Unwrapped spectral doppler envelope tracking in medical ultrasound

Abstract

For spectral Doppler imaging, the search range for tracking the spectral envelope is dynamically set. The limits of the search envelope are established by a spectral-by-spectral placement between frequency bands. This search can be aided by plotting the spectrum from 0 to 2π. The limits vary over time to better separate the frequency bands, thus avoiding aliasing for subsequent tracking.

Description

BACKGROUND

[0001] This embodiment relates to Doppler mode (e.g., spectral) imaging. A Doppler response is generated by transmitting a plurality of pulses (e.g., pulse waves (PW)) at a location. For spectral Doppler, the spectrum of object motion or flow for a single spatial region is estimated and displayed as a function of time. Spectral Doppler ultrasound imaging provides an image of the spectrum as a velocity value (vertical axis) modulated by energy as a function of time (horizontal axis) for a gate. The spectrum can be used to study fluid flow or tissue motion within a patient.

[0002] Spectral Doppler can suffer from aliasing. If spectral aliasing is present, the envelope can be displayed incorrectly. In spectral envelope tracking where spectral aliasing is present, the tracking will not correctly follow the peak. Instead, the trace is trimmed at any location between the Nyquist, baseline, or both, depending on the spectral bandwidth and signal-to-noise ratio (SNR). While aliasing can sometimes be corrected with a baseline offset, correction is not always possible even if the sampling frequency is less than the maximum Doppler shift. This is because the search range for tracking is typically fixed at the Nyquist limit determined by the baseline position. SUMMARY

[0003] By way of introduction, the preferred embodiments described below include a method, system, computer readable medium, and instructions for spectral Doppler imaging. The search range for tracking a spectral envelope is dynamically set. The limits of the search envelope are established by a spectral-by-spectral placement between frequency bands. Plotting the spectrum from 0 to 2π can aid this search. The limits vary over time to better separate the frequency bands, thereby causing subsequent tracking to avoid aliasing.

[0004] In a first aspect, a method for spectral Doppler imaging is provided. An ultrasound system acquires samples representing a response from a range gate. A Doppler estimator estimates a spectrum of the range gate over time from the samples. The spectrum is plotted over time from 0 to 2π. For each spectrum in the spectrum, a point in a noise region is identified between a first signal and a second signal, or between the first signal and 0, or between the first signal and 2π. A spectral envelope is tracked over time. The tracking is limited by the point of the spectrum. The spectral envelope or information derived from the spectral envelope is displayed in an image.

[0005] In one embodiment, the samples are acquired by transmitting at a pulse repetition frequency and acquiring one of the samples in response to each repetition. In another embodiment, a Fourier transform is applied to the samples to make the estimates. Each spectrum in the spectrum is energy as a function of frequency over a time period.

[0006] According to one embodiment, plotting includes plotting each frequency spectrum to start at 0 and end at 2π, such that negative signals extend from 2π toward 0 and positive signals extend from 0 toward 2π.

[0007] In various embodiments, identifying a point by positioning the point midway between a first signal and a second signal, or between the first signal and 0, or between the first signal and 2π; searching for a frequency band having the first signal and a frequency band having the second signal in the spectrum, and (1) positioning a point in a noise region between the first signal and the second signal when the second signal is found in the search, or (2) positioning a point in a noise region between the first signal and 0 or 2π when the second signal is not found; determining a noise level from at least some of the frequency spectra, and using the noise level to search for the first signal and the second signal; and / or positioning a boundary over time, the boundary formed by points of a frequency spectrum and varying with respect to 0 and 2π.

[0008] In another embodiment, tracking includes positioning an edge of the first signal having the noise region and positioning an edge of the second signal having the noise region, the positioning of the edges being over time, and the edges being the frequency spectrum envelope. In yet another embodiment, tracking includes using 0 or 2π for a frequency spectrum that extends to 0 or 2π for the noise region and a first signal or a second signal that does not extend to 0 or 2π. As another embodiment, tracking includes tracking the first signal as a positive signal between 0 and a point of the frequency spectrum, and tracking the second signal as a negative signal between 2π and a point of the frequency spectrum.

[0009] According to one embodiment, displaying includes displaying the frequency spectrum envelope, with positive above a baseline and negative below the baseline. As another embodiment, displaying includes displaying information based on a peak velocity of the frequency spectrum envelope.

[0010] In a second aspect, a method for spectral Doppler imaging is provided. An ultrasound system acquires samples representing a response from a range gate. A Doppler estimator estimates a frequency spectrum of the range gate over time from the samples. The frequency spectrum is plotted from 0 to 2π over time. Positive and / or negative frequency spectrum envelopes are detected over time based on the frequency spectrum plotted from 0 to 2π. The frequency spectrum envelope or information derived from the frequency spectrum envelope is displayed in an image.

