CN113440166A

CN113440166A - Liver disease activity estimation using ultrasound medical imaging

Info

Publication number: CN113440166A
Application number: CN202110270447.0A
Authority: CN
Inventors: Y·拉拜德
Original assignee: Siemens Medical Solutions USA Inc
Current assignee: Siemens Medical Solutions USA Inc
Priority date: 2020-03-12
Filing date: 2021-03-12
Publication date: 2021-09-28
Also published as: KR20210116268A; DE102021105965A1; KR102617857B1

Abstract

Liver disease activity estimation using ultrasound medical imaging. Ultrasound-based estimates (40) of disease activity are provided, such as an estimate of NAS or other activity index of NAFLD for liver disease. Ultrasound measures (30, 32) acoustic scattering and shear wave propagation parameters, such as measuring acoustic backscatter coefficients, shear wave velocity and shear wave damping ratio. From these scatter and shear wave propagation parameters, a score of disease activity is determined (40). In scoring activity of diseases such as NAFLD, physicians may be assisted by relatively inexpensive and rapid ultrasound. Ultrasound imaging (42) is more readily available and less expensive and MRI. Ultrasound is non-invasive.

Description

Liver disease activity estimation using ultrasound medical imaging

RELATED APPLICATIONS

This patent document is a continuation-in-part application of U.S. patent application serial No. 15/716,444 filed on 26.9.2017, which claims benefit of filing date under 35 u.s.c. § 119(e) of provisional U.S. patent application serial No. 62/482,606 filed on 6.4.2017. Both of these applications are incorporated herein by reference.

Background

The present embodiments relate to ultrasound imaging. Ultrasound is used to measure disease-related activity (activity) in tissue such as the liver.

Nonalcoholic fatty liver disease (NAFLD) is the most common liver disease in adults and children in the united states. NAFLD is characterized by excessive liver fat accumulation and liver fibrosis. Fat fraction (fat fraction) can be measured as an indicator of NAFLD. The fat composition and/or other tissue properties (e.g., the degree of fibrosis) in the liver or other tissues such as breast tissue provide diagnostically useful information.

More than 25% of patients with NAFLD develop nonalcoholic steatohepatitis (NASH). NASH can progress to cirrhosis and hepatocellular carcinoma. NAFLD Activity Score (NAS) is used to diagnose and monitor changes or levels of NASH. NAS was provided from liver biopsy histological evaluation and calculated as an unweighted sum of observed steatosis, lobular inflammation and bloating scores.

Magnetic Resonance Imaging (MRI) can measure proton density fat component (PDFF) as a biomarker of liver fat content. MRI can be used to further estimate NAS. However, MRI is not widely available and is expensive.

Disclosure of Invention

By way of introduction, the preferred embodiments described below include methods, instructions and systems for ultrasound-based estimation of disease activity, such as NAS or other activity index (index) for NAFLD. Ultrasound measures acoustic scattering and shear wave propagation parameters, such as measuring acoustic backscatter coefficients, shear wave velocity, and shear wave damping ratio (damping ratio). From these scatter and shear wave propagation parameters, a score of disease activity is determined. In scoring activity of diseases such as NAFLD, physicians may be assisted by relatively inexpensive and rapid ultrasound compared to biopsy or MRI based scoring. Ultrasound is non-invasive and is more readily available and less expensive than MRI.

In a first aspect, a method for non-alcoholic liver disease activity estimation with an ultrasound scanner is provided. A scan of a patient from an ultrasound scanner generates a first measurement of scatter in tissue. The first measure of scattering is the backscatter coefficient. Scanning of the patient from the ultrasound scanner generates second and third measurements of shear wave propagation in the tissue. The second measurement is the shear wave velocity and the third measurement is the shear wave damping ratio. A first value of an ultrasound-derived liver disease activity index is estimated from the backscatter coefficient, the shear wave velocity, and the shear wave damping ratio. Outputting an ultrasound image comprising an indication of the estimated ultrasound-derived first value of the liver disease activity index.

In a second aspect, a system for estimation of disease activity is provided. A beamformer (beamformer) is configured to transmit and receive sequences of pulses within a patient using a transducer. The sequence of pulses is used for the scattering parameter and for the first and second shear wave parameters. The image processor is configured to generate a score for an index of disease activity from a combination of the scattering parameter, the first shear wave parameter, and the second shear wave parameter. The display is configured to display a score of an index of disease activity.

In a third aspect, a method of liver disease activity estimation with an ultrasound system is provided. The ultrasound system determines a plurality of scattering parameters of liver tissue of the patient. The ultrasound system determines a plurality of shear wave parameters of liver tissue of a patient. Fat composition is estimated from at least one of the scattering parameters. Estimating a level of liver disease activity from at least one of the fat component and the shear wave parameter. Indicating the level of liver disease.

The present invention is defined by the following claims, and nothing in this section should be taken as a limitation on those claims. Additional aspects and advantages of the invention are discussed below in conjunction with the preferred embodiments and may be claimed later, either individually or in combination.

Drawings

The components and figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. Moreover, in the figures, like referenced numerals designate corresponding parts throughout the different views.

FIG. 1 is a flow diagram of one embodiment of a method for estimating tissue properties using ultrasound;

FIG. 2 is a block diagram of one embodiment of a system for estimating tissue properties using ultrasound;

FIG. 3 is a flow diagram of one embodiment of a method for estimating disease activity, such as NAS, using ultrasound; and

figure 4 is a plot showing the accuracy in predicting NAS from ultrasound measurements compared to histological NAS.

Detailed Description

Estimating disease activity aids in diagnosing, screening, monitoring, and/or predicting health conditions. For example, NAS or other liver disease activity is estimated. Using ultrasound to estimate the score of the index allows for a fast, inexpensive and non-invasive estimation of disease activity. The ultrasound derived NAFLD activity score was estimated, avoiding biopsy or MRI.

Disease activity may be estimated from measurements used in estimating tissue properties, such as liver fat component and/or estimation of liver fat component. Measurement and estimation of fat composition or other tissue properties is discussed below with reference to fig. 1. The use of these measurements and/or tissue properties in estimating disease activity is then discussed with reference to fig. 3.

With respect to tissue property estimation, Quantitative Ultrasound (QUS) is used to screen, diagnose, monitor, and/or predict health conditions. The complexity of human tissue can be measured using multiple QUS parameters in order to accurately characterize the tissue. For example, liver fat composition is estimated using a multi-parameter approach that combines quantitative parameters extracted from received signals of different wave phenomena, such as scattering and attenuation of longitudinal waves, propagation and attenuation of shear waves, and/or propagation and attenuation of on-axis waves (on-axis waves) from acoustic radiation force pulses (ARFI).

