Estimation of hemodynamic parameters

Abstract

A method for deriving one or more hemodynamic parameters based on blood velocity and arterial diameter metrics, each metric being cyclically or continuously sampled over a time period to obtain a data series (i.e., waveform) spanning a time window for each metric. Additionally, a radial blood velocity profile is calculated, which indicates blood velocity as a function of radial position on a plane perpendicular to the lumen of the vessel. This gives an indication of how blood velocity varies over vessel diameter for an individual patient. This information is in addition to the standard blood velocity and arterial diameter metrics, as input to a transfer function that maps the input to a hemodynamic parameter.

Description

Technical Field

[0001] The present invention relates to hemodynamic parameter estimation using ultrasound data of blood vessels. Background Art

[0002] Critically ill patients in the intensive care unit (ICU) and patients in the operating room (OR) require monitoring of various physiological parameters. These parameters include routine vital signs monitoring, such as heart rate (HR), respiratory rate (RR), arterial oxygen saturation (SpO2), body temperature, and usually invasive or non-invasive blood pressure (BP) measurement.

[0003] In addition, patients often undergo hemodynamic monitoring. Hemodynamic monitoring attempts to measure the forces responsible for the circulation of blood in the body. This is actually a performance measurement of the cardiovascular system. Typical core hemodynamic parameters include cardiac output (CO), stroke volume (SV), and stroke volume variability.

[0004] Hemodynamic monitoring is required for early detection, recognition, and management of life-threatening clinical conditions, such as sepsis and cardiogenic shock, and to assess the efficacy of pharmacological interventions, such as the administration of vasopressors.

[0005] There are many different methods for monitoring hemodynamic parameters. By way of example, three clinically common methods for monitoring cardiac output will now be described.

[0006] The first example is the thermodilution method. This method uses the Swan-Ganz pulmonary artery catheter (PAC). It is widely considered to be the most clinically accurate method for assessing core hemodynamics. It is considered the "gold standard" for hemodynamic monitoring in critically ill cardiac patients and is often the reference measurement against which new technologies are validated and compared. However, despite this, the Swan-Ganz method still exhibits significant imprecision, with variations in cardiac output measurements as high as 10-15%.

[0007] A second paradigm method for cardiac output monitoring that has been recently developed is the PiCCO (Pulse Contour Cardiac Output) system. This is a less invasive alternative to PAC because the location of the arterial line placement is less important, and the arterial line can be an axillary artery, a brachial artery, a femoral artery, or a radial artery. The PiCCO method also requires a central venous catheter. The PiCCO method is based on continuous cardiac output monitoring using a method of pulse contour analysis. This involves estimating CO and SV using arterial waveform information, together with intermittent transpulmonary thermodilution, for calibration purposes. For patients who already have a central arterial line, PiCCO only requires the insertion of an arterial catheter, making it less invasive than the Swan Ganz method. However, the accuracy of the PiCCO measurement is highly dependent on potential changes in systemic vascular resistance (SVR) and the time interval since the last calibration point (i.e., the time point since the last thermodilution). Sudden changes in SVR caused by the administration of vasoactive drugs make PiCCO measurements inaccurate. And over time, SVR may also change due to the gradual regulation of the autonomic nervous system. For this reason, the PiCCO system needs to be calibrated intermittently to obtain accurate measurements.

[0008] As a third example, there are minimally invasive and non-invasive methods for hemodynamic monitoring. These methods include minimally invasive System and non-invasive systems. Each of these methods relies on pulse contour analysis but omits calibration. As a result, clinicians have low confidence in them. Such systems perform better at tracking trends in hemodynamic parameters than obtaining absolute measurements.

[0009] The method uses an invasive peripheral arterial line, The method uses a finger cuff. Usually, The performance of this method is better than method.

[0010] The main limitations of more accurate methods of determining hemodynamic parameters include the highly invasive nature of the measurement modality (which also carries an increased risk to the patient) and the high level of skill required to insert an arterial line, especially in the case of a PAC. In addition, both the PAC and PiCCO methods require intermittent calibration to obtain accurate measurements. This is problematic when using thermodilution for calibration, which cannot be performed frequently. Some PAC methods include a heating element to heat a small volume of blood, thereby allowing continuous thermodilution to calibrate the system.

[0011] about and They are less accurate because they are not calibrated. This also means that they are not reliable for clinically relevant events such as infusions or medications, due to expected changes in SVR.

[0012] Therefore, classical hemodynamic monitoring methods are either highly invasive, inaccurate, or both.

[0013] It would be advantageous to find a non-invasive hemodynamic monitoring method with greater accuracy than currently available methods. Summary of the invention

[0014] In previous studies, the inventors of the present invention have found that a combination of ultrasound acquired parameters related to blood flow in an arterial branch of a subject can be used to estimate hemodynamic parameters. This is based on studies performed by the inventors that found that blood velocity waveforms (as a function of time) and arterial diameter waveforms (acquired and processed as independent signal sources as a function of time) correlate well with central hemodynamics. It has been found that by using features related to blood flow parameters measured at an arterial location (such as the carotid artery), a statistical or machine learning based model can be used to estimate central hemodynamics. A statistical or machine learning model can be used that embodies a functional or learned relationship between input and output hemodynamic parameters. This can be based on a supervised learning or training procedure. It can be based on a regression fitting procedure.

[0015] Previous methods used the following combination as input to the transfer function: a blood velocity-time waveform over at least one cardiac cycle and an arterial diameter-time waveform over at least one cardiac cycle. This method works well in calculating central hemodynamic parameters such as stroke volume and cardiac output.

[0016] After further investigation, the inventors have found that improvements can be made by additionally taking into account the variation of blood velocity over the vessel diameter. For this purpose, the inventors propose to additionally acquire color Doppler data and derive from these data a radial velocity profile over the artery at the measurement location and use one or more features of this profile as one or more additional inputs to the transfer function.

[0017] Therefore, one aspect of the present invention provides a computer-implemented method for deriving one or more hemodynamic parameters of a subject, the method comprising receiving 2D ultrasound data of a blood vessel (e.g., an artery) from the subject, wherein the 2D ultrasound data comprises B-mode data, pulsed wave Doppler data, and color Doppler data.

[0018] The method also includes using the received color Doppler data to derive a radial velocity profile for the blood vessel, the radial velocity profile corresponding to the blood velocity at the measurement location as a function of the radial position on the radial axis of the blood vessel. For example, the radial axis refers to the radial axis of a (cross-sectional) plane oriented relative to the normal of the longitudinal / axial axis of the artery at the measurement location. For example, the radial axis extends from the center point of such a plane to its periphery (i.e., the wall of the artery). The radial axis is sometimes referred to as the short axis of the artery, and the longitudinal axis is sometimes referred to as the long axis of the artery.

[0019] The method further comprises deriving a blood velocity-time waveform from the received 2D ultrasound data, the blood velocity-time waveform representing the blood velocity (as a function of time) of the blood vessel at the measurement location over a time window.

[0020] The method further comprises deriving an arterial diameter waveform from the received 2D ultrasound data, the arterial diameter waveform representing a diameter of the at least one blood vessel at the measurement location over the time window or a parameter proportional to the diameter as a function of time.

[0021] The method further comprises calculating a first set of parameters, comprising: calculating one or more predefined velocity profile parameters from the radial velocity profile, wherein calculating the one or more predefined velocity profile parameters comprises calculating at least a skewness of the velocity profile; calculating a predefined blood velocity parameter from the blood velocity waveform; and calculating a predefined arterial diameter parameter from the arterial diameter waveform. For example, the parameters may be characteristics of the respective waveforms. In some examples, the parameters may be statistical parameters. The skewness reflects an asymmetry of the waveform shape, like the radial position of a velocity maximum.

[0022] The method further comprises extracting a predefined transfer function configured to receive the first set of parameters as input and to generate the one or more hemodynamic parameters as output.

[0023] The method also includes processing the first set of parameters using the transfer function to derive values ​​for the one or more hemodynamic parameters.

[0024] The method may also include generating a data output reflective of the one or more hemodynamic parameters.

[0025] For example, "hemodynamic parameters" may mean central hemodynamic parameters, such as cardiac output, stroke volume, stroke volume variation. In some embodiments, the one or more hemodynamic parameters may include a fluid responsiveness metric.

[0026] Therefore, embodiments of the present invention are based on the use of a combination of ultrasound-acquired parameters related to blood flow in an arterial branch of a subject in order to estimate hemodynamic parameters. In addition, it is also proposed to measure a radial velocity profile for an artery, meaning the blood velocity as a function of position on the diameter of the arterial radius or lumen. The blood velocity is not constant over the arterial diameter. As will be explained in more detail later, it is usually assumed that the radial velocity profile is close to a parabola over the entire cardiac cycle. Due to pulsatility, blood flow may deviate slightly from the parabolic shape, and a fixed correction factor is usually used to estimate the flow. According to the literature, the population average of this factor is about 5.6-5.8. However, this is based on an ideal geometry for the artery. However, for any given patient, the vascular geometry may deviate from a straight line shape, which means that even at low pulsatility (low Womersley number), the flow will not be strictly parabolic. Accordingly, measuring and taking into account the radial velocity profile will improve accuracy when calculating hemodynamic parameters. Therefore, calculating the velocity profile in the artery will improve accuracy. The radial velocity profile may vary over time depending on the velocity-time waveform and may have a shape that may deviate significantly from a standard parabolic profile. This will be explained in more detail below.

[0027] In some embodiments, deriving the radial velocity profile may include: receiving color Doppler data that spans a scan plane oriented at an oblique angle relative to the longitudinal axis of the blood vessel; deriving a first blood velocity profile as a function of position on the scan plane; and, calculating a projection of the first blood velocity profile onto a plane normal to the longitudinal axis of the blood vessel and parallel to the radial axis of the blood vessel to thereby derive the radial velocity profile corresponding to blood velocity as a function of radial position on the radial axis. Doppler ultrasound data must be acquired at an angle to the direction of flow, otherwise a zero velocity measurement (no Doppler shift) is obtained. This results in velocity data as a function of radial position on the angled scan plane in a first instance. This is then projected onto a plane perpendicular to the direction of flow (i.e., normal to the longitudinal axis of the artery).