[0011] In one embodiment, plotting includes plotting each frequency spectrum to start at 0 and end at 2π, such that negative signals extend from 2π toward 0 and positive signals extend from 0 toward 2π. In one embodiment, detecting includes setting a boundary that varies from 0 to 2π from frequency spectrum to frequency spectrum, and searching for a positive frequency spectrum envelope on one side of the boundary and a negative frequency spectrum envelope on the other side of the boundary. According to another embodiment, setting the boundary includes setting the boundary at an intermediate between signals of different frequency bands, where 0 or 2π is used in place of signals for frequency spectra of the frequency bands that have no signal.

[0012] In a third aspect, a system for spectral Doppler imaging is provided. A beamformer is configured to sample a gate at a pulse repetition interval established in response to a velocity scale. A Doppler estimator is configured to generate a plurality of frequency spectra from the sampling of the gate. A signal processor is configured to identify a line in a noise region between 0 and 2π that varies from frequency spectrum to frequency spectrum in the frequency spectra over time, and limit a search for a frequency spectrum envelope of a frequency spectrum by the line. A display is configured to display the frequency spectrum envelope or information derived from the frequency spectrum envelope.

[0013] As one embodiment, the signal processor is configured to plot the frequency spectra from 0 to 2π such that one of the two frequency bands is through 2π and the other of the two frequency bands is through 0, and identify a line plotted in the noise region from the frequency spectra. In yet another embodiment, the signal processor is configured to position the line from frequency spectrum to frequency spectrum at an intermediate between the two frequency bands.

[0014] The present invention is defined by the following claims, and no limitation from any portion of this section shall be regarded as limiting those claims. Other aspects and advantages of the present invention will become apparent from the following detailed discussion, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0015] The components and the drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. Moreover, in the drawings, like reference numerals designate corresponding parts throughout the several views.

[0016] Figure 1 Aliasing in spectral Doppler imaging is shown.

[0017] Figure 2 is a flowchart of one embodiment of a method for spectral Doppler imaging with improved tracking limits to avoid aliasing.

[0018] Figure 3 An example plot of a frequency spectrum from 0 to 2π is shown.

[0019] Figure 4 An example dynamic boundary or search limit identified for Figure 3 plotting of

[0020] Figure 5 is a block diagram of one embodiment of a system for spectral Doppler imaging with reduced aliasing. DETAILED DESCRIPTION

[0021] Figure 1 Three spectral Doppler images are shown with different baseline settings. Each column in each image represents a spectrum. Given the sampling rate used to form the spectrum, the spectrum is plotted in the velocity scale over time based on the Nyquist criterion. In the top spectral band, the baseline is set to 0, so the velocity scale ranges from -π to π. The result is aliasing 100, where some of the positive signals are placed at the negative portion of the band (i.e., below the baseline). The middle band shows a baseline offset of -0.25 (25% negative offset). The result is negative signal aliasing into the positive region. When a trace 110 is detected, this negative signal is tracked as a positive peak. The bottom band shows further aliasing resulting from further baseline offset to -0.5 (50% negative offset). Similarly, a positive baseline offset would result in greater aliasing of positive signals into the negative.

[0022] The method herein avoids aliasing. Even in cases where baseline offset does not work, the spectral Doppler envelope is tracked without aliasing. The search range for positive and negative flow peaks is dynamically adjusted, rather than fixed by the baseline position setting. For each spectral column, the boundary between positive and negative flow is dynamically determined, which is then used as the Nyquist limit for subsequent positive and negative searches. The search limits are dynamically determined based on spectral characteristics such as bandwidth and signal-to-noise ratio (SNR). In this way, any existing peak search works correctly and produces an envelope that is consistent with the spectral content. By dynamically adjusting the search range, true peak velocities in either direction are correctly detected even in the presence of spectral aliasing. Likewise, the envelope can be correctly displayed. Both the positive and negative flow envelopes are correctly tracked in the presence of spectral aliasing. Cases where the envelope becomes scrambled due to aliasing of the Doppler signal itself are handled, improving the ability to extract diagnostically useful information compared to envelope tracking using prior art. When the maximum Doppler offset is less than the sampling rate, both the positive and negative envelopes are correctly displayed without aliasing artifacts.

[0023] Figure 2 One embodiment of a flowchart of a method for spectral Doppler imaging is shown. By plotting the spectrum in a way that distinguishes between positive and negative signals, a variable limit is identified between two bands of the spectrum. This limit is then used to track positive and / or negative signals, thereby avoiding aliasing.

[0024] The method is summarized by Figure 5The system 500 or a different system implementation. For example, an ultrasound system (e.g., a beamformer and transducer) or another ultrasound scanner acquires the samples. A processor, e.g., a Doppler estimator, estimates the spectrum, plots, detects / identifies, and tracks from the spectrum. The ultrasound system or scanner displays an image of the tracked spectral envelope or information from the spectral envelope. One or more actions can be performed through interaction with a user (e.g., a gate placement for sampling). Other actions or all actions can be performed automatically without user input, except for initial activation or gate location determination.

