DE102019125062A1 - Method for the automated determination of hemodynamic parameters in a blood vessel on invasively recorded pulse waves - Google Patents

Abstract

The invention relates to a method for the automated determination of hemodynamic parameters in a blood vessel on invasively recorded pulse waves by analyzing these pulse waves and / or pulse waves selected from the entirety, but at least one pulse wave. The invasive measurement and / or determination of the pulse wave parameters (PWPs), in particular the pulse wave speed (PWV), is carried out with the method steps: inserting a pressure measuring catheter with a pressure measuring sensor up to a designated measuring point within the blood vessel, recording and storing a temporal pressure curve and recording a pulse wave signal at the measuring point, identification and classification of the individual pulse waves from the pulse wave signal, analysis of the pulse waves with regard to different properties, e.g. Times, amplitudes, shapes and / or relevant points, determination of the PWV on the one hand by means of a wave decomposition method by determining and converting a time offset with a given reference distance between a pulse wave basic signal component and the pulse wave echo component of the blood vessel system in the pulse wave signal, determination of the PWV on the other hand by means of a wave analysis without dependence to the reference distance using a given pressure / flow model and a pressure / volume model of the blood vessel system.

Description

The invention relates to a method for the automated determination of hemodynamic parameters in a blood vessel on invasively recorded pulse waves by analyzing these pulse waves and / or pulse waves selected from the entirety, but at least one pulse wave according to claim 1.

Cardiovascular events strongly define human mortality worldwide. In order to identify the associated risk at an early stage and prevent such an event, numerous hemodynamic parameters are collected and evaluated. For example, the elasticity or stiffness of the arteries is an independent predictor of cardiovascular events (heart attack, stroke, death) and a promising way of being able to differentiate between risk persons, especially those with intermediate cardiovascular risk. The pulse wave velocity ( PWV ) describes this vascular stiffness and is a proven, officially recognized parameter that is anchored in the ESH guidelines, for example.

In addition to the pulse wave velocity ( PWV ) a variety of other known hemodynamic parameters play an important role in the assessment of a person's arterial condition, for example the augmentation index, the augmentation pressure, the cardiac output or the PWV -Variability. These parameters make it possible, individually and / or from their entirety, to obtain specific information about the rigidity of the arterial wall of the vessel. In order to determine the parameters, pulse waves are analyzed with regard to various properties, for example times, amplitudes, shapes or relevant points. All parameters generated from the analysis of pulse waves are subsequently referred to as "pulse wave parameters" or " PWP "Are referred to.

Various PWPs describe conclusions about the stiffness of the arterial wall, symbolically the most famous of these parameters, the pulse wave velocity ( PWV ). The evaluation of the stiffness of the arterial wall with the help of the PWPs in turn allows information about structural disturbances of the arteries, which partly precede certain vascular changes and partly develop in parallel.

A measurement of the stiffness of the arterial vessel wall which is customary as the gold standard according to the current state of the art comprises an invasive measurement of the PWV and further PWPs at two measuring points in the aorta with the help of a cardiac catheter. In humans, the cardiac catheter is usually inserted through an access via the femoral artery (A. femoralis). Pressure waves caused by the regular contraction of the heart are then recorded at two points invasively in the blood vessel, for example first in the ascending aorta (ascending artery) and secondly after the catheter has been withdrawn in the direction of the aortic fork ( Bifurcation) in right and left leg arteries. An ECG is often, but not always, registered at the same time.

The further procedure is usually as follows: The time difference between the R wave in the registered ECG and the base point of the pressure wave registered via the catheter results in a so-called transit time ( TT ). The transit time describes the time required for the pressure wave to travel from the contraction of the heart to the respective measuring point. The difference in transit times at the first measuring point and at the second measuring point then gives the transit time of the pressure waves between the first measuring point and the second measuring point. The distance between the first measuring point and the second measuring point is determined by withdrawing the catheter from the blood vessel, the withdrawal length of which being recorded. From the recorded transit time difference at the two measuring points and the withdrawal length of the catheter, the PWV be determined in the blood vessel.

Due to the high volatile fluctuations in the PWV and other PWPs, regular measurements on the persons concerned are recommended. For such measurements, non-invasive measuring systems are required, which are used in everyday life, outside of a catheter laboratory and sometimes also outside of a doctor's office. Non-invasive measurement systems, however, require validation, which can only be achieved by invasive measurements as described.

However, the invasive measurement method is increasingly changing. For cardiac catheter examinations, access via the wrist artery (radial artery) is predominantly chosen. For anatomical and technical reasons, however, the invasive determination of the PWV and some other PWPs cannot be withdrawn from the catheter. However, this means that future non-invasive systems cannot be validated invasively, or only with great effort, since the withdrawal method is a prerequisite for this, but is increasingly being used less.

In addition, studies showed a high interoperator variability in the manual evaluation of the invasive pressure curves using the retraction method. That means that the measured PWV and other PWPs are heavily dependent on the individual execution of the measurement. The reason for this are both fluctuations in the hemodynamics during the sequential measurement with withdrawal, as well as the time measurement between the ECG as the start point and the pulse wave as the end point, which is partly due to fluctuations in the pre-ejection period (PEP, PEP- Variability). This makes the objectivity and accuracy of the measured PWV already limited in the gold standard mentioned.