[0028] In some embodiments, calculating the one or more predefined velocity profile parameters further comprises calculating one or more of: (i) the radial position of the velocity maximum in the velocity profile, e.g. relative to the geometric center of the artery, (ii) the area under the curve of the velocity profile, or the area under one or more predefined segments (e.g. the first half and the second half) of the curve. The general concept is to derive parameters that directly or indirectly reflect the deviation of the shape of the velocity profile from a parabolic shape.

[0029] In some embodiments, the method comprises deriving a radial velocity profile for a series of time points over the time window to thereby derive a radial velocity profile as a function of time.

[0030] In some embodiments, calculating the one or more predefined velocity profile parameters comprises calculating the standard deviation of the radial velocity profile over the time window for each time point (each frame). This means taking the standard deviation of the velocity values ​​in the velocity profile at each time point. In some embodiments, these standard deviation values ​​may be averaged over all time points.

[0031] The goal here is to quantitatively capture the changes in the velocity profile over time by evaluating the standard deviation of the velocity profile for each frame. This method can capture blood flow turbulence, including that caused by arrhythmias.

[0032] In some embodiments, the aforementioned time window includes the systolic and diastolic phases of the subject's cardiac cycle. In some embodiments, calculating the one or more predefined velocity profile parameters includes calculating the difference between the velocity profile at a time point within the systolic phase and the velocity profile at a time point within the diastolic phase.

[0033] In some embodiments, at least one of the one or more predefined speed profile parameters represents a deviation of at least one characteristic of the speed profile from a parabolic speed function.

[0034] In some embodiments, calculating at least one of the one or more predefined velocity profile parameters comprises: extracting features from the velocity profile; and calculating deviations of the extracted features from corresponding features of the parabolic velocity function.

[0035] The various features presented above provide patient specific information and characterize the curvature and overall lumen shape. Including one or more of these parameters as input to the transfer function provides an improvement in the accuracy of the hemodynamic parameter estimate.

[0036] In some embodiments, the blood velocity parameter includes at least one of: an interval range of the velocity-time waveform over the time window; an average value of the blood velocity over the time window; an average value of the peak systolic velocity over the time window; an average value of the blood velocity over the time window normalized by the number of cardiac cycles spanned by the time window; an integral of the velocity-time waveform over the time window with respect to time normalized by the number of cardiac cycles spanned by the time window.

[0037] In some embodiments, the artery diameter parameter comprises at least one of: an average value of the artery diameter over the time window; and an average value of the cross-sectional area of ​​the at least one blood vessel over the time window.

[0038] In some embodiments, the one or more hemodynamic parameters may include one or more of: cardiac output, stroke volume, and stroke volume variability.

[0039] In some embodiments, the time window spans at least one cardiac cycle, and preferably spans multiple cardiac cycles.

[0040] In some embodiments, the transfer function is a machine learning model.

[0041] In some embodiments, the transfer function is a multi-parameter linear regression model.

[0042] The invention may also be embodied in software form. Another aspect of the invention is a computer program product comprising a computer readable code configured to, when executed by a suitable computer or processor, cause the computer or processor to perform a method according to any one of the embodiments outlined in this document.

[0043] Another aspect of the present invention is a computer readable medium having computer readable code embodied therein, the computer readable code being configured such that, when executed by a suitable computer or processor, the computer or processor is caused to perform a method according to any one of the embodiments outlined in this document.

[0044] The present invention may also be embodied in hardware. Therefore, another aspect of the present invention is a processing device comprising: an input / output; and one or more processors adapted to: receive 2D ultrasound data of a blood vessel from the subject, wherein the 2D ultrasound data comprises B-mode data, pulsed wave Doppler data and color Doppler data; use the received color Doppler data to derive a radial velocity profile of the blood vessel, the radial velocity profile corresponding to the blood velocity at a measurement position as a function of the radial position on the radial axis of the blood vessel; derive a blood velocity-time waveform from the received 2D ultrasound data, the blood velocity-time waveform representing the blood velocity of the blood vessel at the measurement position over a time window; derive an artery diameter waveform from the received 2D ultrasound data, the artery diameter waveform representing the diameter of the at least one blood vessel at the measurement position over the time window or a parameter proportional to the diameter.

[0045] The one or more processors are further adapted to calculate a first set of parameters, including: calculating one or more predefined velocity profile parameters based on the radial velocity profile; calculating predefined blood velocity parameters based on the blood velocity waveform; and calculating predefined arterial diameter parameters based on the arterial diameter waveform;

[0046] The one or more processors are further adapted to extract a predefined transfer function configured to receive the first set of parameters as input and to generate the one or more hemodynamic parameters as output; process the first set of parameters using the transfer function to derive values ​​of the one or more hemodynamic parameters; and preferably generate a data output reflecting the one or more hemodynamic parameters.

[0047] Another aspect of the present invention is a system, comprising: a processing device as described above or according to any embodiment detailed in this document; and an ultrasound scanning device, comprising at least one transducer unit and a processing unit, wherein the at least one transducer unit is used to collect ultrasound echo signal data of the at least one blood vessel of the object, and the processing unit is used to process the echo data to derive B-mode data, pulse wave Doppler data and color Doppler data. The input / output of the processing device can be operatively coupled to the output of the ultrasound scanning device for receiving the B-mode data, pulse wave Doppler data and color Doppler data.

[0048] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] For a better understanding of the invention, and to show more clearly how it may be put into effect, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0050] Figure 1 Example blood velocity-time waveforms and diameter-time waveforms are illustrated;

[0051] Figure 2 An example radial velocity profile across a cross section of an artery is shown;

[0052] Figure 3 The relationship between the radial velocity profile and the Dean number is illustrated;

[0053] Figure 4 A comparison of an example radial velocity profile for a vessel with a moderate pulse compared to an ideal parabolic velocity profile is illustrated;

[0054] Figure 5An example variation of the velocity profile correction factor (y-axis) as a function of time (x-axis; seconds) for a single patient is shown;

[0055] Figure 6 Example vascular morphology is shown;

[0056] Fig. 7A The radial velocity profile along a relatively straight vessel is illustrated;

[0057] Figure 7B The radial velocity profile of a blood vessel along a non-straight line is illustrated;

[0058] FIG8 shows a set of example radial velocity profiles at different points along a tortuous vessel;

[0059] Fig. 9 The steps of an example method according to one or more embodiments of the present invention are outlined;

[0060] Fig.10 Outlining components of example processing devices and systems according to one or more embodiments of the present invention;

[0061] Fig.11 The geometry of an example blood vessel is schematically illustrated;

[0062] Fig.12 A radial cross section of an example blood vessel is schematically illustrated;

[0063] Fig.13 The projection of the acquired Doppler scan plane onto a plane normal to the direction of flow through the vessel (longitudinal axis of the vessel) is illustrated;

[0064] Fig.14 illustrates a comparison of an example blood velocity profile to an ideal parabolic profile to extract one or more velocity profile parameters; and

[0065] Fig.15 The architecture of an example ultrasound system is outlined. DETAILED DESCRIPTION

[0066] The present invention will be described with reference to the accompanying drawings.

[0067] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the apparatus, system and method, are intended for illustrative purposes only and are not intended to limit the scope of the invention. These and other features, aspects and advantages of the apparatus, system and method of the present invention will become better understood from the following description, the appended claims and the accompanying drawings. It should be understood that the drawings are only schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the drawings to indicate the same or similar parts.

[0068] The present invention provides a method for deriving one or more hemodynamic parameters based on blood velocity and arterial diameter metrics, each metric being cyclically or continuously sampled over a time period to obtain a data series (i.e., waveform) spanning a time window for each metric. Additionally, a radial blood velocity profile is calculated, which indicates blood velocity as a function of radial position on a plane perpendicular to the vessel lumen. This gives an indication of how blood velocity varies over vessel diameter for an individual patient. This information is in addition to the standard blood velocity and arterial diameter metrics as input to a transfer function that maps the input to a hemodynamic parameter.

[0069] At least one goal is to estimate hemodynamic parameters such as cardiac output and stroke volume based on ultrasound measurements performed on a vessel (e.g., an artery such as the carotid artery). It is also possible to use other vital signs and patient demographics as inputs into the transfer function. From 2D ultrasound data (e.g., pulsed wave Doppler and B-mode), it is possible to calculate the velocity (velocity-time waveform) and the arterial diameter (diameter-time waveform) at a single measurement time point. By way of illustration, Figure 1 A typical velocity-time waveform 12 and diameter-time waveform 14 are shown. Velocity is shown on the left y-axis (in cm / second) and diameter is shown on the right y-axis (in mm). Time is plotted on the x-axis (in seconds).

[0070] It is generally assumed that the radial / diameter velocity profile of the common carotid artery (CCA) is close to a parabola over the entire cardiac cycle at locations distal to the bifurcation (approximately 1.5 cm and below). Due to pulsatility, blood flow may deviate slightly from the parabolic shape, and a fixed correction factor is usually used to estimate flow. According to the literature, population average values ​​of approximately 5.6-5.8 are cited, and the population average value may depend on the specific method by which flow is extracted from the measurements.

[0071] However, in large patient populations, the CCA geometry may deviate from a straight line shape. Therefore, the flow will not be strictly parabolic even at low pulsatility (low Womersley number). In engineering, this effect is called Dean flow, and the deviation of the flow from a parabola is characterized by the Dean number De=Re·(d / 2Rc), where Re is the Reynolds number, d is the arterial diameter, and Rc is the arterial curvature. Therefore, in addition to the velocity-time waveform at a single measurement point, the radial velocity profile in the artery provides valuable additional information. This velocity profile may vary with time depending on the velocity-time waveform and may have a shape that deviates significantly from a simple parabolic profile.