[0025] The actions are performed in the order shown, but other orders are possible. Additional, different, or fewer actions can be provided. For example, action 250 is not performed. As another example, actions 230 and / or 232 are not performed. In another example, action 220 is not performed. In yet another example, actions are provided for gate placement, gate size adjustment, velocity scale setting, pulse repetition frequency setting, filtering, processing, maximum velocity determination over time, or other Doppler functions.

[0026] The method is implemented for pulsed wave (PW) spectral Doppler. In PW, the gate location is sampled using pulsed wave transmission interleaved with echo reception (e.g., 1-50 cycles of a waveform per pulse). The PW can be interleaved with other imaging modes, e.g., B-mode or flow-mode. Alternatively, continuous wave Doppler is used. With continuous wave Doppler, samples are created that can be transformed to create a spectrum.

[0027] For spectral Doppler imaging, a sampling gate, range gate, or spectral Doppler gate is positioned. For example, a B-mode and / or flow-mode scan is performed. The user indicates a gate location on the resulting image. In other examples, the gate is positioned automatically, e.g., at the location of maximum Doppler velocity or energy determined from flow-mode data. The user or an automated process also determines the size of the gate, e.g., a depth or length in range. The lateral extent and / or range extent can be based on data, a user setting, or a default value otherwise selected.

[0028] In action 200, the ultrasound system acquires samples representing a range gate location. The samples are beamformed samples, but can be other raw data (e.g., channel data). In alternative embodiments, the samples are acquired by loading from memory or transferring from another device.

[0029] To acquire by scanning, the transducer emits a plurality of acoustic energy beams in sequence. The acoustic or ultrasound energy of each emission is focused at or near the gate. The sequence of emissions is performed. Enough samples are acquired to perform spectral or other Doppler analysis. Any set number (e.g., 3-512) of emission beams are emitted so that a spectrum of the response from the Doppler gate can be estimated.

[0030] By performing additional emissions, additional information is obtained for estimating the spectrum at other times or periods. A given response to a given beam can be used for different spectra (i.e., different times), such as where a moving window of received responses or samples is used to generate each spectrum.

[0031] For a velocity scale set for spectral Doppler imaging, a sufficient number of samples are acquired. The ultrasound system uses the velocity scale for imaging. The velocity scale defines a frequency range over which the spectrum or velocity is estimated and displayed. The velocity scale is selected by a user, is a default or predetermined value of the system, and / or is determined adaptively by the ultrasound system. Based on the velocity scale, the emissions are performed at a pulse repetition frequency or interval so as to sample the motion or flow signal at the gate location.

[0032] In response to the emissions, the transducer receives acoustic echoes. A receive beamformer samples the echoes to acquire a receive signal for the gate. A receive beam is formed by focusing the receive signal to coherently combine data representing the gate. This combined data representing the gate is a beamformed signal or sample.

[0033] The receive operation occurs repeatedly in response to the repeated emissions. The beamformed samples from the gate location are received at different times. A sample is received in response to each emission. The samples of the same location are acquired in a set over time. For Doppler analysis, a set of samples from the same location is acquired. The samples can be obtained in a continuous manner so that a moving window (e.g., set or stream sample count) of any step size (e.g., every sample or every third sample) is used to estimate the spectrum.

[0034] In act 210, a Doppler estimator estimates a spectrum for the Doppler gate from the samples of the set or stream sample count. By repeating for other set or stream sample counts, a plurality of spectra are estimated for a corresponding plurality of times or periods (i.e., over time).

[0035] Each spectrum represents the energy as a function of frequency or velocity over a same period of time. Frequency and velocity have a known relationship, so velocity is represented with frequency and frequency is represented with velocity.

[0036] The spectrum is estimated by applying a Fourier transform, wavelet transform, or Wigner-Ville distribution to a sequence of ultrasound responses or samples. Any transform can be applied to determine each spectrum.

[0037] The velocity scale is used to estimate the spectrum. The signal from the fluid or tissue is in a range of positive and negative velocities. The range used in the estimation is the velocity scale. Any velocity beyond the velocity scale wraps around or alias. The spectrum provides the energy as a function of frequency in the frequency range set by the velocity scale. The baseline determines the band range within the velocity scale.

[0038] The spectrum is estimated for the Doppler gate. The spectrum is estimated from the ultrasound samples in the sample sequence from the Doppler gate. Each spectrum corresponds to a time period in which the samples were taken.

[0039] Actions 200 and 210 are repeated for different times or time periods. To create a spectrum band, the spectrum is determined for each different time. The spectrum for a given time (period) in the spectrum band is mapped as having velocity on the vertical axis and modulation intensity. The spectrum is distributed in time along the horizontal axis. Other mappings can be used.