In addition, there is the comparatively high effort of manual evaluation. This ties up expensive personnel and represents an additional source of error. In addition, the withdrawal length (i.e. the distance from the measuring point in the ascending artery to the bifurcation) and the measuring points must be measured or determined as precisely as possible. This is often only possible to an insufficient extent under real measurement conditions.

It is therefore the task of specifying a method for the automated determination of hemodynamic parameters in a blood vessel of invasively recorded pulse waves by analyzing these pulse waves and / or of pulse waves selected from the entirety, but at least one pulse wave, in which an invasive measurement of the PWPs , especially the PWV , can be carried out in a simple manner, without a withdrawal method, with minimization of possible sources of error and with high accuracy. The evaluation should take place automatically and should therefore be easy, quick and delegable. The interoperator variability of the measurement process is to be reduced and thus the objectivity of the measurement is increased. The aim is also to increase the accuracy of the invasive gold standard method. Finally, a future-proof validation option for other non-invasive systems for determining PWPs is to be created.

The problem is solved with a method for the automated determination of hemodynamic parameters in a blood vessel on invasively recorded pulse waves by analyzing these pulse waves and / or pulse waves selected from the entirety, but at least one pulse wave with the features of claim 1. The subclaims contain expedient or advantageous embodiments of the method.

• Es erfolgt ein Einführen eines Druckmesskatheters mit einem Druckmesssensor bis zu einem vorgesehenen Messpunkt innerhalb des Blutgefäßes. Anschließend erfolgt ein Aufzeichnen und Speichern eines zeitlichen Druckverlaufs und Erfassen mindestens eines Pulswellensignals an diesem Messpunkt.

As an example for all PWPs, the method for invasively measuring a pulse wave velocity in a blood vessel is carried out with the following method steps:

• A pressure measuring catheter with a pressure measuring sensor is inserted up to a designated measuring point within the blood vessel. This is followed by recording and storing a pressure curve over time and recording at least one pulse wave signal at this measuring point.

• Zum einen wird z.B. ein sogenanntes Wellenzerlegungsverfahren (Wave-Separation) angewandt. Dabei wird durch Extrahieren eines Pulswellenechoanteils aus mindestens einem Pulswellensignal ein Zeitversatz zwischen einem Pulswellengrundsignalanteil und dem Pulswellenechoanteil im Pulswellensignal ermittelt. Aus diesem Zeitversatz wird mittels eines gegebenen Referenzabstandes zwischen einer Echostelle des Blutgefäßsystems und der Position des Druckmesssensors im Blutgefäßsystem eine PWV berechnet.

In a next method step, at least one pulse wave signal is analyzed to determine the PWPs, hereinafter using the example of PWV . This is done in at least two different ways, which can also be combined if necessary, and to make the determination more robust:

• On the one hand, a so-called wave separation method is used, for example. By extracting a pulse wave echo component from at least one pulse wave signal, a time offset between a pulse wave basic signal component and the pulse wave echo component in the pulse wave signal is determined. From this time offset, a given reference distance between an echo point of the blood vessel system and the position of the pressure measurement sensor in the blood vessel system PWV calculated.

On the other hand, one PWV determined with the help of the so-called wave analysis. In particular, the models of the Bramwell-Hill model and a given pressure / flow model of the blood vessel system have proven to be suitable, but are not limited to this.

The basic idea of the method is therefore to record the temporal pressure curve in the blood vessel at only a single measuring point in the blood vessel and to subsequently feed the pressure curve recorded in the process to automated signal processing. It is also essential to the invention to separate an echo signal from the recorded pressure curve, which is generated in the course of the blood vessel system itself. This echo signal is then converted into a time offset from which the PWV can be determined directly. It is also essential to the invention that an analysis is carried out on the at least one pulse wave recorded at the single measurement point with regard to the determination of relevant points and areas on the signal for determining further PWPs, for example the augmentation index, augmentation pressure and cardiac output.

In an advantageous embodiment of the method, a signal filtering of the stored temporal pressure curve is carried out using a high and / or a low pass filter to detect the pulse wave signal. Signal components in the pressure curve that are due to interference or outside a frequency range assumed to be typical are filtered out and are not included in the subsequent signal processing.

Advantageously, when the pulse wave signal is detected, individual pulse waves are classified, with a base point detection for identifying each individual pulse wave being carried out in a first step and a signal analysis for determining physiologically relevant signal points being carried out on the individual pulse wave in a second step. As a result, the shapes of the signal curve can be taken into account and incorporated into the subsequent evaluation.

The extraction of the pulse wave echo component according to the wave decomposition method takes place in an advantageous embodiment such that the classified pulse wave is broken down into a forward signal component and a backward signal component on the detected pulse waves to extract the pulse wave echo component. The forward signal component is the pulse wave basic signal, while the reverse signal component is the pulse wave echo signal.

A time difference is expediently determined as a time offset between the forward signal component and the backward signal component.