In one embodiment, tissue properties (e.g., hepatic fat composition) are estimated by transmitting and receiving a sequence of pulses to estimate scattering parameters, and by transmitting and receiving a sequence of pulses to obtain shear wave parameters. The estimation may also include transmitting and receiving a sequence of pulses to estimate the parameters from axial displacement caused by acoustic radiation force pulses (ARFI). QUS parameters are estimated and combined to estimate tissue properties. Other information, such as non-ultrasound data (e.g., blood biomarkers) may be included in the estimation of the tissue property.

Fig. 1 illustrates a method for tissue property estimation with an ultrasound scanner or system. Tissue responses to different types of waves or wave phenomena are measured. A combination of measurements of these different responses is used to estimate tissue properties.

The method is implemented by the system of fig. 2 or a different system. Medical diagnostic ultrasound scanners perform measurements by acoustically generating waves and measuring responses. An image processor of a scanner, computer, server, or other device makes an estimate based on the measurements. A display device, network or memory is used to output the estimated tissue property.

Additional, different, or fewer acts may be provided. For example, actions 33 and/or 38 are not provided. As another example, acts 36 and 37 are alternative or may be used together, such as averaging the results from both. In another example, an action is provided for configuring an ultrasound scanner and/or scanning.

The actions are performed in the order described or illustrated (e.g., top to bottom or numerically), but may be performed in other orders. For example, acts 30, 32, and 33 may be performed simultaneously, such as using the same transmit and receive pulses, or in any order.

In act 30, the ultrasound scanner generates a measurement of scatter in tissue from a scan of the patient. The measurement of scattering measures the tissue response to longitudinal waves transmitted from the ultrasound scanner. The scattering or echo (echo) of longitudinal waves impinging on the tissue is measured.

Any measurement of scattering may be used. Example scattering parameters include speed of sound, sound dispersion (dispersion), angular scattering coefficient (e.g., backscatter coefficient), frequency dependent attenuation coefficient, attenuation coefficient slope, spectral slope of normalized log spectrum (log-spectrum), spectral intercept of normalized log spectrum (intercept), spectral mid-band of normalized log spectrum, effective scatterer diameter, acoustic concentration, scatterer number density (number diversity), mean scatterer spacing, non-linearity parameters (B/a), and/or ratio of coherent to non-coherent scattering.

More than one measurement may be performed. For example, the ultrasound system determines values of two or more scattering parameters of the patient tissue. In one embodiment, the spectral slope of the logarithm of the acoustic attenuation coefficient, the backscatter coefficient and/or the frequency dependent backscatter coefficient is measured.

To measure scatter, an ultrasound scanner scans the tissue with ultrasound. A sequence of transmit and receive events is performed to acquire signals to estimate quantitative ultrasound scattering parameters. In one embodiment, a one-, two-, or three-dimensional region is scanned in a B-mode sequence (e.g., transmitting a wideband (e.g., 1-2 cycles) transmit beam and forming one or more responsive receive beams). Any scanning format may be used, such as linear, sector, or vector. The transmit and receive operations may be repeated for each scan line. Narrowband pulses (e.g., 3 or more cycles) may be transmitted and received at different center frequencies with or without overlapping frequency spectra. The narrowband transmission pulses may be used in single or multiple transmission and reception events. The transmit pulse and the corresponding receive beam may be formed at different steering angles (steering angles), such as sampling the same location of the tissue from different directions. Different steering (steering) may be performed for transmission only or for reception only. Different transmission beams may have different transmission powers and/or F-numbers. The transmission or transmissions may be focused, unfocused or use plane waves. Any scanning sequence may be used.

Repetitions, with or without different transmission and/or reception settings, may be used to measure scatter once or differently. In case multiple measurements of the same scattering parameter are provided for the same location, the measurements may be averaged or combined. Measurements from different locations, such as neighboring locations or locations within a given range, may be averaged. For example, a measure of scatter is a frequency-dependent measure averaged from multiple transmissions to the same location. Changes in the power spectrum as a function of depth, angle and/or frequency may be measured. As another example, estimates of attenuation coefficients from different transmission and/or reception angles are averaged to reduce variance or to quantify the angular dependence of attenuation.

In one embodiment, the scan to be measured is adaptive. The transmission and/or reception may be adaptive. For example, the result of one measurement is used to set the amplitude, angle, frequency and/or F # for subsequent transmissions.

In one example, the attenuation coefficient is measured. The reference phantom (refer-phantom) method is used, but other measurements of attenuation coefficient may be used. The acoustic energy has an exponential decay as a function of depth. The measurement of the sound intensity as a function of depth is performed before or without depth gain correction. To eliminate systematic effects, the measurements are calibrated based on measurements of acoustic intensity as a function of depth in the phantom. By averaging over a one-, two-or three-dimensional region, the measurement may be subject to less noise. The beamformed samples or acoustic intensities may be converted to the frequency domain and the calculations performed in the frequency domain.

In another example, the backscatter coefficients are measured. An acoustic attenuation is determined. The acoustic attenuation is used to determine a reference calibration. The scattered energy is provided as a backscatter coefficient by calibration for acoustic attenuation. The calculation may be performed in the frequency domain, providing measurements as a function of frequency.

The spectral slope of the logarithm of the frequency-dependent backscatter is measured from the backscatter coefficient. The logarithm of the backscattering coefficient is determined as a function of frequency. A line is fitted (e.g., least squares) to the logarithm of the backscatter as a function of frequency to determine the spectral slope.

In act 32, the ultrasound scanner generates a measurement of shear wave propagation in tissue from the scan of the patient. For shear wave imaging, an acoustic radiation force pulse (ARFI or pushing pulse) is delivered to the tissue. The pulse causes a displacement of the tissue at a location, resulting in the generation of shear waves. The shear wave generally travels transverse (transarsely) to the transmit beam of the push pulse. By tracking tissue displacement at one or more laterally spaced locations, shear waves passing through those locations can be detected. The time at which the shear wave travels from the origin to a later location and the distance between the locations provides the shear wave velocity.

Any shear wave parameter may be determined. For example, shear wave velocity or velocity in tissue is measured. Other shear wave parameters include angle and/or frequency dependent shear wave velocity, angle and/or frequency dependent shear wave attenuation, angle and/or frequency dependent storage modulus (storage modulus), angle and/or frequency dependent loss modulus (loss modulus), viscosity and/or angle and/or frequency dependent acoustic absorption coefficient.

The acoustic absorption coefficient comes from the absorption of the acoustic pulse, not from the absorption of the shear wave. The acoustic absorption is determined as

Where F is the radiation force, I is the strength of the ARFI push pulse, c is the acoustic speed of sound, and α is the acoustic absorption coefficient.