[0072] Figure 2Some of the same radial velocity profile shapes are schematically illustrated, in the form of rough contour plots. The radial direction r is indicated. Example (a) shows a standard parabolic profile. Example (b) illustrates a non-parabolic profile, where the velocity maximum is displaced from the center point of the artery, so that the entire velocity profile is tilted. Example (c) shows a non-parabolic profile, where the velocity maximum is also displaced from the center, and where the gradient lines are distorted into an irregular, asymmetric gradient profile.

[0073] Figure 3 An indicative illustration of how the velocity profile deviates as a function of the Dean number is provided. Each graph shows a respective plurality of radial velocity profiles. In each graph, the y-axis represents velocity and the x-axis represents radial position across the artery, scaled to the vessel radius, with zero at the centerline. Thus, each individual line or waveform in each graph is a single respective radial velocity profile. The arrows 22 in each graph show the direction of increasing Dean number. Thus, it can be seen how the radial profile shape changes as the Dean number increases.

[0074] In the literature, a population average correction factor is often used to derive the beat-to-beat blood flow. However, volunteer studies performed by the present inventors have shown that the flow correction factor can vary significantly from patient to patient, based on the mean blood flow, with a typical range of 0.55 to 0.67.

[0075] By means of diagrams, Figure 4 Example radial velocity profiles are shown. The y-axis shows velocity (m / s) and the x-axis shows radial position (normalized fraction of the total radius). Line 32 shows a standard parabolic radial velocity profile. Line 34 shows an example radial velocity profile for an artery with a Womersley number of 3 corresponding to flow. Line 36 illustrates the average velocity over the entire cardiac cycle for the parabolic 32 profile. Line 38 illustrates the average velocity over the entire cardiac cycle for an artery with a Womersley number of 3.

[0076] Table 1 below shows examples of different correction factors (taking into account the correction for the deviation from the parabolic radial velocity profile) for a group of five different patients and for flow measured at six time points for each patient. The time difference between acquisitions varied from 5 to 10 minutes.

[0077] Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 Time 1 0.67 0.59 0.59 0.62 0.59 Time 2 0.58 0.62 0.61 0.54 0.61 Time 3 0.65 0.59 0.57 0.60 0.57 Time 4 0.68 0.58 0.60 0.55 0.60 Time 5 0.67 0.59 0.57 0.57 0.57 Time 6 0.61 0.60 0.57 0.57 0.57 average 0.64 0.595 0.585 0.575 0.585

[0078] Table 1 - Velocity Profile Correction Factors for Patients 1-5 at Time Points 1-6

[0079] It is not difficult to see that the correction factor varies significantly between different patients and also for the same patient at different time points. Therefore, it can be expected that using a single average correction factor will lead to some inaccuracy.

[0080] In addition, patient-specific instantaneous correction factors based on velocity waveforms may vary between 0.45 and 0.7, e.g. Figure 5 As shown in the example, Figure 5 The velocity profile correction factor (y-axis) is shown as a function of time (x-axis; seconds) for a single example patient. The instantaneous velocity profile correction factor indicates the degree to which the measured radial velocity profile deviates from a standard parabolic velocity profile.

[0081] The above considerations suggest that the transfer function used to map arterial flow parameters to hemodynamic parameters will be made more accurate if the patient's radial velocity profile is taken into account.

[0082] In addition to the above, not only pulsatility but also vessel morphology contributes significantly to the deviation of an individual's arterial flow from the standard parabolic radial velocity profile. For illustration, Figure 6 Different morphologies of the common carotid arteries of a group of different patients are shown. As mentioned above, in known methods, correction of the flow or velocity profile due to pulsatility is usually considered based on the Womersley number. This can be done using different methods, such as averaging correction factors, segment-by-segment correction (for systole and diastole), and finally using instantaneous, inverse Womersley reconstruction.

[0083] However, deviations caused by local or global arterial shape are not typically taken into account, but may cause the velocity profile to deviate significantly from a parabolic shape. For example, a non-straight common carotid artery (CCA) velocity profile will vary at different locations in a single patient, and therefore, the accuracy of any calculations that depend on the velocity profile may vary at different measurement locations if this variable characteristic is not taken into account.

[0084] Vascular stenosis and calcification may also cause changes in the radial velocity profile.

[0085] Additionally, vessel tortuosity (ie, the degree of nonlinearity of the vessel) can result in changes in the radial velocity profile as a function of position along the vessel. Tortuosity tends to increase with patient age.

[0086] For example, Fig. 7A A relatively straight (low tortuosity) example blood vessel 42 is illustrated. Three sample radial velocity profiles 44a, 44b, 44c along a portion of the vessel are shown. All are substantially similar to each other, and all are substantially approximated to parabolic velocity profiles.

[0087] By comparison, Figure 7BAn example vessel 42 is shown which changes direction at a bend point and therefore has a high degree of tortuosity. Two example radial velocity profiles 44a, 44b upstream of the bend point show very similar profile shapes and generally match the parabolic velocity profile. A set 44c of radial velocity profiles downstream of the bend point is also shown. It can be seen that these profiles are different from each other and also have distortions compared to the standard parabolic velocity profile. Thus, it can be appreciated that the velocity profiles vary greatly depending on the upstream and downstream morphology.

[0088] To illustrate this more clearly, FIG. 8 shows a series of graphs indicating radial velocity profiles 52 at different locations along a tortuous vessel compared to a parabolic velocity profile 54. The y-axis shows velocity (m / s), while the x-axis shows radial position on a cross section of the artery measured from the center point of the artery. FIG. 8a shows a velocity profile 52a for an artery segment having a straight inflow path and a straight outflow path. The profile is very similar to the parabolic velocity profile 54. FIG. 8b shows a velocity profile 52b for an artery segment having an inflow and a straight outflow at a small horizontal angle. FIG. 8c shows a velocity profile 52c for an artery segment having a straight inflow path and an outflow path at a small horizontal angle. FIG. 8d shows a velocity profile 52d for an inflow at a small horizontal angle and an outflow at a small horizontal angle. It can be seen that the radial velocity profiles of the velocity profiles with non-straight inflow and / or outflow deviate from the parabolic velocity profile 54 to a greater extent.

[0089] Based on the above observations, the inventors associated with the present application proposed to acquire Doppler data on a radial cross section of a vessel during a cardiac cycle, and obtain blood velocity measurements at different radial positions on the radial cross section of the vessel to thereby derive a radial blood velocity profile. It is proposed to use features extracted from the profile as additional inputs to a transfer function configured to map individual vessel flow / geometry inputs to hemodynamic parameter outputs.

[0090] Fig. 9 The steps of an example computer-implemented method 60 according to one or more implementations are summarized in block diagram form. These steps will be briefly described before further explanation in the form of an example embodiment.

[0091] The method 60 is used to derive one or more hemodynamic parameters of a subject.

[0092] The method includes receiving 62 2D ultrasound data ("U / S data") from a blood vessel (eg, an artery) of a subject, wherein the 2D ultrasound data includes B-mode data, pulsed wave Doppler data, and color Doppler data.

[0093] The method further comprises deriving 64 a radial velocity profile for the blood vessel using the received color Doppler data, the radial velocity profile corresponding to the blood velocity at the measurement location as a function of the radial position on the radial axis of the blood vessel. The radial axis means, for example, the radial axis of a (cross-sectional) plane oriented normally relative to the longitudinal / axial axis of the artery at the measurement location. For example, the radial axis extends from the center point of such a plane to its periphery (i.e., to the wall of the artery). The radial axis is sometimes referred to as the short axis of the artery, while the longitudinal axis is sometimes referred to as the long axis of the artery.

[0094] The method further comprises deriving 66 a blood velocity-time waveform from the received 2D ultrasound data, the blood velocity-time waveform representing the blood velocity (as a function of time) of the blood vessel at the measurement location over a time window. Preferably, the time window spans at least one complete cardiac cycle, such as a plurality of cardiac cycles. This may be derived from pulse wave Doppler (PWD) data. The blood velocity waveform may be represented by a data series of blood velocity measurement samples for the measurement location at regular time intervals over the time window. The pulse wave Doppler data is acquired for a single measurement location, the measurement location being defined according to a gating setting of an acquisition device.

[0095] The method further comprises deriving 68 an arterial diameter waveform from the received 2D ultrasound data, the arterial diameter waveform representing the diameter of at least one blood vessel at the measurement location over a time window or a parameter proportional to the diameter (as a function of time). This may be derived from processing of the B-mode data. Extracting the diameter measurement may be based on an automatic segmentation algorithm or on another image processing algorithm. Those skilled in the art will recognize appropriate technical means for extracting diameter measurements from 2D ultrasound data.

[0096] The method further comprises calculating a first set of parameters, including: calculating 70 one or more predefined velocity profile parameters from the radial velocity profile; calculating 72 predefined blood velocity parameters from the blood velocity waveform; and calculating 74 predefined arterial diameter parameters from the arterial diameter waveform. For example, the parameters may be features of the respective waveforms. In some examples, the parameters may be statistical parameters.

[0097] The method further comprises extracting a predefined transfer function configured to receive the first set of parameters as input and to generate one or more hemodynamic parameters as output. The method further comprises processing 76 the first set of parameters using the transfer function to derive values ​​for the one or more hemodynamic parameters.

[0098] The method may also include generating a data output 78 reflecting one or more hemodynamic parameters. In some examples, the data output may be transmitted to a user interface, such as a patient monitor system. It may be displayed using a display device. It may be transmitted to a data storage unit for subsequent retrieval. It may be transmitted to a remote computer or remote data storage via a network or Internet link.

[0099] As mentioned above, the method 60 may also be embodied in hardware form, for example in the form of a processing unit or device configured to perform the method according to any example or embodiment described in this document or according to any claim of the present application.