The time difference can be determined via the identified base points of the basic pulse wave signal and the pulse wave echo signal or via the cross-correlation method.

In one embodiment, a reflection distance related to a body size to a prominent point that generates a pressure echo within the blood vessel system can be used as the given reference distance.

In addition, can be determined from the wave decomposition process PWV -Value with that determined by the wave analysis PWV -Value a weighted PWV be determined.

The method steps are advantageously carried out independently of the platform, with a variation of the sampling rates of the pressure curve and the pulse wave signal approximating to a standard resolution to achieve comparability between different device-dependent recording techniques.

An arrangement for executing a method according to one of the preceding claims comprises a measurement sensor which can be inserted and placed in the blood vessel, an evaluation and storage unit for receiving, storing and processing the data of the pressure data recorded by the measurement sensor and a control implemented in the evaluation and control unit - and evaluation program.

In an expedient design, the measurement sensor is designed as an indirect measurement sensor consisting of an catheter that can be invasively inserted into the blood vessel and filled with a fluid, and a pressure sensor located outside the blood vessel.

The evaluation and control unit can also have an interface for connecting an EKG detection device. In this way, the pulse wave data can be monitored and evaluated together with EKG data.

The evaluation and control unit can expediently have a selection option for triggering a calibration and calibration process and / or for coordinating with a measuring unit to be connected.

Practical or advantageous embodiments of the method or the device are now given by way of example. The serve to clarify 1 to 10th .

An apparatus and a method for determining pulse wave parameters (PWPs), in particular pulse wave velocity ( PWV ) on invasively recorded pulse waves.

The basic principle of the method is a so-called single-point measurement carried out at only one location. During the measurement, the temporal signal curve of the pressure at the measuring point is recorded and evaluated. The aim of the method is to determine and evaluate a pulse wave echo that results from the Interaction of blood flow, blood pressure and the course of the arteries, for example from the aortic bifurcation located downstream of the body with the forking of the aorta into the left and right femoral artery. In principle, this is done in that a number of parameters, for example a number of functions and support points and their time intervals, are determined on the measured signal curve. For example, at a known reflection distance using the determined time intervals, the PWV be inferred.

The method is carried out, for example, using the following basic method steps, which are explained in more detail below. An invasive measurement in an arterial blood vessel is assumed here as an example.

In a first step, pressure curves over time are measured. These pressure curves over time represent arterial pressure curves. For this purpose, these pressure curves are recorded and stored in a human artery using a measuring unit. The pressure curve measurement is carried out invasively.

The measurement process is carried out, for example, in such a way that, in connection with a cardiac catheter examination, access is made via the artery on the wrist (radial artery) and the pressure measurement sensor is pushed into the artery to a certain point. In principle, however, the method is not restricted to this special access, but can in principle be carried out via any arterial access.

At least one pressure curve course, i.e. a pulse wave recorded. However, it is advisable to record several pressure curves over a certain time interval. The length of this time interval can be 10 seconds, for example.

• Ist der Druckmesssensor in der gewünschten Position in der Arterie platziert (z.B. in der A. Ascendens), werden die Signale mit den Druckschwankungen, das heißt mit den Druckverläufen gespeichert. Die aufgezeichneten Signale werden zunächst durch Filter von Störungen bereinigt. Hierzu können beispielsweise Hochpass- und/oder Tiefpass-Filter zur Anwendung kommen. Die Frequenzen von 0,5 Hz für den Hochpass-Filter und 50 Hz für den Tiefpassfilter haben sich dabei als geeignet herausgestellt, sind jedoch nicht darauf beschränkt. Anschließend werden durch ein Auswerteprogramm die einzelnen Pulswellensignale, insbesondere deren Start und deren Ende, in den Druckverläufen automatisiert erfasst. Dazu eignet sich beispielsweise die Fußpunktmethode, z.B. mit Hilfe der Intersecting-Tangent-Methode oder der Upstroke-Methode. Jedoch ist das Bestimmen der Pulswellen nicht auf diese Methoden beschränkt. Die detektierten Fußpunkte der Pulswellen sind wichtige Parameter bei der Bestimmung der Pulswellengeschwindigkeit.

In a next step, the individual pulse wave signals are extracted from the recorded pressure curves over time. This is done with the following steps:

• If the pressure measurement sensor is placed in the desired position in the artery (e.g. in the Ascendens artery), the signals are saved with the pressure fluctuations, i.e. with the pressure profiles. The recorded signals are first cleaned of interference by filters. For example, high-pass and / or low-pass filters can be used for this purpose. The frequencies of 0.5 Hz for the high-pass filter and 50 Hz for the low-pass filter have been found to be suitable, but are not limited to this. The individual pulse wave signals, in particular their start and end, are then automatically recorded in the pressure profiles by an evaluation program. The base point method is suitable for this, for example, using the intersecting tangent method or the upstroke method. However, the determination of the pulse waves is not limited to these methods. The detected base points of the pulse waves are important parameters when determining the pulse wave speed.