To measure shear waves, push pulses or ARFIs are transmitted to focal locations in the tissue. A reference scan of the stationary state tissue location is performed before the push pulse or after the tissue returns to a stationary (stopping) state. Over time, the displacement or change in position of the tissue at one or more locations spaced from the focal position is measured. The tracking scan is repeatedly performed. Using correlation or other measures of similarity, the axial, 2D or 3D displacement of tissue from a reference time compared to the current tracking time is determined. The time of maximum displacement indicates the time of the shear wave. Other timings may be used, such as the start or end of a displacement. The time of arrival of the shear wave at the tracking location and the distance from the tracking location to the focal position of the push pulse provide the shear wave velocity. Other methods may be used, such as solving for shear wave velocity at multiple locations by determining the displacement of the displacement profile (displacement as a function of time) of different tracked locations or from displacement as a function of location.

The measurement of the shear wave parameters may be a function of frequency and/or angle. The measurements are repeated by transmitting push pulses in the beam from different angles and/or at different frequencies. The measure is determined using a spatio-temporal displacement distribution in the time domain or the frequency domain. The results from the different angles can be used to determine angle-dependent measurements.

The shear wave parameters may be measured at different locations. The measurement may be based on tissue displacement for one or a single push pulse. The measurement may alternatively be based on tissue displacement for multiple push pulses. The measurements are repeated for different regions using different push pulses.

To measure the shear wave parameters, both the push pulse and the tracking transmission occur. The displacement is measured by receiving an acoustic response to the tracking transmission and not the push pulse transmission. The same scan used to measure the scattering parameters can be used to measure the shear wave parameters. For example, a reference scan that precedes the transmission of the push pulse and is used for tracking is used to measure scatter. In other embodiments, the scanning of the shear wave parameters uses different transmission and/or reception than the scattering parameters. The scan for measurements is divided into separate sequences of transmit and receive events for different measurements.

The push pulse has a relatively long duration compared to the tracking pulse, such as tens, hundreds, or thousands of cycles of the push pulse and one to three cycles of the tracking transmission. Where repetition is provided, different focus positions, frequencies, angles, powers and/or F-numbers may be used for the push pulses.

The same measurements may be repeated for the same location and/or different locations. Different frequencies, F-numbers, angles, powers, focus positions, and/or other differences may be used for any repetition. The resulting measurements may be used together to determine another measurement, or may be combined, such as averaged, to reduce noise.

An ultrasound scanner may adapt the scanning to shear wave parameter measurements. For example, for the estimation of the attenuation coefficient of the shear wave, the push pulse is adapted. The center frequency, duration, F-number, or other characteristic of the push pulse may be changed for later transmission. The focus is tighter or weaker. The displacement of the generated shear wave is larger or smaller. As another example, for an estimation of the absorption coefficient with an ARFI push pulse, another push pulse is transmitted with tighter focus or longer duration. The change may improve the signal-to-noise ratio (SNR) and/or reduce variability in the measurements.

The adaptation is based on any information. For example, the displacement profile is compared to a reference or calibration profile. As another example, the amount of displacement for the maximum, average, or median displacement is determined. This information may indicate a need for a stronger or higher intensity push pulse, or may indicate that a fewer intensity push pulse is needed, allowing for a shorter cooling time.

In act 33, the ultrasound scanner generates an ARFI measurement of the axial displacement of the tissue. ARFI transmissions cause tissue to be displaced along the axis or scan line of the transmission beam. Rather than tracking shear waves, tissue displacement along the axis caused by or in response to ARFI is tracked over time.

Any ARFI measurement may be used. For example, the attenuation of the longitudinal wave of the ARFI pulse may be estimated from the displacement tracked at a location spaced from the focal point of the ARFI. The measurement may be at the focal point or other location along the axial scan line.

For measurement, ARFI is transmitted along the scan line. The tracking scan is performed after transmitting the ARFI. Acoustic echoes from tracking transmissions along the scan lines are received. The received data is correlated with references from before or after the ARFI-induced displacement. The amount of displacement is determined as a function of time, position, transmission angle and/or transmission frequency. The maximum displacement, the displacement as a function of depth, and/or the amount of displacement as a function of time are used to calculate the ARFI measurement.

The same measurements may be performed at other times and/or locations. The results from the repetition may be used to derive another measurement or may be averaged.

The transmission may be adapted, for example, to adapt the F number, frequency, duration, power, and/or angle. The adaptation may be in response to any measure, such as the magnitude of the maximum displacement.

Other measurements may be used. The response of the tissue to different types of waves and/or scans is measured. One or more measurements of the same type are used. For a given measurement, a single instance, average, or distribution (e.g., standard deviation over time, duration, frequency, angle, and/or interval) is performed. Any number of measurements of the same or different types may be performed.

In act 34, an ultrasound scanner or other image processor estimates tissue properties of the patient's tissue from the different measurements. Measurements from two or more different wave phenomena are used. Two or more measured values are used to estimate a tissue property. For example, both measurements of scattering and measurements of shear wave propagation are used to estimate tissue properties. In another example, a measurement of on-axis displacement (e.g., an ARFI measurement) is used with a measurement of acoustic scattering and/or a measurement of shear wave propagation.

Other information may be used to estimate tissue properties. For example, clinical information of the patient is used. The clinical information may be medical history, age, body mass index, gender, fasting or not, blood pressure, diabetes or not, and/or blood biomarker measurements. Exemplary blood biomarkers include alanine Aminotransferase (ALT) levels, aspartate Aminotransferase (AST) levels, and/or alkaline phosphatase (ALP) levels. Any information about the patient may be included.

Any tissue property can be estimated. For example, the fat composition of the tissue is estimated. The fat component of the liver, breast or other tissue is diagnostically useful. The fat component in the liver of a patient helps diagnose NAFLD. Other diagnostically useful tissue properties include inflammation, density, fibrosis, and/or nephron characteristics (count and/or diameter). The tissue property is binary, such as present or absent, or an estimate along a scale (i.e., the level or amplitude of the tissue property). In one embodiment, only one tissue property is estimated. In other embodiments, two or more different tissue properties are estimated from the same or different measurements.

Acts

36 and 37 represent two different embodiments for the estimation in act 34. The different embodiments are alternative. Other embodiments may be used. Two or more embodiments may be used, such as determining the value of a tissue property in two ways and then averaging the results or selecting the most likely accurate result.

Values of tissue properties are estimated. In an embodiment of act 36, the machine learning classifier estimates tissue properties. The machine-trained classifier provides a non-linear model. Any machine learning and machine learning classifier resulting therefrom may be used. For example, support vector machines, probabilistic boosting trees, Bayesian networks, neural networks, or other machine learning is used.