[0040] By repeating these actions, the spectrum is acquired for different times. As more samples are acquired, more samples are added to the group and old samples are removed. The spectrum sequence represents the Doppler gate at different times. Other spectra can be estimated for other time periods or different times corresponding to different time periods or sets of acquisitions. The time periods can overlap, for example when using a moving window with a step size that is less than the set time period, or they can be unique. The sample acquisition and estimation for different time periods is repeated to provide a spectrum over time. For the spectrum band, the process and corresponding repetition is ongoing or occurs a number of times.

[0041] In action 220, the signal processor plots the spectrum from 0 to 2π over time. The spectrum is plotted from 0 to 2π rather than -π to π using a baseline of 0 or other offset (e.g., 25% or -50%) in the baseline that is less than 100% (e.g., -.75π to 1.25π). The spectrum can be plotted substantially from 0 to 2π, where substantially accounts for an offset of 10% or less.

[0042] Figure 3 An example is shown. In Figure 3 the "0" on the y-axis is 2π and the "256" on the y-axis is 0. The black or dark regions represent noise and the white or gray regions represent signal.

[0043] By plotting each spectrum starting at 0 and ending at 2π, the negative signals are in the upper region (i.e., extending along or from 2π to 0) and the positive signals are in the lower region (i.e., extending along or from 0 to 2π). In alternative embodiments, the plot is from 2π to 0 so that the positives are in the upper region and the negatives are in the lower region. Either of 0 to 2π or 2π to 0 is plotted from 0 to 2π over time.

[0044] In a typical spectral band, this plotting method would not be used. The spectral band would show positive signals in the upper half and negative signals in the lower half, where, due to baseline offset, separation could be used in addition to the half to avoid aliasing. Unlike this typical method, the plotting of action 220 is designed to better separate the positive and negative signals, with noise in between. This plotting from 0 to 2π or 2π to 0 places the noisy region 340 between the signals and avoids aliasing.

[0045] Essentially, plotting from 0 to 2π or 2π to 0 helps to separate frequency bands without aliasing. In an alternative embodiment, the velocity scale is plotted along with the baseline at any location. Machine learning models or signal processing are applied to distinguish frequency bands to set boundaries, regardless of how they are plotted. These boundaries separate the frequency bands and therefore vary over time or per spectrum.

[0046] exist Figure 2 In action 230, the signal processor detects positive and / or negative spectral envelopes over time based on a spectrum plotted from 0 to 2π. It also detects edges of noisy signals. Figure 3 In the example, there are two frequency bands (positive and negative). The edges of each frequency band are detected separately. If only one frequency band is represented in one or more spectra, the envelope is detected for that single frequency band (e.g., only positive or negative). More than two frequency bands can be detected.

[0047] Action 232 represents a method for individual detection of a spectral envelope or frequency band. Each spectrum is divided into frequency band regions defined by limits in noise region 340. In other embodiments, noise-based thresholding is used such that the spectrum is divided into noise-free frequency bands.

[0048] In action 232, the signal processor identifies points within a noise region—between one signal and another, between one signal and 0, or between one signal and 2π—for each spectral point in the spectrum. These points (over time) define boundaries over time. These boundaries are formed by the points in the spectrum. Because these points are located spectrally, the boundaries formed by the points in the spectrum vary relative to 0 and 2π.

[0049] To identify each point of the boundary, the signal processor searches each frequency spectrum. The signal processor searches one or more frequency bands, such as the bands for positive signals and the bands for negative signals. To search for the edges of the frequency bands, the noise level is determined to distinguish noise from the signal. The amplitude of energy modulation is used to define the noise level. Statistical analysis reveals the noise floor. For example, energy is sorted by magnitude. Transitions in magnitude are determined from the sorted energy, for example, by slope. Strong signals are identified from weak signals to set the noise level or threshold that distinguishes the signal from noise.

[0050] The edge separating the strong signal from the noise signal (i.e., the noise threshold) is then found. The edge can be found using multiple energies in a row that are above the noise. The edge is divided into separate regions. In an alternative approach, the energies of a group or all of the spectra are low pass filtered, removing outliers. A threshold is then applied to identify signal regions or noise regions 340. Other approaches can be used that search for signals or different frequency bands using noise levels.

[0051] Once the edge is located for a spectrum, a point is located in the noise region 340 between the signal, frequency band, or edge. Figure 4 An example is shown. The spectral edges of the positive signal 420 and the negative signal 410 are used to set a point per spectrum, forming a boundary 400 in the noise region 340. In this example, each point and the resulting boundary 400 is placed halfway between the edges or envelopes of the two frequency bands. For example, based on the edge of the negative signal 410 at 10 and the edge of the positive signal 420 at 125, the time point 400 is approximately 57. As another example, based on the edge of the negative signal 410 at 100 and the edge of the positive signal at 256, the time point 1 is 178. Other divisions other than half can be used. More than one boundary can be used, for example, setting a point based on each frequency band being outside of a certain range or by %, (e.g., 1 / 3 of the noise band from the edge 410 and 1 / 3 of the noise band from the edge 420, with 1 / 3 being the noise in between).