The detected pulse waves are then classified. The shape of the individual pulse waves is analyzed. This allows a classification into at least one class, which has a significant influence on the class-related further processing of the signals.

In a next step, a signal analysis is carried out on the classified pulse wave signals. Class-specific physiological points and properties of the pulse wave signal are determined, such as ejection duration ED (Ejection time of the heart), a so-called inflection point IP (Arrival of the reflected wave) or a characteristic impedance Zc (wave resistance). The points P1 (First Shoulder) and P2 (Second Shoulder) have proven to be useful support points for the classification. The Empiric Wavelet Transformation ( EWT ) together with a differentiation and Fourier transformations. The dynamic time equalization ( DTW ) was found to be suitable. However, the analysis steps described are not limited to these methods.

From these classified pulse wave signals, a pulse wave basic signal component and a pulse wave echo component are determined for the wave decomposition method, the time offset of which is determined from one another.

Both signal components can be determined and the pulse wave echo component can be extracted, for example, by determining a so-called flow curve. This flow curve can be statistically averaged, e.g. as a triangle starting at the base point ( F ), the maximum at Inflection Point ( IP ) and the end at the level of ejection duration ED be used. A more robust process is the so-called reservoir Excess model. The excess curve determined can be used as a flow curve. Subsequently, the classified pulse waves can be broken down into the respective forward-looking waves (after ejection), ie into the forward signal component and the backward-oriented waves (after reflection on the vessel walls and vascular branches), ie into the reverse signal component.

The forward signal component is then the pulse wave basic signal component, the reverse signal component then represents the pulse wave echo component in the individual pulse wave signal. In the next step, the time difference between the pulse wave fundamental signal component and the pulse wave echo signal component is determined. The time difference can be determined via the identified base points of the pulse wave basic signal and the pulse wave echo signal. The method of cross-correlation, which determines the time difference by shifting the two signals pulse wave basic signal and the pulse wave echo signal into one another, has proven to be more robust.

In order, for example, to use the wave decomposition method PWV to determine, the determination or the determination of a length is necessary. This length forms a reference distance for the determination of the PWV and is included in the signal processing as an evaluation parameter. In particular, this is the distance between the position of the pressure measurement sensor and the echo point in the blood vessel system, which causes the pulse wave echo signal component.

This reference distance can either be measured and then leads to the specification, in particular, of an arterial length, which is used as suitable for the reflection. However, it can also be e.g. determine an effective reflection distance (EfRD) based on the height of the person examined.

Alternatively, by using wave analysis PWV can be determined without specifying a distance. In particular, the models of the Bramwell-Hill model and a pressure / flow model of the blood vessel system have proven to be suitable, but are not limited to this. This means that the determination of the retreat length can be omitted. The relationship between volume and pressure (PV relation) reflects, for example, the so-called arterial compliance, which is required in the Bramwell-Hill model. With the help of the PU loop, which reflects the relationship between the flow velocity and the pressure in the artery, the increase can be started at the beginning PWV be derived at a local location. The Korteweg equation is also suitable, which also takes into account the dependence of blood pressure via the elasticity module of the blood vessel wall.

All approaches, such as the described wave decomposition method including the length and the wave analysis, which is independent of the length, can then be considered together, whereby the PWV is determined independently of one another using different approaches (fusion approach). The results can then be weighted against each other so that the final pulse wave speed of the real one PWV best matches.

The method is advantageously carried out independently of the platform. For example, by sampling the sampling rate and approximating to a standard resolution, comparability between different recordings (ie in particular different patients) and different measuring units (such as different types of catheters, so-called smart stents and the like) can be compared Measuring sensors) can be achieved. The values of 1,000 Hz and 12 bits have been found to be suitable, but are not limited to this.

• Es ist eine Messeinheit vorgesehen. Die Messeinheit besteht aus einem Messsensor, welcher an einer Steuer- und Speichereinheit angeschlossen ist. Dieser Messsensor kann direkt oder indirekt invasiv eingesetzt und an der gewünschten Stelle in der Arterie platziert werden. Eine Ausgestaltungsform sieht dabei vor, dass ein flüssigkeitsgefüllter Katheter mit dem Zugang über die A. Radialis in der A. Ascendens platziert und an einen Drucksensor angeschlossen wird. Dabei befinden sich der zu dem Katheter zugehörige Druckdomen und die Steuer- und Speichereinheit außerhalb des Messpunktes, d.h. insbesondere außerhalb des Körpers. Optional kann eine EKG-Erfassungseinrichtung vorgesehen sein, über die ein EKG (ein Elektrokardiogramm) parallel geschrieben und gespeichert werden kann.

The following arrangement is expedient for carrying out the method, the characteristics of which are also explained below using examples:

• A measuring unit is provided. The measuring unit consists of a measuring sensor which is connected to a control and storage unit. This measuring sensor can be used directly or indirectly invasively and placed at the desired location in the artery. One embodiment provides that a liquid-filled catheter with access via the radial artery is placed in the ascending artery and connected to a pressure sensor. The pressure dome associated with the catheter and the control and storage unit are located outside the measuring point, ie in particular outside the body. Optionally, an EKG detection device can be provided, via which an EKG (an electrocardiogram) can be written and stored in parallel.

