] An ultrasound measurement system consisting of a large array of transceivers will be placed on or around a patient’s body, encircling the imaged -region of interest (head, torso, abdomen, pelvis or limb pending on the region of interest). The system’s encirclement of the -region of interest is used for a multi-directional acquisition of ultrasound waves.

[0040] During the data acquisition, a predefined emission schedule will be executed. It will include the order in which any given transceiver is activated as a source of the ultrasound energy pulse or as a recorder of the emitted signal. The emitted ultrasound pulse will be broadband (10 kHz – 3 MHz). All sensors simultaneously record the incoming acoustic / elastic waves propagating through the patient’s region of interest in a 3D multidirectional sense.

[0041] In a different embodiment, parts of the encircling array have different transducer densities.

[0042] Building the 3D model

[0043] Processing the data to build a model may be done in 4 steps. These steps can viewed as successive approximations, each improving on the former:

[0044] Initial Model Building: An isotropic travel-time tomography may be used to estimate a low-resolution velocity and attenuation model that will serve as the FWI initial model. The initial model may be built using either isotropic travel-time tomography or a deep neural network.

[0045] Using the information from the acoustic model we can build an initial model transverse isotropic model. For example, segmented volumes with speed of sound >2500m / s are assumed to be made of bone. We can determine the long axis of the organ and assume that the axis of symmetry is parallel to it for all points there. Similar considerations can be applied to some (though not all) of the muscles and ligaments (with or without neural networks). Using this improved initial model will enable faster convergence and more accurate final model.

[0046] Isotropic FWI: Solving for high-resolution acoustic model and then, optionally, solving for elastic isotropic model.

[0047] Initial Anisotropic Model Building: localized / full anisotropic travel-time tomography may be used to estimate a low-resolution velocity model that will serve as the anisotropic FWI initial model.

[0048] Anisotropic FWI: Solving for high resolution acoustic anisotropic model and then, optionally, for elastic anisotropic model.

[0049] In a preferred embodiment, the full pipeline is done only on the region of interest, while the rest of the volume performs only part of the pipeline: Parts of the volume only run the isotropic steps, or only the acoustic steps.

[0050] Initial isotropic model building

[0051] In one embodiment, the initial model building uses travel-time tomography to compute a low-resolution isotropic model. Travel-times of two wave types are analyzed, direct waves and scattered waves. Direct waves are waves that transmit across themedium directly from the emission point to the location of the sensor. Scattered waves are waves that are refracted for an interface between two organs with different acoustic / elastic impedance and change direction.

[0052] In another embodiment the initial isotropic model is obtained directly from received sensors’ data using deep-neural-network (DNN). The DNN is trained offline on a plurality of organs and therefore improves the model by adding prior knowledge about the structure of those organs.

[0053] Isotropic FWI

[0054] In this stage acoustic FWI is used to solve for a high-resolution acoustic isotropic model based on measured waveforms from all receivers and the initial model from the previous stage. The resulting acoustic isotropic model offers a good approximation for most soft tissues and a weak approximation for bones.

[0055] Bones, unlike soft tissues, are stiffer and more elastic. Elastic material responds to both bulk and shear strain and stress. It requires more parameters to accurately describe elastic material. In practice, elastic FWI improves the accuracy of both the hard and the soft tissue. Hence, further improving the model accuracy by running elastic isotropic FWI is an optional step, by using the acoustic FWI results as the initial model.

[0056] In a preferred embodiment, the elastic isotropic FWI will be carried out near bones and near the region of interest.

[0057] Initial anisotropic model building

[0058] This step augments the high-resolution isotropic model found in the former step with low-resolution anisotropic model building.

[0059] In one embodiment, the low-resolution anisotropic model building uses anisotropic travel-time tomography. The same observed travel-time data is used as input to solve for an acoustic anisotropic model. The only difference is the use of a forward model operator of anisotropic ray equations for modeling wave propagation travel times. The tomography results are 3D voxels of anisotropic model properties and orientation angles of the symmetry axes.

[0060] The results of the low-resolution anisotropic travel-time tomography are merged with the high-resolution elastic isotropic model to form an initial low-resolution elastic anisotropic model.

[0061] In another embodiment the initial anisotropic model is obtained directly from received sensors’ data using deep-neural-network (DNN). The DNN is trained offline on a plurality of organs and therefore improves the model by adding prior knowledge about the structure of those organs.

[0062] Anisotropic FWI

[0063] Elastic pressure wave equations in TI media can be approximated with expanded anisotropic acoustic wave equations. The acoustic approximation is only valid for propagating pressure waves and excludes shear waves. Forward modeling of ultrasound waves propagating through anisotropic acoustic media requires less parameters and reduces the computation cost and time.