[0100] To further aid understanding, Fig.10 A schematic representation of an example processing device 82 configured to perform a method according to one or more embodiments of the present invention is presented. The processing device is shown in the context of a system 80 that includes the processing device. The processing device alone represents one aspect of the present invention. The system 80 is another aspect of the present invention. The provided system does not necessarily include all of the illustrated hardware components; it may include only a subset of them.

[0101] The processing device comprises one or more processors 86, which are configured to perform the method according to the method outlined above, or according to any of the embodiments described in this document or in any of the claims of this application. In the illustrated example, the processing unit also includes a communication interface or input / output 84.

[0102] exist Fig.10 In the illustrated example, the system 80 further includes an ultrasound scanning device 92, which includes at least one transducer unit 94 and a processing unit 96, wherein the at least one transducer unit 94 is used to collect ultrasound echo signal data of at least one blood vessel of the subject, and the processing unit 96 is used to process the echo data to derive B-mode data, pulse wave Doppler data, and color Doppler data. Fig.15 Methods for performing such processing are described in more detail. In some embodiments, the transducer unit may include an ultrasound probe. In some embodiments, the transducer unit includes a wearable ultrasound patch, the wearable ultrasound patch including an ultrasound transducer array integrated in the patch, the patch can be attached to the surface of the patient's body for monitoring purposes. For example, the patch can be applied over a vessel of interest (e.g., the common carotid artery).

[0103] Of course, the system does not necessarily include all of the hardware components mentioned above. The system according to the present invention may be provided without including these components or only including one or more of these components.

[0104] The system 80 may also include a memory 88 for storing computer program code (e.g., computer executable code or instructions) configured to cause one or more processors 86 of the processing device 82 to perform the method as outlined above, or a method according to any embodiment described in the present disclosure, or a method according to any claim.

[0105] As mentioned above, the present invention can also be embodied in software form. Therefore, another aspect of the present invention is a computer program product comprising code, wherein the code is configured to, when running on a processor, cause the processor to perform a method according to any example or embodiment of the present invention described in this disclosure, or a method according to any claim of this patent application.

[0106] Although pulsed wave Doppler (PWD) data can produce high temporal resolution blood velocity information, it only acquires measurements for a single sample volume in the artery, which is typically small compared to the diameter of the artery.

[0107] In contrast, color Doppler can be used to obtain velocity measurements at sample points across the entire lumen of the artery (over the entire radial axis, such as over the entire radius or the entire diameter), thereby providing a velocity profile. It is noted that two-dimensional (2D) color Doppler is contemplated in the present disclosure. By extracting (one or more) specific features from the velocity profile, the transfer function can be supplemented with information that directly or indirectly indicates deviations from a parabolic flow profile. This is unique spatial information about flow through the artery that is not available from PWD and B-mode data, and this information can improve transfer function performance. This provides significant improvements, especially in the case of non-straight arteries. Non-straight arterial paths can be both natural anatomical features and caused by plaque formation such as atherosclerosis.

[0108] The method is also of particular benefit to patients suffering from cardiac arrhythmias. During arrhythmias, rather than a coordinated contraction of the heart, a quivering motion is often observed, with insufficient filling of the heart, which may result in irregular blood flow. As a result, the blood flow behavior appears more pulsatile (albeit with a lower peak flow velocity).

[0109] Fig.11 A blood vessel 42 is schematically shown with a blood flow along the longitudinal axis z of the vessel. The direction r of the radial axis of the vessel is shown. The blood flow has a velocity v along the longitudinal direction z of the vessel. Fig.12 Shown along Fig.11 The radial cross section is taken along the line X indicated in FIG. The radius r and the circumferential Axial direction. In this schematic illustration, the vessel 42 lumen has an overall radius R and a diameter 2R. The radial dimension can be measured from a center point 154 of the vessel lumen, such that the lumen diameter extends from a radial position -R on one side to +R on the other side.

[0110] As outlined above, embodiments of the present invention relate to obtaining a radial velocity profile for a blood vessel, the radial velocity profile corresponding to the blood velocity as a function of the radial position on the radial axis r of the blood vessel at a measurement position. In this case, the measurement position means a position along the longitudinal axis of the blood vessel, the longitudinal axis of the blood vessel meaning the axis in the same direction as the blood flow. For example, the measurement position may be taken as Fig.11 The position x indicated in is obtained to obtain a radial velocity profile indicating the blood velocity in the longitudinal direction z as a function of the radial position along the radial axis r of the blood vessel.

[0111] As mentioned above, in at least some embodiments it is proposed to obtain a radial velocity profile from color Doppler data.Color Doppler data may be acquired at position X using an ultrasound scanning device.

[0112] Preferably, Doppler ultrasound data is acquired at an angle to the direction of blood flow, otherwise a zero velocity measurement (no Doppler shift) is obtained. In a first example, this produces velocity data as a function of radial position on the angled scan plane. This velocity data can then be projected onto a plane perpendicular to the direction of flow (i.e., normal to the longitudinal axis of the artery).

[0113] Thus, in some embodiments, deriving the radial velocity profile may include receiving color Doppler data across a scan plane oriented at an oblique angle relative to the longitudinal axis of the vessel. Fig.13 As illustrated in , a scan plane A may be acquired along a direction Y. The method may further include deriving a first blood velocity profile as a function of position on the scan plane. The method may then further calculate a projection of the first blood velocity profile onto a plane B, normal to the longitudinal axis of the vessel and parallel to the radial axis of the vessel, to thereby derive a radial velocity profile corresponding to blood velocity as a function of radial position on the radial axis.

[0114] like Fig.13 As illustrated in , for example, it is proposed to first measure the radial velocity profile along the minor axis of an ellipse generated by slicing the artery with an ultrasound scan plane at an angle relative to the longitudinal axis.

[0115] In order to project the velocity profile measured on the elliptical plane A onto a plane B normal to the long axis of the vessel, the method comprises estimating the angle ψ of the scan plane relative to the plane normal to the longitudinal axis. A (rA )→V B (r B )=V B (cosψr A ) resamples the velocity measurement results and projects the velocity profile from the inclined plane A onto the vertical plane B, where V A is the radial position r on the radial axis of the angled plane A B The velocity profile measured as a function of V B is the radial position r in the plane normal to the longitudinal axis of the vessel B In other words, if the initial velocity profile V A A series of radial positions r on the radial axis of the angled plane A The multiple blood velocity measurements at the locations are projected onto the vertical plane, including each velocity measurement V A (r A ) is resampled to the normal plane at V B (r B =cosψr A ) at the new measurement sample.

[0116] In operation, when ultrasound data is acquired, color Doppler measurements are acquired on a 2D grid of focal spots / points that collectively define a scan plane. In operation, the ultrasound scanning device can be operated so that the scan plane intersects the target vessel at an oblique angle relative to the longitudinal axis of the vessel. The angle of the scan plane can be adjusted manually by the operator manipulating the orientation of the ultrasound transducer unit, or it can be adjusted electronically using beam steering. Then, once measurement data has been acquired for each grid of measurement points, an ellipse is fit to the color Doppler flow image and the main geometric parameters of the ellipse (axis, orientation) are identified. The angle ψ of the scan plane relative to the plane normal to the longitudinal axis can be calculated based on the length of the major and minor axes of the ellipse using the following equation:

[0117]

[0118] Each velocity measurement point on an axis radial to the scan plane (eg, on the major axis of the ellipse) may then be mapped to a velocity measurement for the normal plane by multiplying the coordinate along the major axis of the ellipse by cosψ, as discussed above.

[0119] As discussed above, it is also proposed to obtain pulse wave Doppler (PWD) ultrasound data in addition to color Doppler data. Pulse wave Doppler data can be used to derive a blood velocity-time waveform, which represents the blood velocity at the measurement position of the blood vessel over a time window. Pulse wave Doppler data is usually collected only at a single relatively small measurement position, such as at the center of the blood vessel. Compared with color Doppler (multiple beams), pulse wave Doppler data is collected using a smaller aperture (single beam). PWD enables sampling in a (small) volume at a higher repetition rate (higher sampling frequency), which allows excellent statistics and higher SNR in the velocity measurement results to become possible. Therefore, compared to color Doppler with higher spatial resolution and SNR, PWD exhibits a trade-off at the expense of a smaller spatial sampling area.

[0120] As discussed above, it is also proposed to obtain B-mode ultrasound data. The B-mode ultrasound data can be used to derive an arterial diameter waveform, which represents the diameter of the blood vessel at the measurement position over a time window or a parameter proportional to the diameter. The B-mode data includes a sequence of scan lines acquired on a scan plane. The B-mode scan plane can be the same plane as the color Doppler scan plane. In some embodiments, the B-mode scan plane can be perpendicular to the color Doppler scan plane. Therefore, the B-mode ultrasound data according to the present disclosure is 2D B-mode ultrasound data.

[0121] In some embodiments, three-mode ultrasound acquisition is used to jointly acquire color Doppler data, pulsed wave Doppler data, and B-mode data in an interleaved manner over a measurement time window. In three modes, color Doppler frames (at a lower frame rate) are synchronized with PWD frames (which have a higher frame rate). PWD frames interrogate flow velocity and direction along a single scan line at a depth (defined by sample volume or gating). Color flow Doppler interrogates multiple sample volumes simultaneously along an array of scan lines (each pixel represents a sample volume).

[0122] In some embodiments, the derived radial velocity profile may be modified or adjusted based on pulse wave Doppler (PWD) data to improve accuracy. As mentioned above, PWD velocity measurements have a higher SNR (signal to noise ratio). In some embodiments, a normalization constant may be derived based on at least one PWD derived velocity measurement for adjusting each velocity measurement point.

[0123] For example, adjusting a radial velocity profile derived from a color Doppler data set may include the following steps. First, identify the PWD velocity measurement that is closest in time to the color Doppler data set. The PWD velocity measurement will correspond to a particular gating location, such as the center of a vessel. Then, derive a normalization constant C_Normalized = (PWD_Gating Location) / (Color Doppler velocity at the PWD gating location). For example, if the PWD gating location is the center of a vessel, the Color Doppler measurement corresponding to that same location is used as the denominator.