[0052] In the case where there is no positive or negative signal for a given spectrum, 0 or 2π is used as the envelope or edge 410, 420. In the example, 0 is used for the positive edge 420 from approximately times 0-75, 125-375, 525-630, 700-950, 1075-1225, and 1275-1500, and 2π is used for the negative edge 410 from 75-125, 700-725, 950-1050, and 1225-1275. For each spectrum, the point is located in the noise region 340 between the positive signal or the negative signal and (1) the other of the negative signal or the positive signal when the other of the negative signal or the positive signal is found in the search, or (2) 0 or 2π when the other signal is not found. Figure 4

[0053] The boundary 400 is a limit used in the search for tracking. The search for the envelope of a spectrum uses the limit or velocity defined by the boundary 400, rather than using a baseline set to separate the frequency bands. The boundary 400 acts as a dynamic or variable Nyquist limit, which has different values (points) for different times, avoiding aliasing.

[0054] In the case where there is no positive or negative signal for a given spectrum, 0 or 2π is used as the envelope or edge 410, 420. In the example, 0 is used for the positive edge 420 from approximately times 0-75, 125-375, 525-630, 700-950, 1075-1225, and 1275-1500, and 2π is used for the negative edge 410 from 75-125, 700-725, 950-1050, and 1225-1275. For each spectrum, the point is located in the noise region 340 between the positive signal or the negative signal and (1) the other of the negative signal or the positive signal when the other of the negative signal or the positive signal is found in the search, or (2) 0 or 2π when the other signal is not found. Figure 2 ​In act 240, the signal processor tracks the spectral envelope over time. The same methods used to detect the spectral envelope or otherwise find the edges of the band or signal in act 230 can be used. Different methods can be used. For example, a less-processed but less-precise band edge or signal-to-noise boundary detection is performed in act 230 to find the boundaries in act 232, and a more-processed and more-precise band edge or signal-to-noise boundary detection is performed in act 240 to perform the tracking. As another example, the detection of act 230 is performed without limits other than 0 and 2π, and the tracking of act 240 limits where the search for edges occurs. Any now-known or later-developed spectral envelope tracking can be used.

[0055] In another embodiment, a peak velocity search is performed. The tracking is performed over all or only a portion of time (one or a small number of spectra). The tracking can be less than all of the spectra.

[0056] The tracking locates the edges of the signal over time or at a certain time. For example, the maximum positive signal and / or the maximum negative signal over one or more cardiac cycles is determined. A noise threshold is found from the spectrogram and used as a threshold to identify the signal. A trajectory is placed at a location where the signal is primarily on one side of the trajectory and the noise is primarily on the other side of the trajectory. The trajectory locates the edges of the signal on the spectrum. The maximum of that trajectory is then found.

[0057] The tracking of act 240 is limited by the points that form the band separation. Rather than relying on a baseline location to define where positive and negative are separated, the dynamic or variable boundary identified in act 232 is used. Since the boundary varies over the spectrum, only positive signals are provided on the positive signal side of the boundary and only negative signals are provided on the negative signal side of the boundary. The search for the positive spectral envelope is on one side of the boundary and the search for the negative spectral envelope is on the other side of the boundary. Aliasing is avoided.

[0058] The signal processor locates the edges of each band (e.g., positive and negative signals) with one or more noise regions. The edges extend over time or the spectrum. The located edges are the spectral envelope. Using the located edges, the signal processor can find the maximum positive signal and the maximum negative signal over time or at a certain time. Figure 4 As an example of tracking rather than an example of identifying boundaries 400, trajectories 410 and 420 are located as spectral edges. The spectrum can be re-plotted, for example, based on a baseline location (e.g., baseline 0, so re-plot from -π to π). Boundaries 400 are used to limit the tracking so that aliasing is avoided even in the case of re-plotting. Any aliased signals are not included in the false signals because boundaries 400 indicate the actual band membership. The positive signals are tracked between 0 to the point or boundary 400, and the negative signals are tracked between 2π and the point or boundary 400, resulting in a spectral envelope without aliasing regardless of the baseline location.

[0059] In another embodiment, a peak velocity search is performed. The tracking is performed over all or only a portion of time (one or a small number of spectra). The tracking can be less than all of the spectra.Figure 2 In act 250, the signal processor generates an image on the display. The image has the spectral envelope or has information as a function of the spectral envelope.

[0060] In one embodiment, a spectral band or spectral Doppler image is generated for the Doppler gate. Filtering can be applied to smooth the spectrum along the time and / or frequency dimension or in energy. The spectral band shows the frequency modulated by energy as a function of time. Any now known or later developed spectral band mapping can be used, for example with a gray scale mapping representing the intensity of the energy. The energy modulates the pixels. The gray scale or color is mapped from the energy values.

[0061] The tracked spectral envelope is also displayed. The tracked spectral envelope is included in the image, for example highlighted by coloring or overlaying a graphic (e.g., a curve or line along the edge). Different colors can be used for different frequency bands. Additional information can be included, for example an annotation with alphanumeric text showing the maximum value (e.g., maximum positive velocity) of one or more frequency bands.