Furthermore, a control and evaluation program is provided, which carries out the described steps for processing, processing and evaluating the stored pressure signals automatically.

An interface for exporting the data can also be provided, as can a GUI interface for displaying the information, such as, for example, the pressure signals and certain parameters, such as the base points P0 , the transit time TT , the ejection time of the heart ED , the inflection point IP , the pulse wave peaks e.g. P1 and P1 and / or the wave resistance Zc. An interface for performing updates is also useful.

In addition, in one embodiment, a switch for calibrating or calibrating the measuring system can be provided before each measurement (for example pressure calibration). In addition, depending on the measuring unit used (catheter, smart stent and such devices), an individual determination of the damping coefficient and the resonance frequency is useful.

The method mentioned and the example arrangement described are used for extracting and analyzing aortic pulse waves and for determining hemodynamic parameters (PWPs) of pulse wave analysis, such as e.g. Pulse wave speed used.

The method and the arrangement are additionally explained below with reference to the attached figures.

• 1 eine beispielhafte Grundkonfiguration der Messanordnung,
• 2 einen beispielhaften gemessenen Druckverlauf in einem Blutgefäß
• 3 eine Darstellung der Fußpunktmethode,
• 4 eine Darstellung einzelner, über die Fußpunkte erfasster Pulswellen,
• 5 eine prinzipielle Darstellung verschiedener Signalanteile an einem einzelnen Pulswellensignal,
• 6, 7 und 8 jeweils eine beispielhafte Darstellung möglicher Klassifizierungen der Pulswellenverläufe,
• 9 eine beispielhafte Darstellung markanter Punkte in einer Pulswelle,
• 10 eine beispielhafte Wellenzerlegung einer Pulswelle bei der Signalanalyse.

It shows:

• 1 an exemplary basic configuration of the measuring arrangement,
• 2nd an exemplary measured pressure curve in a blood vessel
• 3rd a representation of the base point method,
• 4th a representation of individual pulse waves detected via the base points,
• 5 a basic representation of different signal components in a single pulse wave signal,
• 6 , 7 and 8th each an exemplary representation of possible classifications of the pulse waveforms,
• 9 an exemplary representation of prominent points in a pulse wave,
• 10th an exemplary wave decomposition of a pulse wave in the signal analysis.

1 shows a basic configuration for an arrangement for carrying out the method. The arrangement includes a catheter 1 , a pressure sensor 2nd and a control and storage unit 3rd . The catheter 1 and the pressure sensor form a measuring sensor unit. In the example given here, it is used indirectly invasively and placed at the desired location in the artery. One embodiment provides that a liquid-filled catheter 1 with access through the A . Radial 4th in the A . Ascendens 5 placed and to the external pressure transducer 2nd is connected. The pressure transducer associated with the catheter is located here 2nd and the control and storage unit 3rd outside the measuring point, ie especially outside the body. Optionally, an EKG detection device 6 be provided, via which an EKG (an electrocardiogram) can be recorded and stored in parallel.

Furthermore, a control and evaluation program 7 be provided, which carries out the steps described in the method for processing, processing and evaluating the stored pressure signals. The control and evaluation program can be imported and updated externally into the evaluation and control unit.

It can also be an interface 8th be provided for the implementation of updates. There is also an interface 9 useful for exporting the data, as well as an internal or external GUI interface for a display 10th or an external evaluation unit 11 , for example a PC, with an external evaluation program 12th to display the information, such as the pressure signals and certain parameters, such as determined base points F , a transit time TT , an ejection time of the heart ED , an inflection point IP and / or the support points of pulse wave peaks such as P1 and P2 .

In addition, in one embodiment, a switch 13 be provided with which a procedure for calibrating or for calibrating the measuring system can be triggered before each measurement. This applies in particular to pressure calibration. In addition, when designing the control and evaluation program, a procedure is useful with which damping coefficients of different designs of the measuring unit used can be determined. This allows the properties of catheters, a so-called Take smart stents and other components into account when evaluating signals. An individual determination of the damping coefficient and the resonance frequencies of the above-mentioned components can thus be carried out and is useful in terms of measurement technology.

2nd shows the raw data initially acquired by the method. A pressure curve as a function of time is shown here. These signals are registered at the measuring point by the pressure sensor and recorded by the storage unit. Withdrawal and simultaneous registration of an EKG are not necessary. The data can thus also be collected, for example, via the access of the radial artery and evaluated without the influence of the PEP variability originating from the R peaks of the EKG. PEP variability is understood to mean fluctuations in the time between the electrical excitation of the heart (R peak in the ECG) and the actual start of ejection (foot point). This scatter leads to a blur in the PWV -Determination by the withdrawal method

The signal should be recorded in a sufficient resolution and sampling rate so that details in the pulse waves are retained. A sampling rate of 1,000 Hz and a processing width of 12 bits have been found to be suitable, but these parameters are not limited to this.