Machine learning learns from training data. The training data includes various examples, such as tens, hundreds, or thousands of samples, and a ground true value (ground true). Examples include input data to be used, such as values of scattering and shear wave propagation parameters. The ground truth value is the value of the tissue property for each example. In one embodiment, machine learning classifies fat components based on scatter and shear wave propagation parameters. Magnetic Resonance (MR) scans that provide proton density fat component (PDFF) are used to provide the basic true values of fat component. MR-PDFF provides the percentage of fat at a location or region. The percentage of fat is used as a ground truth value so that machine learning learns to classify the percentage of fat from the input values of the ultrasound parameters. Other sources of ground truth values may be used for a given tissue property, such as from a biopsy, modeling, or other measurement.

In one embodiment, machine learning trains neural networks. The neural network includes one or more convolutional layers that learn a filter kernel to distinguish between values of tissue properties. Machine training learns what weighted combination of input values (e.g., convolution using a learned kernel) indicates an output. The resulting machine learning classifier uses the input values to extract discriminative information and then classifies tissue properties based on the extracted information.

The training provides one or more matrices. The one or more matrices relate input information to output classes. Hierarchical training and derived classifiers can be used. Different classifiers may be used for different tissue properties. Multiple classifiers may be used for the same tissue property, and the results averaged or combined.

In the embodiment of act 37, a linear model is used instead of or in addition to the machine learning model. A predetermined or programmed function relates an input value to an output value. The function and/or the weights used in the function may be determined experimentally. The weights are obtained, for example, by least squares minimization using MR-PDFF values.

Any linear function may be used. For example, values of tissue properties are estimated from one or more scattering parameters and one or more shear wave propagation parameters. Any combination of addition, subtraction, multiplication, or division may be used.

In one embodiment, two or more functions (e.g., a weighted combination of measurements) are provided. One of the functions is selected based on a value of one of the parameters. For example, an ultrasound-derived fat composition (UDFF) estimate includes two functions, represented as a weighted combination:

where d and δ are constants, a, b, c, α and β are weights, and P is a measure of the parameter. A parameter P_i,kFor determining which function to select. Possible functions include two or three other parameters and weights. Additional, different, or fewer numbers of functions, parameters in functions, weights, and/or constants may be used. Different selection criteria may be used. The selection parameters may be of one type and the weighting parameters of each function may be of another type. Alternatively, different types (e.g., scattering and shear wave propagation) are included as weighting parameters, regardless of the type or types of parameters used for selection.

In one example, AC is an acoustic attenuation coefficient (e.g., a scattering parameter), BSC is a backscatter coefficient (e.g., a scattering parameter), and SS is a spectral slope of a logarithm of a frequency-dependent backscatter coefficient (e.g., also a scattering parameter). SWS is the shear wave velocity (e.g., shear wave propagation parameter). Two functions based on scattering parameters are used, where the function for a given estimate is selected based on the shear wave propagation parameters, as follows:

the weights and constants are based on minimizing the difference from the fat component provided by MR-PDFF. Expert selected or other weights and/or constants may be used.

In act 38, the ultrasound scanner or display device displays the estimated tissue parameters. For example, an image of the fat component is generated. A value representing the estimated fat component is displayed on the screen. Alternatively or additionally, a graph (e.g., a curve or an icon) representing the estimated fat composition is displayed. A reference to a scale or other reference may be displayed. In other embodiments, the fat component as a function of position is displayed in a one-dimensional, two-dimensional, or three-dimensional representation by color, brightness, hue, luminosity (luminosity) or other modulation of display values. Tissue properties may be mapped to pixel colors linearly or non-linearly.

The tissue properties may be indicated separately or together with other information. For example, shear wave imaging is performed. Displaying shear wave velocity, modulus, or other information determined from tissue response to the shear wave. Any shear imaging may be used. The displayed image represents shear wave information for the region of interest or the entire imaged region. For example, where shear rate values are determined for all grid points in a region or field of view of interest, the pixels of the display represent the shear wave rate for that region. The display grid may be different from the scanning grid and/or the grid for which displacements are calculated.

The shear wave information is used for color overlay or other modulation of the display value. Color, brightness, luminosity, hue or other display property is modulated as a function of shear wave characteristics, such as shear wave velocity. The image represents a two-dimensional or three-dimensional region of the location. The cropped data is in a display format or may be scan converted to a display format. The cut data is color or gray scale data, but may be data before being mapped with gray scale or color scale (color scale). The information may be mapped to the display values linearly or non-linearly.

The image may include other data. For example, shear wave information is displayed on top of or together with B-mode information. B-mode or other data representing tissue, fluid, or contrast agent in the same region may be included, such as displaying B-mode data for any location with a shear wave rate below a threshold or with poor quality. Other data helps the user to determine where to cut the information. In other embodiments, the shear wave characteristics are displayed as an image with no other data. In other embodiments, B-mode or other image information is provided without shear wave information.

Additional estimates of tissue properties are displayed substantially simultaneously with shear waves, B-mode, color or flow mode, M-mode, contrast agent mode, and/or other imaging. Substantially accounting for the visual perception of the view. Displaying two images sequentially with sufficient frequency may allow a viewer to perceive that the images are displayed simultaneously. Component measures (component measures) for estimating tissue properties may also be displayed, for example, in a table.

Any format for substantially simultaneous display may be used. In one example, the shear wave or anatomical image is a two-dimensional image. The value of the tissue property is a text, chart, two-dimensional image, or other indicator of the estimated value. A cursor or other position selection may be positioned relative to the cropped or anatomical image. The cursor indicates the selection of the location. For example, a user selects pixels associated with an interior region of a lesion, cyst, inclusion body (inclusion), or other structure. The tissue property at the selected location is then displayed as a value, pointer along a scale, or other indication. In another example, tissue properties are indicated in the region of interest (a sub-portion of the field of view) or over the entire field of view.

In another embodiment, the shear wave or B-mode and the fat component images are displayed substantially simultaneously. For example, a dual screen display is used. A shear wave image (e.g., shear wave velocity) and/or a B-mode image is displayed in one area of the screen. The fat component as a function of position is displayed in another area of the screen. The user can view different images on the screen for diagnosis. Additional information or indications of the nature of the tissue assist the user in diagnosing the region.

In one embodiment, the tissue estimate is provided as a real-time digital or quantitative image. Since the tissue parameter can be estimated quickly, the value of the tissue parameter is estimated and output within 1-3 seconds of the start of the scan. Tissue properties may be estimated at different times, such as before, during, and/or after treatment. Estimates from different times are used to monitor the progression of the disease and/or the response to treatment. For example, the percentage change in the value of the tissue property over time is calculated and output.