[0062] A velocity scale defines the vertical range on the spectral band. A baseline defines the separation of positive and negative within the velocity scale. The spectral envelope is displayed as positive above the baseline and negative below the baseline. As additional samples are acquired, the resulting spectrum at the different times is added to the spectral band, for example adding the spectrum to the right of the band, shifting the remaining spectrum to the left by one time step, and removing the leftmost spectral band. Another update or scrolling of the spectral band can be used.

[0063] The spectral band can or can not be displayed with a spatial image, for example a one-dimensional M-mode, two-dimensional B-mode, two-dimensional F-mode (flow mode), or a combined image thereof. The location of the gate can be indicated graphically in the image, for example by a circle, double line, or other graphic in the field of view.

[0064] In additional or alternative embodiments, information derived from the spectral envelope is displayed in an image. For example, the spectral envelope is used to detect the maximum positive and / or negative velocity. The maximum value(s) are displayed as alphanumeric text or graphically in any image (e.g., with reference to the patient's maximum value to a disease grading and / or population chart). As another example, the maximum values and / or other statistical information (e.g., standard deviation) of the energy and / or velocity are determined and displayed. In yet another example, the statistical information from the spectral envelope is used for a disease grading or another diagnostic or prognostic rating, for example a strandness criterion. The grading, diagnostic, or prognostic result is displayed.

[0065] Figure 5A system 500 for spectral Doppler imaging is shown. The system 500 uses dynamically determined boundaries (e.g., based on a mapping or plotting of the spectrum from substantially 0 to 2π) to track the spectral envelope. Aliasing can be avoided using the dynamically determined boundaries and / or the substantially 0 to 2π plotting.

[0066] The system 500 is a medical diagnostic ultrasound imaging system. Other imaging systems can be used, such as a workstation that loads samples from memory or other sources.

[0067] The system 500 includes a transmit beamformer 510, a transducer 520, a receive beamformer 530, a Doppler processor 540, a display 550, a signal processor 560, and a memory 570. In addition, different or fewer components can be provided, such as a system 500 without the front-end beamformers 510, 530 and / or the transducer 520 or a system 500 with a scan converter. The Doppler processor 540 and the signal processor 560 can be combined into one device that functions as both processors or additional processors for sequential or parallel processing can be used. User input can be provided for placing a gate and / or determining the size of the gate.

[0068] The system 500 implements a method of Figure 1 The beamformers 510, 530 and the transducer 520 are used to acquire samples. The Doppler processor 540 estimates a spectrum for a gate from the samples. The Doppler processor 540 and / or the signal processor 560 substantially plots from 0 to 2π and / or determines boundaries between frequency bands, tracks, and causes to be displayed a tracked spectral envelope or information derived therefrom. Other methods can be implemented. Doppler processing can be performed before or after CINE.

[0069] The transducer 520 is an array of multiple elements. The transmit beamformer 510 is shown as separate from the receive beamformer 530. Alternatively, the transmit and receive beamformers 510, 530 can be provided with some common components. Together or separately operating, the transmit and receive beamformers 510, 530 form a beam 524 of acoustic energy for sampling a range gate 522 and / or scanning a one-dimensional, two-dimensional, or three-dimensional region.

[0070] The transmit beamformer 510 is configured to transmit a sequence of transmit beams 524 of ultrasound energy. The acoustic energy is directed at the Doppler gate 522, but can be focused elsewhere (e.g., the Doppler gate is along a scan line but not at a focal point). The acoustic energy beams 524 are transmitted to the Doppler gate 522 or other locations.

[0071] A sequence of ongoing transmit beams 524 is generated with a PRF. The PRF determines the interval between adjacent transmits or transmit beams 524 in time. The PRF can be low enough to have no transmit-free periods needed for propagation time, interleaving with other imaging modes, and reverberation reduction. In one embodiment, the PRF is established based on a velocity scale, propagation time, interleaving, and reverberation reduction. In other embodiments, the PRF is set based on a velocity scale and a Nyquist criterion.

[0072] A receive beamformer 530 forms one or more receive beams 526 in response to each transmit beam 524. Although shown as parallel, the receive beams 526 can be collinear or angled with the transmit beams 524. Acoustic echoes are received by the transducers 520 in response to the transmitted acoustic energy. The echoes are converted to electrical signals by the transducers 520, and the receive beamformer 530 forms receive beams 526 from the electrical signals to generate samples representing one or more locations within the range gate 522.

[0073] Given the ongoing transmit beams 524 at a PRF or PRI (pulse repetition interval) for each location, the samples are also generated on an ongoing basis. The response over time is acquired. The response creates a collection or accumulation of samples representing a spectrum for the same time period.