[0064] Anisotropic acoustic FWI input includes sensed ultrasound waveforms and the initial anisotropic model from the former stage. The acoustic anisotropic model computed by FWI is a 3D estimated high-resolution volume of model properties at or near the region of interest itself and not for the whole volume area covered by the measurement system. Each point in the volume will include the estimated acoustic anisotropic properties and the orientation angles of the symmetry axes.

[0065] During this stage, the FWI will also update and improve the accuracy of the localized soft / hard tissue attenuation.

[0066] As an optional stage, elastic anisotropic FWI is carried out on a need basis, given the same region of interest described above, and by using the high-resolution acoustic anisotropic FWI model as an input. The purpose of this stage is to improve the estimation accuracy of the localized anisotropic properties.

[0067] Post Processing

[0068] The elastic stiffness of a TI medium can be characterized by the bulk density and 5 independent elastic coefficients as shown in Eq. 1

[0011] . In it, C33and C44are associated with the pressure wave and shear wave propagation velocity along the axis of symmetry, respectively. C11and C66are related to the velocity of the pressure wave and horizontally polarized shear wave propagating normal to the axis of symmetry. And C13is related to the off axes wave propagation velocities. Isotropic medium is a particular case of TI symmetry where C11= C33, C44= C66, and C13= C33– 2C44.

[0070] The elastic stiffness of an orthotropic medium with 3 planes of symmetry can be characterized by bulk density and 9 independent elastic coefficients as shown in Eq. 2

[0015] .

[0072] A more general anisotropic symmetry is triclinic symmetry which can be characterized by bulk density and 21 independent elastic coefficients as shown in Eq 3

[0016] . ^ ^^^ C^^ ^^^^14^15^16C^^^^^^^^^24^ ^ ^ ^25 26^^^^^^^^^^ ^ ^ ^

[0073] ^ =^43 35 36^ (3) ^ ^14^24^34^^^^45^46^ ^ ^15^^^^35^45^^^^56^ ^ ^^^^^^^36^46^56^^^^

[0074] From a medical practitioner point of view, having 6 different properties (or 6 images) used to define the transverse isotropy, is not useful for diagnosis, and there is a need to produce fewer images that combine or highlight the data in a useful manner.

[0075] In the post-processing stage, the anisotropic FWI results are analyzed. At each point in the high-resolution anisotropic 3D model, elastic properties of strain-stress relationship along given axis of symmetry are compared to properties along another symmetry axis and properties along the off-principal axes direction. The magnitude of anisotropy is then calculated from this comparison (e.g. anisotropy intensity is defined as the ratio between two values from different axes). The output is a 3D volume identical in size and resolution to the model solved by the anisotropic FWI of derived anisotropy magnitude.

[0076] The output volume can be used to segment different organs based on anisotropy contrast and magnitude.

[0077] A specific implementation of ‘localized’ elastic / anisotropic FWI of the knee (see Fig.6):

[0078] For clarifying the implementation of ‘localized’ elastic / anisotropic FWI, herein we use a case where the target organ is the knee’s ACL. The ACL is located between the femur and the tibia. Once a high-resolution acoustic model is obtained from the acoustic FWI, hard and soft tissue can be identified. With the prior knowledge of the bones’ location around the ACL elastic FWI is carried out to update the bones’ elastic properties within a distance of 3 cm to 4 cm above and below the target organ. Bones outside that zone will be assigned similar elastic properties as the localized solution. In a similar manner, the anisotropic FWI will target and update the mechanical model around the connecting area between the femur and the tibia. Outside that zone, both soft and hard will be assigned relevant anisotropic values based on the final model and organ classification.

[0079] References

[0080] [1] Backus, G. E., 1962, Long-wave elastic anisotropy produced by horizontal layering, Journal of Geophysical Research, 67, 11, 4427 - 4440.

[0081] [2] Viren, A. R., 1989, Ultrasonic attenuation estimation for tissue characterization, Retrospective Theses and Dissertations. 17318. https: / / lib.dr.iastate.edu / rtd / 17318

[0082] [3] Romero, A. E. J., McEvilly, T. V., and Majer, E. L., 1997, 3-D microearthquakes attenuation tomography at the Northwest Geysers geothermal region, California, GEOPHYSICS, 62, 1, 149 - 167.

[0083] [4] Kamei, R. and Pratt, R.G., 2013. Inversion strategies for visco-acoustic waveform inversion, Geophys. Journal International, 194, 859–884.

[0084] [5] Ngo, H. H. P., Poulard, T., Brum, J., & Gennisson, J. L. 2022. Anisotropy in ultrasound shear wave elastography: An add-on to muscles characterization. Frontiers in Physiology, 2020.