The tissue properties and/or the measurements used to derive the tissue properties may be used in the estimation of disease activity. FIG. 3 is a flow diagram of one embodiment of a method for disease activity estimation with an ultrasound system. For example, the method is for ultrasound-derived non-alcoholic liver disease activity estimation. Ultrasound scanners measure the scattering and/or shear wave propagation in the tissue of a patient to directly or indirectly estimate disease activity.

For example, non-alcoholic fatty liver disease activity score (NAS) is predicted using quantitative ultrasound. NAS is predicted based on ultrasound estimation of tissue mechanical and acoustic properties. The model predicts NAS based on tissue mechanical and acoustic properties estimated using medical ultrasound. In one embodiment, the ultrasound system non-invasively obtains an ultrasound-derived NAFLD activity (UDNA) index as a predictor of NAS. The ultrasound system is configured to execute a pulse sequence for generating measurements of scattering and shear wave propagation. UDNA is determined using a model of at least three properties of the liver, including acoustic backscatter coefficients, shear wave velocity and shear wave damping ratio.

Histologic NAS is the sum of the histological scores for steatosis, lobular inflammation and ballooning, but requires a biopsy. In one embodiment, the proposed ultrasound-derived model pairs appropriate mechanical and acoustic properties with NAS features. Based on the backscatter, attenuation, and/or speed of sound, the ultrasonically derived fat components are used as a measure of the level of steatosis. Shear wave damping ratio is used as a measure of inflammation and shear wave velocity is used as a measure of ballooning. Other ultrasonic measurements may be used.

The method of fig. 3 is implemented by the system of fig. 2 or a different system. Medical diagnostic ultrasound scanners perform measurements by acoustically generating waves and measuring responses. An image processor of a scanner, computer, server, or other device makes an estimate based on the measurements. A display device, network or memory is used to output the estimated disease activity score.

Additional, different, or fewer acts may be provided. For example, act 33 from fig. 1 is included, such as using ARFI measurements in the estimation of fat composition or other tissue properties. As another example, act 34 is not included, such as where the disease activity index score is estimated from measurements without separately estimating tissue properties (e.g., fat components). In yet another example, act 38 is not provided. In another example, an action is provided for configuring an ultrasound scanner and/or scanning.

The actions are performed in the order described or illustrated (e.g., top to bottom or by number), but may be performed in other orders. For example, acts 30 and 32 are performed simultaneously, such as using the same transmit and receive pulses, or in any order.

In act 30, the ultrasound scanner generates one or more measurements of scatter in tissue from an ultrasound scan of the patient. Any acoustic scattering parameter may be used, such as a measure of acoustic interaction with liver tissue. For example, the ultrasound scanner or system measures the backscatter coefficient, frequency dependent backscatter coefficient, attenuation, speed of sound, and/or any other scattering measurement discussed above with respect to fig. 1. The measurements may be frequency dependent, such as averaged from multiple transmissions. Adaptive scanning may be used.

The measurement of scatter may be used to estimate fat composition, such as using acoustic backscatter (e.g., frequency dependent acoustic backscatter) and acoustic attenuation. Measurements of scatter are alternatively or additionally used in the estimation of disease activity, such as using acoustic backscatter or frequency dependent acoustic backscatter.

In act 32, the ultrasound scanner generates one or more measurements of shear wave propagation in tissue from an ultrasound scan of the patient. Any shear wave propagation parameter may be used. For example, shear wave velocity and shear wave damping ratio are used. Any of the shear wave propagation measurements discussed above with respect to fig. 1 may be used. The measurement is for ARFI-induced shear waves in a patient's tissue of interest, such as liver tissue. Adaptive scanning may be used.

As discussed above with respect to fig. 1, the ultrasound scans used to measure scatter and to measure shear wave propagation use the same or different transmit and receive events. For example, separate transmission and reception are used to measure scatter as compared to those used to generate shear waves and to measure tissue response to shear waves.

In one embodiment, a shear wave damping ratio is generated. Ultrasound scanning is performed to measure tissue response to shear waves in order to determine shear wave viscosity as a complex number, such as the ratio of storage modulus to loss modulus. The complex representation uses the real part of the viscosity as storage modulus and the imaginary part of the viscosity as loss modulus.

In one approach, spatiotemporal displacement measurements are acquired during shear wave propagation. These measurements are fourier transformed into the frequency domain, such as using a fast fourier transform, and used to determine the complex wavenumbers. The logarithm of the spectrum of the displacement as a function of time may be determined for each of the various locations subject to shear or other waves. Solving using a logarithm as a function of position provides a complex wave number (complex wave number). Various viscoelastic parameters, such as loss modulus and storage modulus, are determined from the complex wavenumbers. In one embodiment, measurements that determine complex wavenumber, viscosity, or other damping ratio measurements disclosed in U.S. published patent application No. 2016/0302769 are used.

As another method, shear wave attenuation and shear wave dispersion are measured. Dispersion is the change in shear wave velocity or velocity as a function of frequency. Shear wave attenuation may also be measured as a function of frequency. For a given frequency or for a combination (e.g., average) from multiple frequencies, a phasor (phasor) is generated based on the attenuation and spread values. The phasor is converted into a complex number, from which the real and imaginary parts are used as damping ratios. Other measurements of the damping ratio from the tissue response to shear waves may be used.

An image processor estimates a value for an ultrasound-derived liver disease activity index based on the backscatter coefficient, the shear wave velocity, and the shear wave damping ratio. Other measurements may be used. Non-ultrasound information, such as information from a patient medical record, may additionally be used.

Disease activity is estimated directly or indirectly from the measurements. For direct measurements, the measurements are input to a model that outputs an estimate of the value of the disease activity in act 40. The score is estimated directly from the measurements. For indirect measurements, one or more types of measurements are used to determine another value or estimate (e.g., fat composition) in act 34, and then the estimate is used alone, with other types of estimates, with other measurements, or with other types of estimates and other measurements to estimate disease activity in act 40.

In one embodiment, in act 34, the fat component is estimated using the backscatter coefficient, the acoustic attenuation, and/or the shear wave velocity (e.g., using the backscatter coefficient (acoustic scatter) and the acoustic attenuation without shear wave velocity). The fat component is estimated using one or more scattering and/or one or more shear wave propagation parameters. Any of the embodiments discussed above with respect to fig. 1 for estimating fat composition may be used. For example, the acoustic attenuation coefficient (e.g. scattering parameter), the backscatter coefficient (e.g. scattering parameter) and the spectral slope of the logarithm of the frequency-dependent backscatter coefficient (e.g. also scattering parameter) are used for estimating the fat component. Shear wave velocity (e.g., shear wave propagation parameters) may be used, such as to select a function for estimating fat composition from scattering parameters.