[0124] Each velocity value in the radial velocity profile is then multiplied by a normalization constant C_Normalized to derive an adjusted radial velocity profile.

[0125] As discussed above, the method involves: calculating one or more predefined velocity profile parameters from a radial velocity profile; calculating a predefined blood velocity parameter from a blood velocity waveform; and calculating a predefined arterial diameter parameter from an arterial diameter waveform. There are different options for the calculated blood velocity parameters, arterial diameter parameters, and velocity profile parameters, which serve as inputs to the transfer function.

[0126] Regarding the blood velocity parameter, artery diameter parameter, and velocity profile parameter, the meaning of these labels is that these parameters are calculated or extracted from the blood velocity waveform, artery diameter waveform, and radial velocity profile, respectively. In other words, they are blood velocity-derived parameters and artery diameter-derived parameters and radial velocity profile-derived parameters. For example, they can be renamed as the first parameter, the second parameter, and the third parameter.

[0127] Different options for the radial velocity profile parameters will now be discussed.The purpose of deriving the radial velocity profile parameters is to provide a quantitative representation of the profile in a form that can serve as an input to a transfer function.

[0128] Preferably, the extracted features each represent the degree to which the velocity profile deviates from a parabolic velocity profile.

[0129] Calculating one or more predefined velocity profile parameters includes calculating the skewness of the velocity profile. For example, the skewness can be calculated as the absolute difference between the areas under the "left" and "right" parts of the velocity profile. The left part refers to the part of the velocity profile that spans radial position values ​​less than the center position value, while the right part refers to the part of the velocity profile that spans position values ​​greater than the center position value. In other words, it can be calculated as the difference between the respective areas under the velocity profile on both sides of the center point of position.

[0130] In some embodiments, calculating one or more predefined velocity profile parameters includes calculating the radial position of the velocity maximum in the velocity profile. For example, this may include calculating the radial position Δr of the velocity maximum relative to the centerline of the vessel. For example, Δr is equal to the radial distance of the velocity maximum from the radial center of the vessel. In some embodiments, this may be further normalized by the radius R of the vessel, i.e., the one or more predefined velocity profile parameters may include a value of Δr / R.

[0131] In some embodiments, calculating one or more predefined speed profile parameters comprises calculating an area under a curve of the speed profile.

[0132] In some embodiments, calculating one or more predefined velocity profile parameters includes calculating a ratio of an area under a left side of the curve to an area under a right side of the curve, wherein the left side refers to a portion of the curve to the left of a radial center point of the blood vessel and the right side of the curve refers to a portion of the curve to the right of the radial center point of the curve.

[0133] In some embodiments, the method comprises deriving a radial velocity profile for a series of time points over a time window to thereby derive a radial velocity profile as a function of time.

[0134] In some embodiments, calculating one or more predefined velocity profile parameters comprises calculating a standard deviation of the radial velocity profile over the time window. In some embodiments, calculating one or more predefined velocity profile parameters comprises calculating a standard deviation of the radial velocity profile over the time window for each time point (i.e., each frame), and then calculating an average of the standard deviation values ​​over all time points.

[0135] The concept in this case is to quantitatively capture the changes in the velocity profile over time by evaluating the standard deviation of the velocity profile for each frame. The method can thereby capture turbulent blood flow, including that caused by arrhythmias.

[0136] In some embodiments, the aforementioned time window includes systole and diastole.In some embodiments, calculating one or more predefined velocity profile parameters includes calculating the difference between the velocity profile at a time point in systole and the velocity profile at a time point in diastole.

[0137] Systole and diastole can be identified in the velocity-time waveform by identifying unique waveform features. For example, a trough to peak double trough corresponds to systole, while a double trough to a subsequent trough corresponds to diastole. In some embodiments, an average value of the velocity profile during systole can be calculated, and an average value of the velocity profile during diastole can be calculated. These average values ​​can be included in the calculated velocity profile parameters.

[0138] In some embodiments, as a more general principle, at least one of the one or more predefined velocity profile parameters represents a deviation of at least one characteristic of the velocity profile from a parabolic velocity function.

[0139] In some embodiments, calculating at least one of the one or more predefined velocity profile parameters comprises extracting features from the velocity profile and calculating deviations of the extracted features from corresponding features of the parabolic velocity function.

[0140] By means of diagrams, Fig.14 An ideal parabolic radial velocity profile 54 is shown, compared to an example measured radial velocity profile 52 over 61 frames in a simulation. Line 162 indicates the radial center of the artery (radius R), while line 164 indicates the radial location of the velocity maximum of the measured radial velocity profile 52. In some embodiments, calculating one or more predefined velocity profile parameters may include calculating a radial distance Δr between the radial location 164 of the velocity maximum of the measured velocity profile and the radial center 162 of the vessel (which corresponds to the radial location of the velocity maximum of the parabolic velocity profile).

[0141] This represents only one example. More generally, it is proposed to derive features of the measured radial velocity profile that, independently and in conjunction with each other, characterize the deviation of the velocity profile from a parabola. For example, these features may include any of the features mentioned above. One or more of these features may be provided as input to a transfer function. In some embodiments, the difference relative to the corresponding feature of the parabolic velocity profile may be explicitly calculated and used as input to the transfer function.

[0142] Reference has been made herein to a parabolic velocity profile. By this, the parabolic velocity profile refers to a profile indicating the blood velocity v as a function of the position r on the cross section of the vessel, which has the form of a parabolic function, e.g.

[0143] v(x)=b0+b 1r +b 2r 2

[0144] Wherein b2 is generally negative because the blood velocity is generally maximum near the center of the blood vessel, and where b0 may be equal to the maximum blood velocity value on the blood vessel.

[0145] With respect to speed parameters and diameter parameters, different options will now be discussed with reference to Table 2 below, which lists various possible options for these parameters. Parameters 1 and 3-6 are examples of speed parameters, while parameters 2 and 7 are examples of diameter parameters.

[0146]

[0147] Table 2

[0148] With respect to the blood velocity parameter, by way of a non-limiting set of examples, the parameter may be any one or more of the following:

[0149] The range of the velocity waveform in the time window (IDRVelWave in Table 2);

[0150] The average value of blood velocity over the time window (meanVelWave in Table 2);

[0151] The mean of the peak systolic velocity over the time window (PSV in Table 2 );

[0152] The mean value of blood velocity over the time window, normalized by the number of cardiac cycles spanned by the time window (meanVelNormPerBeat in Table 2);

[0153] The velocity waveform is integrated with respect to time over a time window, normalized by the number of cardiac cycles spanned by the time window (VTINormPerBeat in Table 2).

[0154] In some embodiments, a combination of two or more of these parameters may be used as input to the transfer function.

[0155] Regarding the artery diameter parameter, according to a set of non-limiting examples, this parameter may be one or more of the following:

[0156] The mean value of arterial diameter over the time window (Dia in Table 2 );

[0157] The cross-sectional area of ​​at least one vessel is averaged over the time window (CSArea in Table 2).

[0158] In some embodiments, the transfer function may be adapted to receive as additional input one or more demographic characteristics of the subject, such as age, gender and / or body mass index (BMI).

[0159] Table 2 provides a summary of a non-limiting set of example parameters, some or all of which may be selected as inputs to be obtained and provided to a transfer function in order to derive one or more hemodynamic parameters.

[0160] With respect to the derived one or more hemodynamic parameters, these parameters may include, by way of non-limiting example, one or more of the following: cardiac output, stroke volume, and stroke volume variability.

[0161] In some embodiments, the method may further include obtaining at least one predefined additional parameter for the object over the time window, and wherein the transfer function is further configured to receive the predefined additional parameter for the object as an additional input (in addition to the first set of parameters 70, 72, 74 discussed above).

[0162] By way of non-limiting example, Table 3 below lists a set of example predefined additional parameters that may be used in the present context.

[0163] Before further explanation below, these parameters are briefly described in the table.

[0164]

[0165] Table 3

[0166] In particular, with respect to at least one additional parameter, in some examples, this may include the subject's heart rate (HR in Table 3). In some embodiments, the method may include receiving Doppler ultrasound data of at least one vessel, and processing the Doppler ultrasound data to derive a measure of the subject's heart rate.

[0167] In some examples, the at least one additional parameter may include a parameter derived from processing a stroke volume waveform. For example, the method may include processing a velocity waveform and an arterial diameter waveform to derive an arterial stroke volume waveform. The arterial stroke volume waveform may be calculated by processing the velocity waveform to derive an area under the waveform over a single cardiac cycle (e.g., by calculating a velocity-time integral of the waveform over a single complete cardiac cycle); and processing the diameter waveform to derive a cross-sectional area waveform over the same cardiac cycle. The stroke volume waveform may be derived based on the area under the velocity waveform and based on the cross-sectional area waveform.

[0168] For example, at least one further parameter may include the area under the arterial stroke volume waveform over the time window, normalized by the number of cardiac cycles spanned by the time window (ArtSVNormperBeat in Table 3 above).

[0169] Additionally or alternatively, at least one further parameter may comprise an average of the arterial stroke volume waveform over a time window normalized by the number of cardiac cycles spanned by the time window (meanArtSVWaveNormPerBeat and meanArtSVWave in Table 3 above, respectively).

[0170] In some examples, at least one additional parameter may include a mean arterial blood flow value (volume flow per unit time) over a time window. This parameter corresponds to meanArtFlow in Table 3 above. This parameter may be calculated by: processing an arterial diameter waveform to derive an arterial cross-sectional area waveform for a time window; deriving an arterial blood flow waveform for a time window based on calculating the product of a velocity waveform and an arterial cross-sectional area waveform; and, processing the arterial blood flow waveform to derive a mean arterial flow value over a time window.

[0171] In some examples, the at least one additional physiological parameter may include one or more vital signs, such as heart rate, respiratory rate, and / or blood pressure. In some examples, these parameters may be obtained based on sensor signals received from one or more physiological parameter sensors.