Both the sampling rate and the resolution are not fixed per se, but rather only device-specific sizes. To ensure platform independence and thus the best possible processability, the measured value quantity of the temporal pressure curve is then transformed either in the control and storage unit itself or in the external evaluation unit to a uniform standard so that it can be processed independently of the platform and thus standardized. For this purpose, the signal of the temporal pressure curve can, for example, be sampled up to a sampling frequency of 1,000 Hz. Various approximation methods can be used for this sampling process. In particular, e.g. the so-called cubic spline approximation is used. With this approximation, two neighboring sample points are approximated.

For the cubic spline approximation, polynomials of degree are used for two neighboring sample points (x _i , y _i ) and (x _{i + 1} , y _{i + 1} ) 3rd determined: $$s_i\left(x\right)=a_i+b_i\left(x-x_i\right)+c_i{\left(x-x_i\right)}^{\mathrm{2nd}}+d_i{\left(x-x_i\right)}^{\mathrm{3rd}}$$

$$s_i\text{'}\text{'}\left(x_{i+1}\right)=s_{i+1}\text{'}\text{'}\left(x_{i+1}\right)$$

Because of the continuity of the signal to be sampled, the continuity requirement s _i (x _{i + 1} ) = s _{i + 1} (x _{i + 1} ) applies, where s must be continuously differentiable twice: $$s_i\text{'}\left(x_{i+1}\right)=s_{i+1}\text{'}\left(x_{i+1}\right)$$

$$s_i\text{'}\text{'}\left(x_{i+1}\right)=s_{i+1}\text{'}\text{'}\left(x_{i+1}\right)$$

When a new Δx is selected, the signal can be resampled. If the output signal is sampled to a sampling frequency of 1,000 Hz, the following applies in this case: $$\mathrm{\Delta }x=\frac{1}{\mathrm{1,000}}$$

The choice of the type of splints with regard to the edges does not play a decisive role.

Signal interference can be eliminated by using suitable filters. The recorded signals are first cleaned of interference by filters. For example, high-pass and low-pass filters can be used for this. The frequencies of 0.5 Hz for the high-pass filter and 50 Hz for the low-pass filter have been found to be suitable, but are not limited to this.

Then the individual cardiac cycles, i.e. the individual pulse wave signals, in particular their start and end, are identified and extracted. Carrying out the first time derivative with a subsequent determination of the base point has turned out to be suitable, but is not limited to this.

3rd shows illustratively the base point method "Intersecting-Tangent" used for the extraction of the individual pulse waves PW from the waveform. In the footpoint method described here as an example, all turning points w (i) in the signal curve are determined with a positive increase. In the first derivation, these are characterized by the fact that the value of the first temporal derivation of the pressure curve is in their place P _PW extremal is: $$w=\text{bad}max\frac{d{P}_{PW}}{dt}$$

The variable w denotes the respective positions of the "positive" turning points on the time axis. The result is a set of all identified rashes, i.e. Peaks, within the pulse-like signal component.

Starting from the turning point determined in each case, the foot points become F of the individual pulse waves PW localized, for example, with the help of intersecting tangents in the vicinity of each individual maximum in the 1st derivative, that is to say in the vicinity of the previously determined turning point. In addition to the determined tangent t _W at the turning point w (i) becomes the first local minimum M searched in the time range before the individual pulse wave. This local minimum M has a horizontal tangent t _M on. Then the intersection S between the horizontal tangent t _M and the regression line, ie the tangent t _W calculated. This intersection S , projected onto the signal curve, is then the base point F the pulse wave PW . This point can also be adjusted subsequently.

4th shows the corresponding extracted pulse waves PW in the registered pressure history. These are through the base points F separated from each other and identified.

dabei ist s hier die sortierte Folge der ermittelten Wendepunkte.Then all peaks of the 1st derivative are sorted by size in descending order. $$s=sOrt\left(\frac{d{P}_{PW}}{dt}\left(w\right)\right)$$

s is the sorted sequence of the determined turning points.

This sorted sequence of peaks is analyzed step by step. Each peak is then recorded as a turning point relevant for the signal evaluation if its minimum height is greater than 0 and if its distance from all turning points already detected is greater than a defined time interval with respect to a predetermined sample rate FS.

• (a) s(i) > 0 und
• (b) $idx=\left\{w\left(i\right)|s\left(i\right)>0\wedge \left|\frac{w\left(i\right)-idx\left(j\right)}{FS}\right|\cdot 1.000>D,mitj<i\right\}$
  

Each peak w (i) from the sequence s (i) is therefore a turning point idx if the following conditions (a) and (b) are fulfilled at the same time:

• (a) s (i)> 0 and
• (b) $idx=\left\{w\left(i\right)|s\left(i\right)>0\wedge \left|\frac{w\left(i\right)-idx\left(j\right)}{FS}\right|\cdot \mathrm{1,000}>D,mitj<i\right\}$
  

Condition (b) expresses that a peak w (i) a certain minimum distance D to an already identified peak idx (j) must have. This minimum distance is independent of the sample rate. In the present example, this is, for example, 350 ms.

A subsequent ascending sorting of idx puts the indices in the correct chronological order.