[0074] The Doppler processor 540 is a spectral Doppler estimator. Other imaging detectors can be included, such as a B-mode detector. In one embodiment, the Doppler processor 540 is a digital signal processor or other device for applying a transform to the receive beam sample data. A sequence of transmit and receive events is performed over a period of time. A buffer (e.g., a corner turn memory) or memory 570 stores the receive beamforming data from each transmit and receive event. Wall filters (e.g., programmable filters to distinguish between tissue and fluid motion) can filter the samples before the transform is applied. Any number of transmit and receive events can be used to determine the spectrum, such as three or more. The Doppler processor 540 estimates a spectrum for the gate. By applying a discrete or fast Fourier transform or other transform to the ultrasound samples for the same gate, a spectrum representing the response from the gate is determined. A histogram or data representing the energy level at different frequencies over the period of time the samples were acquired is obtained. Velocity can be determined from the frequency, or the frequency is used without conversion to velocity.

[0075] By repeating the process, the Doppler processor 540 can obtain different spectra for a given gate at different times. Overlapping data can be used, such as with a moving window of selected ultrasound samples to compute each spectrum. Alternatively, each ultrasound sample is used for a single time period and a corresponding spectrum.

[0076] The Doppler processor 540 applies a transform to the frequency range. The range of the frequency or velocity scale limits the positive and negative velocities produced by the estimation. Any of a variety of velocity scales can be used, up to and including a velocity scale equal to the transmit PRF. The frequency spectrum is estimated using a given velocity scale. Similarly, the baseline or center of the velocity scale can be set.

[0077] The signal processor 560 can be part of the Doppler processor 540 or a separate processor. The signal processor 560 is a general purpose processor, a control processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array, a graphics processing unit, an analog circuit, a digital circuit, a combination thereof, or another now known or later developed means for processing. The signal processor 560 is configured by software, firmware, and / or hardware.

[0078] The signal processor 560 is configured to identify a line in the two frequency bands or one of the frequency bands and in the noise region between 0 and 2p. Other lines can be identified between other frequency band or velocity scale limits. The line can be determined between other plots, such as from -p to p.

[0079] The signal processor 560 identifies the line as a curve (i.e., the line changes over time within the velocity scale). The line is continuous, but can be discontinuous. A line fit can be performed. Filtering of the points can be used. The line changes over time on a frequency spectrum by frequency spectrum in the frequency spectrum, but can change with groups of frequency spectra. To avoid aliasing, the line is positioned in the noise between the frequency bands and / or between the frequency bands and the velocity scale limits (e.g., 0 or 2p). The line is between two frequency bands (between the frequency band edges) or between a frequency band and a velocity scale limit (e.g., 0 or 2p). Other locations can be used, such as an interval closer to one frequency band than the other.

[0080] The signal processor 560 can plot the frequency spectrum from 0 to 2p so that one of the two frequency bands is through 2p and the other of the two frequency bands is through 0. This can help identify the line separating the frequency bands. The plotted frequency spectrum is used to position the line in the noise region.

[0081] The signal processor 560 is configured to limit the search for a spectral envelope of the frequency spectrum by the line. The line defines or separates the different frequency bands. The dynamic line is used to define the search region, rather than tracking based on a baseline position of a given velocity scale. The search for each frequency band is limited using the identified line or boundary, instead of searching for one frequency band above the baseline and another frequency band below the baseline. Thus, the search or tracking does not include aliasing.

[0082] The display 18 is a CRT, monitor, LCD, plasma screen, projector, printer, or another now known or later developed display for displaying an image of the spectral envelope or information derived from the spectral envelope. For a spectral Doppler image, a range of velocities is provided as a function of time, with each velocity modulated as a function of energy.

[0083] The spectral envelope is displayed as a trace or pattern on the spectral Doppler image. Peak velocities or other information derived from the spectral envelope can be displayed on the spectral Doppler image (e.g., annotations) or as a separate image.

[0084] Other image configurations can be provided, including color spectral Doppler images. Color or flow mode images can be generated, for example, to show average velocity as a function of location in the region of interest in gray scale B mode.

[0085] The memory 570 stores ultrasound samples, estimated spectra, settings, image data, spectral plots, split lines or points forming dynamic lines, signal edges, traces, annotations, and / or other information. The memory 570 can store information from any processing stage or information used to generate a display.

[0086] In one embodiment, the memory 570 is a non-transitory computer readable storage medium having stored therein data representing instructions executable by the programmed Doppler processor 540 and / or signal processor 560 for Doppler imaging. The instructions for implementing techniques discussed herein are provided on a computer readable storage medium or memory, such as cache, buffer, RAM, removable media, hard disk, or other computer readable storage media. Computer readable storage media includes various types of volatile and non-volatile storage media. The functions, acts or tasks illustrated in the figures or described herein are implemented in response to the presence of one or more sets of instructions stored in or on a computer readable storage medium. The functions, acts or tasks are independent of the particular type of instructions set, storage media, processor or processing strategy and can be performed by software, hardware, integrated circuits, firmware, micro-code and the like, operating individually or in combination. Similarly, the processing strategy can include multiple processing or task operation, parallel or serial, and the like.