[0172] Thus, a process for obtaining a first set of parameters has been described above, the first set of parameters comprising: one or more predefined velocity profile parameters; at least one predefined blood velocity parameter; and, at least one predefined artery diameter parameter. As mentioned, optionally, at least one further parameter may be derived and used as an additional input to the transfer function, but this is optional.

[0173] As discussed above, the method also includes: extracting a predefined transfer function, which is configured to receive at least the first set of parameters as input and generate one or more hemodynamic parameters as output; and processing the first set of parameters using the transfer function to derive values ​​of one or more hemodynamic parameters.

[0174] According to some embodiments, processing these parameters to derive one or more hemodynamic parameters may be performed using a machine learning model comprising one or more machine learning algorithms that have been trained to map a predefined set of input parameters to a set of one or more hemodynamic parameters.

[0175] A machine learning algorithm is any self-training algorithm that processes input data to generate or predict output data. Here, the input data includes at least a blood velocity parameter, an arterial diameter parameter, and one or more velocity profile parameters (i.e., a first set of parameters), and the output data includes one or more hemodynamic parameters. Optionally, the input data may additionally include one or more of the additional parameters discussed above in connection with Table 3.

[0176] Suitable machine learning algorithms used in the present invention will be apparent to those skilled in the art. Examples of suitable machine learning algorithms include linear regression algorithms, decision tree algorithms, and artificial neural networks. Other machine learning algorithms such as logistic regression, support vector machines, or naive Bayes models are suitable alternatives.

[0177] In the examples provided below, a machine learning algorithm in the form of a multi-parameter linear regression model is used to demonstrate the principles of the inventive concept. However, it is to be understood that in each example, the machine learning model can be replaced by a different type of machine learning model without affecting the advantageous technical effects.

[0178] In particular, for a multi-parameter linear regression model, the model is established by first constructing a model or algorithm incorporating each of the desired input parameters as an (independent) parameter of the model with a corresponding coefficient or weight; and secondly training the constructed model based on a training data set to thereby fit the model coefficients or weights to provide the best fit between the input parameters of the training data set and the corresponding output parameters of the training data set. The desired input parameters form the independent variables of the model, while the target hemodynamic parameters are the dependent variables of the model. The model represents the estimated relevant hemodynamic parameters as a linear sum of a constant term (intercept) and each dependent variable multiplied by the respective weight or coefficient.

[0179] The training data set will include training input data items and corresponding training output data items. In this case, the training input data items correspond to preselected example values ​​of blood velocity parameters, artery diameter parameters, and one or more velocity profile parameters. Optionally, the training input data items may additionally include one or more of the additional parameters discussed above in connection with Table 3. The training output data items correspond to predetermined one or more hemodynamic parameters.

[0180] The initialized machine learning algorithm is applied to each input data item to generate a predicted output data item. The error between the predicted output data item and the corresponding training output data item is used to modify the machine learning algorithm. This process can be repeated until the error converges and the predicted output data item is sufficiently similar to the training output data item (e.g., ±1%). This is generally referred to as a supervised learning technique.

[0181] For a multi-parameter regression model, the training process is the process of fitting the model weights / coefficients to the training data set. Once the training or fitting process is completed, the model can be deployed using the weights or coefficients obtained in the training or fitting process to map the input parameters (independent variables) to the output hemodynamic parameters.

[0182] After training, the performance or accuracy of the model can be evaluated by running the resulting machine learning model on a test dataset and evaluating the error between the output predictions generated by the model and the actual ground truth. For example, for a linear regression model, performance measures can include the goodness of fit (R2) of the linear regression, the root mean square error (RMSE) obtained from a correlation scatter plot, and the reproducibility coefficient (rpc) obtained from a Buran-Altman plot.

[0183] According to one or more embodiments, the method may further include deriving a measure of blood flow based on the use of the radial velocity profile, the derivation of which has been discussed above.

[0184] In some embodiments, the method may include generating a real-time instantaneous flow signal representing a measure of instantaneous blood flow at the measurement location (as a function of time). In the context of the present disclosure, flow refers to the volume of fluid flowing through a measurement point per unit time.

[0185] Determining a measure of blood flow from Doppler data generally involves calculating a blood velocity value for the measurement location and multiplying that value by a value for the cross-sectional area of ​​the blood vessel at the measurement location. This operation then provides an estimated measure of the volume of blood flowing through the measurement location per unit time.

[0186] If only PWD data is acquired, only a single velocity value is obtained at the measurement location and this value can then be multiplied by the cross-sectional area, or an estimated average velocity value is calculated from the PWD measurements based on an assumed parabolic velocity profile and this average multiplied by the cross-sectional area.

[0187] In the context of the method of the present invention, the derived radial velocity profile provides a way to obtain a more accurate estimate of the average velocity through the vessel (meaning the average velocity over a radial cross section of the vessel). In particular, according to one or more embodiments, it is proposed to: calculate and store the radial velocity profile in the manner discussed above; receive real-time PWD data for a gated position in the vessel (e.g., the center of the vessel); extract the stored radial velocity profile; scale the velocity profile based on instantaneous PWD velocity measurements; use the radial velocity profile to calculate the average velocity over the vessel diameter (i.e., the velocity values ​​of the velocity profile are averaged); and calculate the average instantaneous flow by multiplying the calculated average velocity by a measure of the cross-sectional area of ​​the vessel at the measurement location. The cross-sectional area can be calculated based on the measured vessel diameter, which is obtained based on the B-mode data discussed above.

[0188] In some embodiments, multiple radial velocity profiles may be stored, such as one radial velocity profile for each of multiple time points across at least one cardiac cycle. In some embodiments, the particular velocity profile extracted may be determined based on the instantaneous phase of the cardiac cycle. Since the radial velocity profile may vary at different time points over the cardiac cycle, this improves the accuracy of the instantaneous blood flow metric.

[0189] In some embodiments, the stored radial velocity profile may be updated periodically, for example, it may be updated for each new color Doppler frame acquired. If multiple profiles corresponding to different phase points of the cardiac cycle are stored, each profile may be updated for each new color Doppler frame acquired for the corresponding phase point in the cardiac cycle.

[0190] In some embodiments, a transfer function may be defined that allows for mapping between PWD velocity measurements acquired at predefined gating positions and the estimated radial velocity profile. A different transfer function may exist for each of a plurality of cardiac cycle phase points.

[0191] The above-mentioned scaling of the velocity profile based on the instantaneous PWD measurement can be understood as follows. The purpose is to maintain the shape or form of the profile, but to adjust the absolute value so that the PWD value measured at the PWD gating position of the artery matches the velocity value indicated by the scaled profile at the same gating position.

[0192] In some embodiments, an average flow value for each cardiac cycle may be calculated. This may be accomplished by performing the above process for each color Doppler frame acquired over the cardiac cycle (i.e., for each of a series of time points across the cardiac cycle). The average flow value for each cardiac cycle may be calculated by averaging the instantaneous flow values ​​calculated in the above manner for each scaled color Doppler frame over each cardiac cycle. The radial velocity profile may vary with changing flow. Therefore, by averaging over multiple time points across the cardiac cycle, a more accurate flow value may be calculated.

[0193] To achieve this feature, the frame rate of the color Doppler data acquisition is preferably set to a value greater than 10 frames per heartbeat. Additionally or alternatively, in some embodiments, at least one radial velocity profile may be calculated for the systolic period and at least one radial velocity profile may be calculated for the diastolic period, and wherein at least these two velocity profiles may be used in the manner described above to calculate the average flow over the cardiac cycle.

[0194] As described above, the method according to an embodiment of the present invention can be advantageously applied to the continuous or continuous monitoring of the hemodynamics of a subject. The method may include: generating a data output indicating one or more estimated hemodynamic parameters generated by the model; and transmitting the data output to a patient monitor system, and wherein the patient monitor system includes a display device and is suitable for displaying a visual representation of the derived one or more hemodynamic parameters on the display device. Additionally or alternatively, the patient monitor system may store or cache the derived hemodynamic parameters (locally or remotely). It can export the parameters to a remote system, such as an institutional website or server.

[0195] According to another aspect of the present invention, a method for providing a machine learning model for deriving one or more hemodynamic parameters is provided. The method includes generating an initial machine learning model, the initial machine learning model is suitable for receiving a predefined parameter set as input, and processing the received parameters to generate an estimate of one or more hemodynamic parameters. The predefined input parameter set includes: parameters calculated from a blood velocity-time waveform, the blood velocity-time waveform representing the blood velocity at a measurement position of at least one blood vessel on a time window; parameters extracted from an arterial diameter-time waveform, the arterial diameter-time waveform representing the diameter of the at least one blood vessel at a measurement position on a time window; and, one or more parameters derived from a radial velocity profile, the radial velocity profile representing the blood velocity as a function of the position on the radial axis of the blood vessel at the measurement position.

[0196] The method also includes providing a training data set, the training data set comprising a plurality of training input data items and a corresponding plurality of training output data items, the training input data items each comprising a value for each of a predefined set of input parameters, and the training output data items each comprising a corresponding value for one or more hemodynamic parameters.

[0197] The method also includes applying a machine learning algorithm to the training input data items and adjusting internal parameters of the machine learning model to minimize the error between the output of the generated model and the training output data items.

[0198] As described above, certain embodiments may include an ultrasound scanning device 92, and / or a device for processing ultrasound echo data to derive additional data.

[0199] By further explaining in more detail, reference will now be made to Fig.15 To describe the general operation of an exemplary ultrasound system.

[0200] The system includes an array transducer probe 104 having a transducer array 106 for transmitting ultrasound waves and receiving echo information. The transducer array 106 may include CMUT transducers; piezoelectric transducers formed of materials such as PZT or PVDF; or any other suitable transducer technology. In this example, the transducer array 106 is a two-dimensional array of transducers 108 that can scan a 2D plane or a three-dimensional volume of a region of interest. In another example, the transducer array may be a 1D array.