In order to avoid any noise and minor interference in the signal, and even more so when differentiating, the 1st derivative can be calculated using the "Savitzsky-Goley" method.

The location of their respective base points is of crucial importance for the identification of the individual pulse wave in the pulsating signal component. As already described, these can be determined from the turning points.

wobei unter gap die jeweilige Lücke in den Samples zu verstehen ist, und y_i = P_PW(x_i) mit i =1,...,n die Regressionspunkte neben dem Wendepunkt idx(j). Dabei ist n die Anzahl der Regressionspunkte. Basierend auf den Koordinaten (x_i, y_i) der Fußpunkte F_j können nun Koeffizienten a,b einer Geradengleichung g(x) = ax + b ermittelt werden.For every foot point F with the coordinates F _j (x, y) $$x_i=isx\left(j\right)-i\cdot Gap$$

where gap is the respective gap in the samples, and y _i = P _PW (x _i ) with i = 1, ..., n the regression points next to the inflection point idx (j). Where n is the number of regression points. Based on the coordinates (x _i , y _i ) of the base points F _j , coefficients a, b of a straight line equation g (x) = ax + b can now be determined.

The pulse wave signals extracted in this way PW are analyzed further below. The essential core of signal processing is the separation of the individual pulse wave signal PW into a basic pulse wave signal PG and into a pulse wave echo signal PE .

5 shows the basic principle of the wave decomposition method of the single extracted pulse wave signal PW into the basic pulse wave signal PG and the pulse wave echo signal PE . The basic pulse wave signal PG describes the pressure curve of the pulse wave signal, which would occur if the blood vessel had no echo point. However, the pulse wave is partially reflected at the echo point of the blood vessel and is at the location of the invasive pressure measurement as a pulse wave echo signal PE registrable. Both signal components have a time offset from one another T on. This time offset allows, for example, a conclusion about the pulse wave speed PWV between the location of the invasive pressure measurement and the known echo site in the blood vessel system. In addition to determining the PWV the pulse wave signals can be further analyzed and classified in their form by breaking them down into the basic pulse wave signal and the pulse wave echo signal.

The 6 , 7 and 8th show examples of different categories of pulse waves. These differ in particular in their form. The pulse waves are classified accordingly. Decisive for the classification are the different forms in the course of each individual pulse wave signal PW , or the existence of points and features of a first shoulder P1 and a second shoulder P2 each individual pulse peak as well as a strength of a dicrotia or incisor I. the respective pulse wave.

In the classification, the first and the second shoulder of each individual pulse wave peak in each individual pulse wave individual signal describe the strength of the “double peak” of a peak. It describes the change in the pulse wave due to the arrival of a reflected wave. This reflected component is the pulse wave echo component. The so-called inflection point is located between the double peaks of the pulse wave signal IP . The correct position of IP has, for example, a high influence on the determined one in the case of wave decomposition experience PWV . The speed at which the pulse wave echo is reflected and thus the time interval between the support points P1 and P2 thus describes how “round” the individual period, ie the individual pulse wave, is in its course and how strongly the individual pulse wave signal deviates from the round shape. If the pulse wave echo comes back very quickly due to a stiffened blood vessel system, you can see a double peak. Here is usually the amplitude of P1 lower than from P2 . If the pulse wave echo comes back particularly slowly, one also speaks of a double peak, but here is P1 higher than P2 . Between these two forms the mixed form can be observed, which has only one peak. Based on this, each individual pulse wave can be classified into form classes. The number of calculation functions, i.e. the number of classes, is k , in which k is a natural number greater than 0.

The assignment to a class can be made according to the "nearest neighbor" principle. For example, the individual pulse wave signal belonging to a pulse wave PW have a left shoulder, while a first calculation function of class k = 1 has no left shoulder and a second calculation function of class k = 2 has a left shoulder. The shape of the individual pulse wave thus deviates more from the shoulderless calculation function than from the calculation function with the left shoulder. So the pulse wave belongs PW exemplary to class k = 2 and is thus represented by the calculation function of class k = 2. The class specification thus indicates which functions that have to be used according to certain properties, courses or symmetries must be used in order to “fit” the individual pulse wave signal.

The classification according to function types is not the only classification option. Another possibility to assign to a class is to perform the similarity analysis with the help of dynamic time equalization ( DTW ) to execute. The period lengths of the individual pulse wave signals and their times are designed to be invariant and thus enable a comparison. Through this comparison the forms of the pulse waves can be compared with each other and with the calculation functions and assigned to the classes.

A class-specific algorithm can be specified for each of these classes, which determines the hemodynamic parameters for the class. For example, those in the 6 , 7 and 8th pulse wave signals shown can be categorized into three different classes.

The individual pulse wave signals PW the pulse wave signals in 6 are, for example, two peaks, with the low peak on the left P1 has a shoulder and the closest peak P2 has a significantly higher amplitude. The incisor I. clearly describes the end of expectoration from the heart. This pulse waveform belongs, for example, to a class that belongs to a function set of type k = 1.