[0087] In one embodiment, the instructions are stored on removable media devices for reading by a local or remote system. In other embodiments, the instructions are stored in a remote location for transmission over a computer network or through telephone lines. In yet other embodiments, the instructions are stored within a given computer, CPU, GPU, or system.

[0088] While the application has been described above by reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of the application. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that all changes and modifications are intended to be included within the spirit and scope of the application as defined by the following claims (including all equivalents).

Claims

1. A method for spectral Doppler imaging, the method comprising: Using an ultrasound system to acquire samples representing responses from range gating; The spectrum of the range-gated spectrum over time is estimated from the sample using a Doppler estimator; Plot the spectrum from 0 to 2π over time; For each spectrum in the spectrum, a point in the noise region is identified between the first signal and the second signal, or between the first signal and 0, or between the first signal and 2π. The spectral envelope is tracked over time, and the tracking is limited by points in the spectrum; as well as Display the spectral envelope or information derived from the spectral envelope in the image. The identification includes searching in the spectrum for the frequency band containing the first signal and the frequency band containing the second signal, and (1) when the second signal is found in the search, positioning the point in the noise region between the first signal and the second signal, or (2) when the second signal is not found, positioning the point in the noise region between the first signal and 0 or 2π.

2. The method of claim 1, wherein acquiring includes transmitting at a pulse repetition frequency and acquiring one of the samples in response to each repetition.

3. The method of claim 1, wherein the estimation comprises applying a Fourier transform to the sample, each spectrum in the spectrum comprising energy as a function of frequency over a time period.

4. The method of claim 1, wherein the plotting comprises plotting each spectrum to begin at 0 and end at 2π, such that negative signals extend from 2π toward 0 and positive signals extend from 0 toward 2π.

5. The method of claim 1, wherein the identification includes positioning the point between the first signal and the second signal, or between the first signal and 0, or between the first signal and 2π.

6. The method of claim 1, wherein identification includes determining a noise level from at least some of the spectrum, and using the noise level to search for the first signal and the second signal.

7. The method of claim 1, wherein the identifier includes a time-varying boundary, the boundary being formed by points of the spectrum and varying relative to 0 and 2π.

8. The method of claim 1, wherein tracking includes locating the edge of the first signal having the noise region and locating the edge of the second signal having the noise region, the location of the edge being time-varying, and the edge including the spectral envelope.

9. The method of claim 1, wherein tracking includes using 0 or 2π for a first or second signal where the noise region extends to 0 or 2π and the noise region does not extend to 0 or 2π.

10. The method of claim 1, wherein tracking includes tracking the first signal as a positive signal between 0 and a point in the spectrum, and tracking the second signal as a negative signal between 2π and a point in the spectrum.

11. The method of claim 1, wherein the display includes displaying the spectral envelope, wherein it is positive above the baseline and negative below the baseline.

12. The method of claim 1, wherein display includes displaying the information, the information being based on the peak velocity of the spectral envelope.

13. A method for spectral Doppler imaging, the method comprising: Using an ultrasound system to acquire samples representing responses from range gating; The spectrum of the range-gated spectrum over time is estimated from the sample using a Doppler estimator; Plot the spectrum from 0 to 2π over time; Positive and / or negative spectral envelopes are detected over time based on the spectrum plotted from 0 to 2π; as well as The image displays the spectral envelope or information derived from the spectral envelope.

14. The method of claim 13, wherein the plotting comprises plotting each spectrum to begin at 0 and end at 2π, such that negative signals extend from 2π toward 0 and positive signals extend from 0 toward 2π.

15. The method of claim 13, wherein the detection comprises setting a boundary that varies spectrally from 0 to 2π, and searching for a positive spectral envelope on one side of the boundary and searching for a negative spectral envelope on the other side of the boundary.

16. The method of claim 15, wherein setting the boundary includes setting the boundary in the middle between signals in different frequency bands, wherein 0 or 2π is used to replace the spectrum of the frequency band in which there is no signal.

17. A system for spectral Doppler imaging, the system comprising: The beamformer is configured to sample the gating at pulse repetition intervals established in response to a velocity scale; A Doppler estimator is configured to generate multiple spectra from the gated samples; A signal processor is configured to identify lines in a noise region between 0 and 2π in two frequency bands or one of the frequency bands, the lines varying over time according to the spectrum, and limiting the search for the spectral envelope of the spectrum by the lines; as well as The display is configured to display the spectral envelope or information derived from the spectral envelope.

18. The system of claim 17, wherein the signal processor is configured to draw the spectrum from 0 to 2π such that one of the two frequency bands passes through 2π and the other of the two frequency bands passes through 0, and to identify the line drawn in a noise region from the spectrum.

19. The system of claim 17, wherein the signal processor is configured to position the line spectrally between the two frequency bands.

Images