[0201] The transducer array 106 is coupled to a microbeamformer 112, which controls the reception of signals by the transducer elements. The microbeamformer can at least partially beamform signals received by subarrays (generally referred to as "groups" or "chips") of transducers, as described in U.S. Patents 5,997,479 (Savord et al.), 6,013,032 (Savord), and 6,623,432 (Powers et al.).

[0202] It should be noted that the microbeamformer is generally entirely optional. In addition, the system includes a transmit / receive switch 116 (T / R switch 116) to which the microbeamformer 112 can be coupled and which switches the array between transmit and receive modes and protects the main beamformer 120 from high energy transmit signals in the event that the microbeamformer is not used and the transducer array is operated directly by the main system beamformer. Transmission of ultrasound beams from the transducer array 106 is directed by a transducer controller 118, which is coupled to the microbeamformer via the T / R switch 116 and the main transmit beamformer (not shown), and which can receive input from user operation of a user interface or control panel 138. The controller 118 may include transmit circuitry arranged to drive the transducer elements of the transducer array 106 during transmit mode (either directly or via the microbeamformer).

[0203] In this exemplary system, according to an embodiment of the present invention, the functions of the control panel can be facilitated by an ultrasonic controller.

[0204] In a typical line-by-line imaging sequence, the beamforming system within the probe may operate as follows. During transmit, a beamformer (which may be a microbeamformer or a main system beamformer, depending on the implementation) activates a transducer array or a subaperture of a transducer array. A subaperture may be a one-dimensional transducer row or a two-dimensional transducer sheet within a larger array. In transmit mode, the focusing and steering of the ultrasound beam generated by the array or a subaperture of the array is controlled as described below.

[0205] In the case of receiving backscattered echo signals from the object, the received signals are subjected to receive beamforming (described below) in order to align the received signals, and in the case of using a subaperture, the subaperture is then shifted, for example, by one transducer element. The shifted subaperture is then activated, and the process is repeated until all transducer elements of the transducer array have been activated.

[0206] For each row (or sub-aperture), the total received signal used to form the associated row in the final ultrasound image will be the sum of the voltage signals measured by the transducer elements in a given sub-aperture during the receive period. After the subsequent beamforming process, the resulting row signal is typically referred to as radio frequency (RF) data. Each row signal (RF data set) generated by each sub-aperture then undergoes additional processing to generate the row in the final ultrasound image. The change in the amplitude of the row signal over time will contribute to the change in the brightness of the ultrasound image with depth, where high amplitude peaks will correspond to bright pixels (or sets of pixels) in the final image. Peaks that appear near the beginning of the row signal will represent echoes from shallow structures, while peaks that appear gradually later in the row signal will represent echoes from structures at gradually increasing depths within the object.

[0207] One of the functions controlled by the transducer controller 118 is the direction of beam steering and focusing. The beam can be steered to be directly in front of (orthogonal to) the transducer array, or at different angles to achieve a wider field of view. The steering and focusing of the transmit beam can be controlled as a function of the transducer element activation time.

[0208] In general ultrasound data acquisition, two methods can be distinguished: plane wave imaging and "beam steering" imaging. The two methods differ in the presence of beam forming in transmit mode ("beam steering" imaging) and / or receive mode (plane wave imaging and "beam steering" imaging).

[0209] First, let's look at the focusing function. By activating all the transducer elements at the same time, the transducer array generates a plane wave that diverges as it travels through the object. In this case, the beam of ultrasound waves remains unfocused. By introducing a position-dependent delay to the activation of the transducer, it is possible to cause the wavefront of the beam to converge at a desired point (called the focal zone). The focal zone is defined as the point where the lateral beam width is less than half the transmit beam width. In this way, the lateral resolution of the final ultrasound image is improved.

[0210] For example, if the delay causes the transducer elements to be activated sequentially, starting with the outermost elements of the transducer elements and ending at the (one or more) center elements, a focal zone will be formed at a given distance from the probe, aligned with the (one or more) center elements. The distance of the focal zone from the probe will change according to the delay between each subsequent round of transducer element activation. After the beam passes through the focal zone, it will begin to diverge, forming a far-field imaging region. It should be noted that for a focal zone located close to the transducer array, the ultrasound beam will diverge rapidly in the far field, resulting in beam width artifacts in the final image. Typically, the near field between the transducer array and the focal zone shows almost no details due to the large ultrasound beam overlap. Therefore, changing the position of the focal zone may result in significant changes in the quality of the final image.

[0211] It should be noted that in transmit mode, only one focus may be defined unless the ultrasound image is divided into multiple focal regions (each of which may have a different transmit focus).

[0212] Furthermore, in the case of receiving echo signals from within the subject, it is possible to perform the inverse of the above process to perform receive focusing. In other words, the incoming signal can be received by the transducer elements and electronically delayed before being transmitted to the system for signal processing. The simplest example of this operation is called delayed-sum beam imaging. It is possible to dynamically adjust the receive focus of the transducer array based on time.

[0213] Turning now to the beam steering function, by properly applying delays to the transducer elements, it is possible to impart a desired angle to the ultrasound beam as it leaves the transducer array. For example, by activating the transducers on a first side of the transducer array, followed by sequential activation of the remaining transducers, ending up on the opposite side of the array, the wavefront of the beam will be tilted toward the second side. The magnitude of the steering angle relative to the normal to the transducer array depends on the magnitude of the delays between subsequent transducer element activations.

[0214] Additionally, it is possible to focus the steered beam, wherein the total delay applied to each transducer element is the sum of both the focusing and steering delays. In this case, the transducer array is called a phased array.

[0215] In the case of CMUT transducers (which require a DC bias for their activation), the transducer controller 118 may be coupled to control a DC bias control 145 for the transducer array. The DC bias control 145 sets the DC bias voltage(s) applied to the CMUT transducer elements.

[0216] For each transducer element in the transducer array, an analog ultrasound signal (commonly referred to as channel data) enters the system through a receive channel. In the receive channel, a partial beamforming signal is generated by the microbeamformer 112 according to the channel data, and then the partial beamforming signal is transmitted to the main receive beamformer 120, where the partial beamforming signals from the individual slices of the transducer are combined into a complete beamforming signal (referred to as radio frequency (RF) data). The beamforming performed at each stage can be performed as described above, or additional functions can be included. For example, the main beamformer 120 can have 128 channels, each channel receiving a partial beamforming signal from a slice having tens or hundreds of transducer elements. In this way, the signals received by thousands of transducers of the transducer array can effectively contribute to a single beamforming signal.

[0217] The beamformed received signal is coupled to a signal processor 122. The signal processor 122 can process the received echo signals in various ways, such as: bandpass filtering, decimation, I and Q component separation; and harmonic signal separation, which is used to separate linear and nonlinear signals, thereby enabling identification of nonlinear (higher harmonics of the fundamental frequency) echo signals returned from tissue and microbubbles. The signal processor can also perform additional signal enhancements, such as speckle reduction, signal compounding, and noise elimination. The bandpass filter in the signal processor can be a tracking filter, and as the echo signal is received from an increasing depth, the passband of the tracking filter slides from a higher frequency band to a lower frequency band, thereby filtering out higher frequency noise from greater depths (which noise generally does not contain anatomical information).

[0218] The beamformers for transmission and for reception are implemented in different hardware and may have different functions. Of course, the receiver beamformer is designed to take into account the characteristics of the transmit beamformer. Fig.15 In the illustration, only the receiver beamformer is shown for simplicity. In a complete system, there will also be a transmit chain with a transmit micro-beamformer and a main transmit beamformer.

[0219] The function of the microbeamformer 112 is to provide preliminary combining of the signals to reduce the number of analog signal paths. This is typically performed in the analog domain.

[0220] Final beamforming is done in the main beamformer 120, typically after digitization.

[0221] The transmit and receive channels use the same transducer array 106 with fixed frequency bands. However, the bandwidth occupied by the transmit pulses may change depending on the transmit beamforming used. The receive channel can capture the entire transducer bandwidth (which is the classical approach) or by using bandpass processing, which can extract only the bandwidth containing the desired information (e.g., harmonics of the main harmonic).

[0222] The RF signal can then be coupled to a B-mode (i.e., brightness mode, or 2D imaging mode) processor 126 and a Doppler processor 128. The B-mode processor 126 performs amplitude detection on the received ultrasound signal for imaging structures in the body (e.g., organ tissue and blood vessels). In the case of line-by-line imaging, each line (beam) is represented by an associated RF signal, whose amplitude is used to generate a brightness value to be assigned to a pixel in the B-mode image. The exact position of the pixel within the image is determined by the position measured along the associated amplitude of the RF signal and the number of lines (beams) of the RF signal. B-mode images of such structures can be formed in harmonic or fundamental image modes or a combination of the two, as described in U.S. Pat. No. 6,283,919 (Roundhill et al.) and U.S. Pat. No. 6,458,083 (Jago et al.). The Doppler processor 128 processes the temporally different signals generated by tissue movement and blood flow for detecting moving matter, such as the flow of blood cells in the image field. The Doppler processor 128 typically includes a wall filter whose parameters are set to pass or filter out echoes returning from selected types of material in the body.

[0223] The structure and motion signals generated by the B-mode and Doppler processors are coupled to the scan converter 132 and the multi-planar reconstructor 144. The scan converter 132 arranges the echo signals in a spatial relationship, according to which the echo signals are received in a desired image format. In other words, the scan converter acts to convert the RF data from the cylindrical coordinate system to a Cartesian coordinate system suitable for displaying an ultrasound image on the display 140. In the case of B-mode imaging, the brightness of the pixel at a given coordinate is proportional to the amplitude of the RF signal received from that position. For example, the scan converter can arrange the echo signals into a two-dimensional (2D) sector format or a cone three-dimensional (3D) image. The scan converter can superimpose colors on the B-mode structural image, the colors corresponding to the motion at a point in the image field, where the Doppler estimated velocity is used to generate a given color. The combined B-mode structural image and the color Doppler image depict the motion of tissue and blood flow within the structural image field. A multiplanar reconstructor converts echoes received from points within a common plane in a volumetric region of the body into an ultrasound image of that plane, as described in U.S. Pat. No. 6,443,896 (Detmer). A volume renderer 142 converts echo signals of a 3D data set into a projected 3D image viewed from a given reference point, as described in U.S. Pat. No. 6,530,885 (Entrekin et al.).