In act 40, the image processor estimates a level of liver disease activity (e.g., NAS or UDNA). For indirect estimation, the level of disease activity is estimated from at least one of the fat component and the shear wave parameter. For example, three inputs are used in the estimation of disease activity. The ultrasound mechanical and acoustic properties replace the histological NAS features. The ultrasonically derived fat component is used as a measure of the level of steatosis, such as based on backscatter, attenuation, and/or speed of sound. Shear wave damping ratio was used as a measure of inflammation and shear wave velocity was used as a measure of ballooning. Values for liver disease activity used to assist the index of the physician are estimated, directly or indirectly, from various measurements and/or estimates such as fat composition, damping ratio and shear wave velocity.

Other measurements and/or estimations may be used as an alternative. Multiple estimates and/or measurements may be suitably used on one histological NAS variable (used in place on). The activity may be estimated from different estimates and/or measurements than the histological variables, providing different methods of determining disease activity.

The image processor uses the model to estimate a score or value of the liver disease activity index. The fraction or value is part of an exponent, such as over a range of integers. Any range may be used, such as three levels (e.g., steatohepatitis, cirrhosis, or hepatocellular carcinoma), two levels (e.g., steatosis or fibrosis), or four or more levels. The score may relate to a particular stage of the disease, or may indicate a level of the disease without relating to a particular stage (e.g., 0-7 provide 8 levels with different amounts of activity for each level). The estimated value or score is a phase and/or numerical representation.

The model may be a function, such as using estimates and/or measurements in variables. In other embodiments, the model is a machine learning classifier. Machine learning, such as using a fully-connected neural network, a convolutional neural network, or a support vector machine, trains a model to classify-output values of disease activity given input estimates and/or measurements. In another embodiment, a logistic regression model is used. Values of disease activity are estimated from measurements and/or estimates using logistic regression. For example, logistic regression of fat composition, shear wave velocity and damping ratio is used to estimate disease activity. As another example, disease activity is estimated as a logistic regression of backscatter coefficients, shear wave velocity, and shear wave damping ratio.

The image processor may use other input information to estimate disease activity. For example, measurements of tissue response to longitudinal waves from ARFI may be used. As another example, clinical information of the patient may be used.

In act 42, the image processor generates an estimate of the disease activity and a display (e.g., a display screen) displays the estimate of the disease activity. The level of liver disease activity (e.g., UDNA) is displayed to the user to aid in disease diagnosis or monitoring. Because ultrasound is used, assistance is provided without the time and expense of invasive biopsy and/or MRI.

The level (e.g., value) of the UNDA index or other scale of disease activity is output as alphanumeric text, as a chart, or as a color coding or marker in an image representing the patient's tissue. For example, an ultrasound image of liver tissue is displayed. The image includes an indication of the value of the estimated ultrasound-derived liver disease activity index. Fat composition, other tissue property estimates and/or measurements may also be output. Instead of or in addition to the fat component, any of the outputs discussed above with respect to act 38 may be used with the disease activity.

Figure 4 shows a graph comparing predicted NAS with histological NAS. The predicted NAS is an estimate of UDNA based on logistic regression from fat composition, shear wave velocity and shear wave damping ratio based on shear wave propagation. The index uses eight levels for scoring (0-7). The root mean square error between the predicted NAS and the histological NAS was 1.14 based on 82 patients. The predicted correlation with tissue NAS is good, especially for a moderate score of 2-5. Providing good diagnostic performance.

Fig. 2 illustrates one embodiment of a system 10 for estimating tissue properties and/or disease activity from measurements in response to different types of waves. The system 10 implements the method of fig. 1, the method of fig. 3, or other methods. The system 10 includes a transmit beamformer 12, a transducer 14, a receive beamformer 16, an image processor 18, a display 20, and a memory 22. Additional, different, or fewer components may be provided. For example, user input is provided for user interaction with the system.

The system 10 is a medical diagnostic ultrasound imaging system. In alternative embodiments, the system 10 is a personal computer, workstation, PACS station, or other arrangement at the same location or distributed on a network for real-time or post-acquisition imaging.

The transmit and receive

beamformers

12, 16 form a beamformer for transmission and reception using the transducer 14. The sequence of transmit pulses is configured based on the operation of the beamformer or and the responses are received. The beamformer scans to measure scatter, shear wave and/or ARFI parameters. The

beamformers

12, 16 are configured to transmit and receive sequences of pulses in the patient using the transducer 14. The sequence of pulses is for one or more scattering parameters and for one or more shear parameters.

The transmit beamformer 12 is an ultrasound transmitter, memory, pulse generator (pulser), analog circuit, digital circuit, or a combination thereof. The transmit beamformer 12 is operable to generate waveforms for multiple channels with different or relative amplitude, delay and/or phasing (phasing). Upon transmission of acoustic waves from the transducer 14 in response to the generated electrical waveforms, one or more beams are formed. A sequence of transmission beams is generated to scan a two or three dimensional region. Sector, Vector, linear, or other scan formats may be used. The same region may be scanned multiple times using different scan line angles, F-numbers, and/or waveform center frequencies. For flow or doppler imaging and for shear imaging, a sequence of scans along one or more of the same lines is used. In doppler imaging, a sequence may include multiple beams along the same scan line before scanning adjacent scan lines. For shear imaging, scanning or frame interleaving may be used (i.e., the entire region is scanned before scanning again). A line or group of lines may be used for interleaving. In an alternative embodiment, the transmit beamformer 12 generates plane waves or diverging (diverging) waves for faster scanning.

The same transmit beamformer 12 generates pulsed excitation or electrical waveforms for generating acoustic energy to cause displacement. An electrical waveform for the acoustic radiation force pulse is generated. In an alternative embodiment, a different transmit beamformer is provided for generating the pulsed excitation. The transmit beamformer 12 causes the transducer 14 to generate a push pulse or acoustic radiation force pulse.

The transducer 14 is an array that generates acoustic energy from electrical waveforms. For an array, the acoustic energy is focused with relative delays. A given transmission event corresponds to the transmission of acoustic energy by different elements at substantially the same time at a given delay. The transmission event may provide a pulse of ultrasonic energy for displacing tissue. The pulses may be pulsed excitation, tracking pulses, B-mode pulses, or pulses for other measurements. The pulsed excitation includes a waveform having many cycles (e.g., 500 cycles), but which occurs in a relatively short time to cause tissue displacement over a longer time. The tracking pulse may be a B-mode transmission, such as using 1-5 cycles. The tracking pulses are used to scan a region of the patient.