[0085] [6] Knight, A. E., Trutna, C. A., Rouze, N. C., Hobson-Webb, L. D., Caenen, A., Jin, F. Q., ... & Nightingale, K. R., 2021, Full characterization of in vivo muscle as anelastic, incompressible, transversely isotropic material using ultrasonic rotational 3D shear wave elasticity imaging. IEEE Transactions on Medical Imaging, 41(1), 133-144.

[0086] [7] Schuster. G. T., 2017, Seismic Inversion Chapter 5: Raypath Traveltime Tomography, 70, Society of Exploration Geophysicists.

[0087] [8] Chapman, C. H. and Pratt, R. G., 1992, Travel time tomography in anisotropic media - I Theory, Geophysical Journal International, 109, 1 - 19.

[0088] [9] Virieux, J. and Operto S., 2009, An overview of full-waveform inversion in exploration geophysics, GEOPHYSICS, 74, WCC1-WCC26.

[0089]

[0010] Warner, M., Ratcliffe, A., Nangoo, T., Morgan, J., Umpleby, A., Shah, N., VInje, V., Stekl, I., Guasch, L., Win, C., Conroy, G., and Bertrand, A.,, 2013. Anisotropic 3D full-waveform inversion, Geophysics, 78, R59–R80

[0090]

[0011] Postmat G. W., 1955, Wave propagation in stratified medium, GEOPHYSICS, 4, 780 – 806.

[0091]

[0012] Jull, G., Moore A., Falla D., Lewis J., McCarthy C., and Sterling M., 2005, Grieve’s modern musculoskeletal physiotherapy, Elsevier.

[0092]

[0013] A wearable Garment adapted for ultrasound sensing and methods thereof for full wave inversion with imprecise sensor positions, WO2022123464A1

[0093]

[0014] Hazari A., Maiya A., and Nagda T., 2021, Conceptual biometric and kinesiology: Applied biomechanics on joint, muscle, tendon, and ligament. Singapore: Springer.

[0094]

[0015] Tsvankin, I., 1997, Anisotropic parameters and P-wave velocity for orthorhombic media: GEOPHYSICS, 4, 62, 1292 – 1309.

[0095]

[0016] Bond, W., 1943, The mathematics of the physical properties of crystals: The Bell System Technical Journal, 22, 1 – 72.

Claims

CLAIMS 1. A method to generate images comprising: transmitting a pressure wave to a region in a living creature, said region containing fibers that are oriented with transverse isotropy, in which a stress or strain relationship of said fibers is assumed to be rotationally invariant with respect to a single axis of symmetry said fibers, receiving received waves that propagate after said pressure wave has impinged upon said fibers; inputting data from said received waves to a wave propagation simulator to create an initial velocity and attenuation model that takes into account the transverse isotropy of the fibers; and using Full Waveform Inversion (FWI) to iteratively minimize a difference between the received data and the initial velocity and attenuation model to create an image of said region that takes into account the transverse isotropy of the fibers, said image having a higher resolution than if created using said initial velocity and attenuation model.

2. The method according to claim 1, wherein said image of said region comprises an image which displays principal axes orientation and a measurement of anisotropy is discernible from said image which is a ratio between stress or strain along an axis of symmetry and stress or strain along an axis which is orthogonal to said axis of symmetry.

3. The method according to claim 1, wherein said initial velocity and attenuation model is created by using travel-time tomography.

4. The method according to claim 1, further comprising using the FWI to segment near isotropic inclusions in said region.

5. The method according to claim 1, wherein using the FWI comprises first carrying out acoustic FWI and then carrying out isotropic or anisotropic FWI to improve resolution accuracy.

6. The method according to claim 1, wherein using the FWI comprises first carrying out elastic FWI and then carrying out isotropic or anisotropic elastic FWI to improve resolution accuracy.

7. The method according to claim 1, wherein said initial velocity and attenuation model is derived using a deep neural network.

8. The method according to claim 1, wherein a physical orientation of the patient is used to induce strain in said region.

9. The method according to claim 8, wherein said region comprises soft tissues around limbs or joints, and wherein limb muscles away from the joints are flexed to increase muscle strain during measurement.

10. The method according to claim 8, wherein said region comprises soft tissues around limbs or joints, and wherein limb muscles away from joints extended to reduce muscle strain during measurements.

11. The method according to claim 8, wherein said region comprises soft tissues around limbs or joints, and wherein limb joints are kept unbent during measurement to maintain specific strain and position of tissue while the limb is extended.

12. The method according to claim 8, wherein said region comprises soft tissues around limbs or joints, and wherein limb joints are kept bent during measurement to allow different exposure of ligaments and tendons to transceiver locations and to apply additional background strain that will impact the intensity magnitude of the anisotropy and its orientation.

Images