[0224] The 2D or 3D images are coupled from the scan converter 132, the multiplanar reconstructor 144, and the volume renderer 142 to the image processor 130 for further enhancement, buffering, or temporary storage for optional display on the display 140. The image processor may be adapted to remove certain imaging artifacts from the final ultrasound image, such as: acoustic shadowing, e.g. caused by strong attenuation or refraction; rear enhancement, e.g. caused by weak attenuation; echo artifacts, e.g. in close proximity to highly reflective tissue interfaces; etc. In addition, the image processor may be adapted to process certain speckle reduction functions to improve the contrast of the final ultrasound image.

[0225] In addition to being used for imaging, the blood flow values ​​produced by the Doppler processor 128 and the tissue structure information produced by the B-mode processor 126 are coupled to a quantification processor 134. The quantification processor produces metrics for various flow conditions, such as volumetric rate of blood flow in addition to structural measurements such as organ size and fetal age. The quantification processor may receive input from a user control panel, such as the point in the anatomical structure of the image at which the measurement is to be made.

[0226] The output data from the quantization processor is coupled to the graphics processor 136 for the reproduction of measurement graphs and values ​​together with the image on the display 140, and for the audio output from the display 140. The graphics processor 136 can also generate graphic overlays for display with the ultrasound image. These graphic overlays can contain standard identification information, such as the patient's name, the date and time of the image, the imaging parameters, etc. For these purposes, the graphics processor receives input from the user interface, such as the patient's name. The user interface is also coupled to the transmit controller 118 to control the generation of ultrasound signals from the transducer array 106, and therefore the images produced by the transducer array and the ultrasound system. The transmit control function of the controller 118 is only one of the functions performed. The controller 118 also considers the mode of operation (given by the user) and the corresponding required transmitter configuration and the bandpass configuration in the receiver analog-to-digital converter. The controller 118 can be a state machine with fixed states.

[0227] A user interface is also coupled to the multi-planar reconstructor 144 for selecting and controlling the planes of a plurality of multi-planar reconstructed (MPR) images, which may be used to perform quantitative measurements in the image field of the MPR images.

[0228] The ultrasound system may be operatively coupled to the processing device 82. The processing device may receive Doppler ultrasound data and spatial ultrasound data (e.g., B-mode) from the ultrasound system. For example, in some examples, the ultrasound system may be used to implement Fig.10 The ultrasonic scanning device 92 of the system 80 is shown in FIG.

[0229] The embodiments of the present invention described above adopt processing equipment. Processing equipment can generally include a single processor or multiple processors. It can be located in a single containing device, structure or unit, or it can be distributed between multiple different devices, structures or units. Therefore, the reference to the processing equipment suitable for or configured to perform a specific step or task can correspond to the step or task performed by any one or more of the multiple processing components alone or in combination. Those skilled in the art will understand how such distributed processing equipment can be implemented. The processing equipment includes a communication module or an input / output unit for receiving data and outputting data to other components.

[0230] The one or more processors of the processor device can be implemented in a variety of ways using software and / or hardware to perform the required functions. The processor usually adopts one or more microprocessors, which can be programmed to perform the required functions using software (e.g., microcode). The processor can be implemented as a combination of dedicated hardware that performs some functions and one or more programmed microprocessors and associated circuits that perform other functions.

[0231] Examples of circuits that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs).

[0232] In various embodiments, the processor may be associated with one or more (non-transitory) storage media, such as volatile or non-volatile computer memory, such as RAM, PROM, EPROM, and EEPROM. The storage medium may be encoded with one or more programs that, when run on one or more processors and / or controllers, perform the desired functions. Each storage medium may be fixed within the processor or controller, may be portable, or may be provided on demand (e.g., via the cloud) so that one or more programs stored thereon may be loaded into the processor or controller.

[0233] Variations to the disclosed embodiments will be understood and implemented by those skilled in the art in practicing the claimed invention based on a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

[0234] A single processor or other unit may fulfill the functions of several items recited in the claims.

[0235] The mere fact that certain measures are recited in mutually different claims does not indicate that a combination of these measures cannot be used to advantage.

[0236] The computer program may be stored / distributed on a suitable (non-transitory) computer-readable medium, such as an optical storage medium or solid-state medium provided together with or as part of other hardware, but may alternatively be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0237] If the claims or specification use the term "suitable for," it should be noted that the term "suitable for" is intended to be equivalent to the term "configured to."

[0238] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A computer-implemented method (60) for deriving one or more hemodynamic parameters of a subject, the method comprising: receiving 2D ultrasound data of a blood vessel from the subject, wherein the 2D ultrasound data includes B-mode data, pulsed wave Doppler data, and color Doppler data; using the received color Doppler data to derive a radial velocity profile for the vessel, the radial velocity profile corresponding to blood velocity at a measurement location as a function of radial position on a radial axis of the vessel; deriving a blood velocity-time waveform from the received 2D ultrasound data, the blood velocity-time waveform representing the blood velocity of the blood vessel at the measurement position over a time window; deriving an arterial diameter waveform according to the received 2D ultrasound data, wherein the arterial diameter waveform represents the diameter of the blood vessel at the measurement position in the time window or a parameter proportional to the diameter; The first set of calculation parameters includes: calculating one or more predefined velocity profile parameters based on the radial velocity profile, wherein calculating the one or more predefined velocity profile parameters comprises calculating at least a skewness of the velocity profile; calculating a predefined blood velocity parameter based on the blood velocity-time waveform; and calculating a predefined arterial diameter parameter according to the arterial diameter waveform; extracting a predefined transfer function configured to receive as input the first set of parameters and to generate as output the one or more hemodynamic parameters; and The first set of parameters is processed using the transfer function to derive values ​​for the one or more hemodynamic parameters.

2. The method of claim 1, wherein: Calculating the one or more predefined velocity profile parameters further comprises calculating one or more of: (i) radial positions of velocity maxima in the velocity profile, (ii) an area under a curve of the velocity profile.

3. The method according to claim 1 or 2, wherein: The method comprises deriving a radial velocity profile for a series of time points over the time window to thereby derive a radial velocity profile as a function of time.

4. The method of claim 3, wherein: Calculating the one or more predefined velocity profile parameters comprises calculating a standard deviation of the radial velocity profile over the time window.

5. The method of claim 3, wherein: The time window includes a systolic phase and a diastolic phase of the subject's cardiac cycle, and wherein calculating the one or more predefined velocity profile parameters includes calculating a difference between the velocity profile at a time point within the systolic phase and the velocity profile at a time point within the diastolic phase.

6. The method of claim 4, wherein: The time window includes a systolic phase and a diastolic phase of the subject's cardiac cycle, and wherein calculating the one or more predefined velocity profile parameters includes calculating a difference between the velocity profile at a time point within the systolic phase and the velocity profile at a time point within the diastolic phase.

7. The method according to claim 1 or 2, wherein: At least one of the one or more predefined speed profile parameters represents a deviation of at least one characteristic of the speed profile from a parabolic speed function.

8. The method of claim 7, wherein: Calculating at least one of the one or more predefined velocity profile parameters comprises: extracting features from the velocity profile; and calculating deviations of the extracted features from corresponding features of the parabolic velocity function.

9. The method according to claim 1 or 2, wherein: The blood velocity parameter includes at least one of the following: The interval range of the speed-time waveform in the time window; The average value of blood velocity over the time window; The peak contraction velocity is averaged over the time window; an average of the blood velocity over the time window normalized by the number of cardiac cycles spanned by the time window; The velocity-time waveform is integrated with respect to time over the time window, normalized by the number of cardiac cycles spanned by the time window.

10. The method according to claim 1 or 2, wherein: The artery diameter parameter includes at least one of the following: The average value of the artery diameter over the time window; The cross-sectional area of ​​the blood vessel is averaged over the time window.

11. The method according to claim 1 or 2, wherein: Deriving the radial velocity profile includes: receiving color Doppler data spanning a scan plane oriented at an oblique angle relative to a longitudinal axis of the blood vessel; deriving a first blood velocity profile as a function of position on the scanning plane; and A projection of the first blood velocity profile onto a plane normal to the longitudinal axis of the blood vessel and parallel to a radial axis of the blood vessel is calculated to derive therefrom the radial velocity profile corresponding to blood velocity as a function of radial position on the radial axis.

12. The method according to claim 1 or 2, wherein: The one or more hemodynamic parameters include one or more of: cardiac output, stroke volume, and stroke volume variability.

13. The method of claim 1 or 2, wherein: The transfer function is a machine learning model; and / or The transfer function is a linear regression model.

14. A computer program product comprising instructions which, when executed by a suitable computer or processor, cause the computer or processor to perform the method according to any one of claims 1 to 13.

15. A processing device (82), comprising: Input / output (84); as well as One or more processors (86) adapted to perform the steps of any of claims 1 to 13.

16. The processing device (82) of claim 15, wherein: The one or more processors (86) are further adapted to generate a data output indicative of the one or more hemodynamic parameters.

17. A system comprising: A processing device (82) as claimed in claim 15 or 16; as well as An ultrasound scanning device (92), comprising at least one transducer unit (94) for collecting ultrasound echo signal data of the blood vessel of the object and a processing unit (96) for processing the ultrasound echo signal data to derive B-mode data, pulse wave Doppler data and color Doppler data; Wherein, the input / output (84) of the processing device is operatively coupled to the output of the ultrasound scanning device for receiving the B-mode data, pulse wave Doppler data and color Doppler data.

Images