The transducer 14 is a 1, 1.25, 1.5, 1.75 or 2 dimensional array of piezoelectric or capacitive film elements. The transducer 14 includes a plurality of elements for transducing between acoustic and electrical energy. The receive signals are generated in response to the ultrasonic energy (echoes) impinging on the elements of the transducer 14. The elements are connected to channels of the transmit and receive

beamformers

12, 16. Alternatively, a single element with mechanical focusing is used.

The receive beamformer 16 includes a plurality of channels having amplifiers, delays (delays) and/or phase rotators and one or more summers (summers). Each channel is connected to one or more transducer elements. The receive beamformer 16 is configured by hardware or software to apply relative delay, phase and/or apodization (apodization) to form one or more receive beams in response to each imaging or tracking transmission. For echoes from pulsed excitations used to displace tissue, receive operations may not occur. The receive beamformer 16 outputs data representing spatial locations using the receive signals. The relative delays and/or phasing and summing of the signals from the different elements provides beamforming. In an alternative embodiment, the receive beamformer 16 is a processor for generating samples using fourier or other transforms.

The receive beamformer 16 may include a filter, such as a filter for isolating information at the second harmonic or other frequency band relative to the transmission frequency band. Such information may more likely include desired tissue, contrast agent, and/or flow information. In another embodiment, the receive beamformer 16 includes a memory or buffer and a filter or summer. Two or more receive beams are combined to isolate information at a desired frequency band, such as a second harmonic, a third fundamental (fundamental), or another frequency band.

In coordination with the transmit beamformer 12, the receive beamformer 16 generates data representing a region. To track shear waves or axial longitudinal waves, data representing regions at different times is generated. After excitation by the acoustic pulses, the receive beamformer 16 generates beams representing positions along one or more lines at different times. Data (e.g., beamformed samples) are generated by scanning the region of interest with ultrasound. By repeating the scan, ultrasound data representing regions at different times after the pulse excitation is acquired.

The receive beamformer 16 outputs beam summation data representing spatial locations. Outputting data for a single location, a location along a line, a location of a region, or a location of a volume. Dynamic focusing may be provided. The data may be used for different purposes. For example, different portions of the scan are performed for B-mode or tissue data rather than displacement. Alternatively, the B-mode data is also used to determine the displacement. As another example, different types of measured data are acquired with a series of shared scans, and B-mode or doppler scans are performed separately or using some of the same data.

The image processor 18 is a B-mode detector, doppler detector, pulsed wave doppler detector, correlation processor, fourier transform processor, application specific integrated circuit, general purpose processor, control processor, image processor, field programmable gate array, digital signal processor, analog circuit, digital circuit, combinations thereof or other now known or later developed device for detecting and processing information from beamformed ultrasound samples for display. In one embodiment, the image processor 18 includes one or more detectors and a separate image processor. The individual image processors are control processors, general processors, digital signal processors, application specific integrated circuits, field programmable gate arrays, networks, servers, groups of processors, data paths, combinations thereof, or other now known or later developed devices for calculating values of different types of parameters from beamformed and/or detected ultrasound data and/or for estimating from values from different types of measurements. For example, a separate image processor is configured by hardware, firmware, and/or software to perform acts 34-38 shown in FIG. 1 and/or acts 34-42 shown in FIG. 3.

The image processor 18 is configured to estimate values of tissue properties and/or disease activity from a combination of different types of parameters. For example using a measured scattering parameter and one, two or more measured shear wave parameters. Different types of parameters are measured based on the transmission and reception sequences and calculations from the results. One or more measured values for each of at least two types of parameters (e.g., scattering, shear wave propagation, or axial ARFI) are determined for fat component estimation. Values of one or more scattering and one or more (e.g., two) shear wave propagation parameters are determined for liver disease activity estimation.

In one embodiment, the image processor 18 estimates tissue properties based on different types of parameters or measurements of tissue response to different types of wavefronts. The estimation applies a machine learning classifier. The measured input values, with or without other information, are used by the learning matrix to output values of tissue properties. In other embodiments, the image processor 18 uses a weighted combination of the values of the parameters. For example, two or more functions are provided. One of the functions is selected using values of one or more parameters (e.g., shear wave velocity). The selected function uses values of the same and/or different parameters to determine the value of the tissue property. The linear or non-linear mapping relates the values of one or more parameters to the values of the tissue property. For example, two or more scattering parameters are used to determine the value of a tissue property using a shear wave propagation selection function.

In another embodiment, the image processor 18 is configured to generate a score for an index of disease activity from a combination of one or more scattering parameters and one or more (e.g., two) shear wave parameters. For example, the image processor 18 estimates the fat composition of the patient's liver from one or more scatter parameters. The image processor 18 generates a score of non-alcoholic liver disease activity derived as ultrasound from the fat component and the two or more shear wave parameters. The scores are generated using a machine learning classifier or logistic regression model. For example, logistic regression models relate scattering (e.g., acoustic backscatter coefficients) and two or more shear wave parameters (e.g., shear wave velocity and shear wave damping ratio) to the level of disease activity.

The processor 18 is configured to generate one or more images. For example, shear wave velocity, B-mode, contrast agent, M-mode, flow or color mode, ARFI, and/or another type of image is generated. Shear wave velocity, flow, or ARFI images may be presented alone or as an overlay or region of interest within a B-mode image. Shear wave rate, flow, or ARFI data modulates color at locations in the region of interest. In the event that the shear wave rate, stream, or ARFI data is below a threshold, the B-mode information may be displayed without modulation by the shear wave rate.

Other information is included in the image or displayed sequentially or substantially simultaneously. For example, the tissue property estimation image and/or the disease activity level are displayed simultaneously with other images. One or more values of the tissue property and/or disease activity map may display information. Where tissue properties and/or disease activity are measured at different locations, values of the tissue properties and/or disease activity may be generated as a color overlay in a region of interest in the B-mode image. Shear wave velocity, tissue properties, and/or disease activity data may be combined into a single overlay on one B-mode image. Alternatively, one or more values of tissue properties and/or disease activity are displayed as text or numerical value(s) adjacent to or overlaid on the B-mode or shear wave imaging image. The image processor 18 may be configured to generate other displays. For example, shear wave rate images are displayed alongside graphs, text, or graphical indicators of tissue properties, such as fat composition and/or degree of fibrosis, and/or disease activity, such as index values indicating UDNA levels. Tissue property information and/or disease activity is presented for one or more locations of the region of interest without being presented in a separate two-dimensional or three-dimensional representation, such as where a user selects a location and the ultrasound scanner then presents the tissue property and/or disease activity for that location.

The image processor 18 operates in accordance with instructions stored in the memory 22 or another memory for estimating tissue response measurements for different types of waves (e.g., scattering from transmitted ultrasound, on-axis tissue displacement, and/or shear waves caused by tissue displacement). The memory 22 is a non-transitory computer-readable storage medium. The instructions for implementing the processes, methods, and/or techniques discussed herein are provided on a computer-readable storage medium or memory, such as a cache, buffer, RAM, removable media, hard drive, or other computer-readable storage medium. Computer-readable storage media include various types of volatile and nonvolatile storage media. The functions, acts or tasks illustrated in the figures or described herein are performed in response to one or more sets of instructions stored in or on computer readable storage media. The functions, acts or tasks are independent of the particular type of instructions set, storage media, processor or processing strategy and may be performed by software, hardware, integrated circuits, firmware, micro code and the like, operating alone or in combination. Likewise, processing strategies may include multiprocessing, multitasking, parallel processing and the like. In one embodiment, the instructions are stored on a removable media device for reading by local or remote systems. In other embodiments, the instructions are stored at a remote location for transmission over a computer network or over telephone lines. In other embodiments, the instructions are stored within a given computer, CPU, GPU, or system.

The display 20 is a device for displaying one-or two-dimensional images or three-dimensional representations, such as a CRT, LCD, projector, plasma, or other display. The two-dimensional image represents a spatial distribution in the area. A three-dimensional representation is rendered from data representing a spatial distribution in a volume. The display 20 is configured by the image processor 18 or other device by inputting signals to be displayed as an image. The display 20 displays an image representing tissue properties and/or disease activity (e.g., averaged from tissue property estimates including neighboring locations) of a single location in the region of interest, or the entire image. For example, the display 20 displays the value of the fat component and/or the score of the disease activity index. The display of tissue properties and/or disease activity based on different types of waves provides a more accurate level of tissue properties or disease information for diagnosis.

Although the invention has been described above by reference to various embodiments, it should be understood that many changes and modifications may be made without departing from the scope of the invention. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this invention.

Claims

1. A method for non-alcoholic liver disease activity estimation with an ultrasound scanner, the method comprising:

generating (30) a first measure of scatter in tissue from a scan of a patient by the ultrasound scanner, the first measure of scatter comprising a backscatter coefficient;

generating (32) second and third measurements of shear wave propagation in the tissue from the scan of the patient by the ultrasound scanner, the second measurement comprising a shear wave velocity and the third measurement comprising a shear wave damping ratio;

estimating (40) a first value of an ultrasound-derived liver disease activity index from the backscattering coefficient, the shear wave velocity and the shear wave damping ratio; and

outputting (42) an ultrasound image comprising an indication of the estimated first value of the ultrasound-derived liver disease activity index.

2. The method of claim 1, wherein generating (30, 32) the first, second, and third measurements from the scan comprises separate transmit and receive events for (1) the first measurement of scatter and (2) the second and third measurements of shear wave propagation.

3. The method according to claim 1, wherein generating (30) the first measure of scatter comprises generating (30) a frequency-dependent backscattering coefficient as the backscattering coefficient.

4. The method according to claim 1, wherein generating (32) the third measurement comprises generating (32) the shear wave damping ratio as a ratio of real and imaginary parts of a complex number from a fourier transform of a spatio-temporal displacement caused by the shear wave propagation.

5. The method according to claim 1, wherein generating (32) the third measurement comprises generating (32) the shear wave damping ratio as a function of shear wave attenuation and shear wave dispersion.

6. The method according to claim 1, wherein estimating (40) the first value comprises estimating (34) a fat component of the patient's liver from acoustic attenuation, the backscatter coefficient and the shear wave velocity, and estimating (40) the first value of the ultrasound-derived liver disease activity index from the fat component, the damping ratio and the shear wave velocity.

7. The method according to claim 6, wherein estimating (40) the first value comprises estimating (40) using logistic regression of the fat component, the shear wave velocity and the damping ratio.

8. The method of claim 1, wherein estimating (40) comprises estimating (40) with a machine learning classifier.

9. The method of claim 1, wherein estimating (40) comprises estimating (40) using a logistic regression model.

10. The method of claim 8, wherein estimating (40) with the logistic regression model comprises estimating (40) as logistic regression of the backscatter coefficients, the shear wave velocity, and the shear wave damping ratio.

11. The method of claim 1, wherein generating (30) the measure of scatter comprises generating (30) the measure of scatter as a frequency-dependent measure averaged from a plurality of transmissions.

12. The method of claim 1, wherein generating (30, 32) the first, second, and third measurements comprises adaptive scanning.

13. The method according to claim 1, wherein estimating (40) includes estimating (40) as a function of clinical information for the patient.

14. A system for estimation of disease activity, the system comprising:

a transducer (14);

a beamformer (12, 16) configured to transmit and receive with the transducer (14) a sequence of pulses in a patient, the sequence of pulses being for a scatter parameter and for a first and a second shear wave parameter;

an image processor (18) configured to generate a score for an index of the disease activity as a function of a combination of the scattering parameter, the first shear wave parameter and the second shear wave parameter; and

a display (20) configured to display the score of the index of the disease activity.

15. The system as recited in claim 14, wherein the image processor (18) is configured to generate the score with a machine learning classifier.

16. The system according to claim 14, wherein the image processor (18) is configured to generate the score using a logistic regression model of the scattering parameter, the first shear wave parameter, and the second shear wave parameter.

17. The system of claim 14, wherein the scattering parameter comprises an acoustic backscatter coefficient, the first shear wave parameter comprises a shear wave velocity, and the third shear wave parameter comprises a shear wave damping ratio.

18. The system as set forth in claim 14, wherein the image processor (18) is configured to estimate a fat component of the patient's liver from the scattering parameters, and the image processor (18) is configured to generate the fraction of non-alcoholic liver disease activity derived as ultrasound from the fat component, the first shear wave parameter, and the second shear wave parameter.

19. A method for liver disease activity estimation with an ultrasound system, the method comprising:

determining (30), by the ultrasound system, a plurality of scattering parameters of liver tissue of a patient;

determining (32), by the ultrasound system, a plurality of shear wave parameters of the liver tissue of the patient;

estimating (34) a fat component from at least one of the scattering parameters;

estimating (40) a level of liver disease activity from at least one of the fat component and the shear wave parameter; and

displaying (42) the level of liver disease activity.

20. The method according to claim 19, wherein estimating (34) the fat component includes estimating (34) from acoustic scattering and acoustic attenuation as the scattering parameters, and wherein estimating (40) the level of liver disease activity includes estimating (40) from the fat component and from shear wave velocity and shear wave damping ratio as the shear wave parameters.