The individual pulse wave signals of the pulse wave signals in 7 are, for example, also two-peaked with a peak P1 and a peak P2 and a minimum in between. The characteristics of the amplitude of the peaks P1 and P2 have compared to the execution in 6 reversed. Now the peak has P1 a higher amplitude than P2 . The incisor I. clearly describes the end of expectoration from the heart. This pulse waveform can therefore be assigned to a function set of a class of type k = 2.

The individual pulse wave signals of the pulse wave signals in 8th are, for example, single-peaked and do not have clear peaks for P1 and P2 on. Due to the shift in the pulse wave echo and the resulting increase in the total amplitude of the pulse wave PW are the positions for P1 and P2 in the pulse wave PW hidden. The incisor I. clearly describes the end of expectoration from the heart. This pulse waveform can therefore be assigned, for example, to a function set of a class of type k = 3.

In order to avoid discontinuities in the transformation, the intermediate areas between the classes are kept smooth. This means that if, for example, a pulse wave can be assigned to two classes, each of these two classes is treated as equally likely.

For example, there is a pulse wave PW a little closer to one of these classes, the assignment is clear. The following then applies to a coefficient c̃: $$\tilde{c}=\alpha c_i+\left(1-\alpha \right)c_{i+1},\alpha =\left[0.1\right]$$

Here, c _i and c _{i + 1 are} the coefficients of the two nearest neighbors.

9 shows an exemplary representation of shape features that are detected on a single pulse wave signal by function classes and that can thus be classified. It shows an exemplary representation of a pulse wave with the points F (Base point), P1 , P2 , IP (Inflection Point; arrival of the reflected wave) and ED (Ejection duration; ejection time of the heart; F to incisure I. ). These class-specific physiological points and properties are determined, among other things, in the signal and serve as useful support in determining the pulse wave velocity and the PWPs.

The points P1 (First Shoulder) and P2 (Second Shoulder) have proven to be useful support points. To classify the individual pulse wave signal and translate it into a set of classified individual functions, the so-called dynamic time equalization ( DTW ) together with a cluster analysis and a Fourier transformation were found to be suitable. However, the signal processing is not limited to this.

$$P_b=\frac{P-z\cdot Q}{2}$$

10th shows an example of the wave decomposition method for determining the ejected and reflected pulse wave. The ejected pulse wave is the above-mentioned pulse wave basic signal component PG and the reflected pulse wave is the pulse wave echo component PE . In addition to this information, the so-called flow curve and the curve of the so-called reservoir pressure can also be determined and shown here. From the Ascendens Pressure curve AP the reservoir pressure curve is based on the 3-element wind boiler model using the reservoir excess model RP calculated. By subtracting the reservoir pressure curve RP from the Ascendens curve AP becomes the flow curve FLW (Flow) won. Using the flow curve ( Q ) and the Ascendens curve ( P ) become a forward wave FW ( P _f ) and a backward wave BW ( P _b ) separated. The values Q of the flow curve and the Ascendens curve P depend on the values P _f the forward wave and the values P _b the backward wave about the following relationships: $$P_f=\frac{P+\mathrm{e.g.}\cdot Q}{\mathrm{2nd}}$$

$$P_b=\frac{P-\mathrm{e.g.}\cdot Q}{\mathrm{2nd}}$$

The size z is a characteristic impedance of the blood vessel and serves as a scaling coefficient of the flow curve Q . The backward wave BW is the returning component of the pulse wave reflected at an echo point in the blood vessel, the forward wave FW the component of the pulse wave traveling in the direction of flow as a result of the amount of blood expelled from the heart in the blood vessel. The forward wave FW represents the forward signal component of the pulse wave signal, the backward wave BW the reverse signal component of the pulse wave signal PW .

A so-called transit time TT can be determined in different ways. Firstly, the difference in the base points between the backward wave and the forward wave. Since the base points in the backward wave can only be determined fuzzy, cross correlation is a suitable alternative. The required degree of shift of the backward wave towards the forward wave represents the TT . The dynamic time equalization ( DTW ) has proven to be suitable. Another method of determining the TT can be determined via the time interval between the first zero crossings of the averaged forward and backward wave. $$TT=NulldurcHGanG\left(P_b-mean\left(P_b\right)\right)-NulldurcHGanG\left(P_f-mean\left(P_f\right)\right)$$

The "mean" function here means that the mean value for the forward wave Pf and the backward wave Pb be calculated and then from the overall signal Pb respectively. Pf is subtracted. The signal thus swings around zero. The zero crossings are then used around the TT to determine.

The length of the blood vessel for the determination of the PWV , for example the aortic length, is measured independently. Alternatively, this can also be estimated using the following formula. The aortic length (distance) in relation to the height of the person (height) results from the empirical anatomical relationship: $$Distance=\frac{HeiGHt}{\mathrm{4th}}-7.28cm$$

The pulse wave speed PWV then results from the aortic length (distance) and the previously determined transit time TT to: $$PWV=\frac{\mathrm{2nd}\ast Distance}{TT}$$

Another approach to determining the PWV After the wave analysis, the so-called Moens-Korteweg equation can be used: $$PWV=\sqrt{\frac{E_{inc}\cdot H}{\mathrm{2nd}r\cdot \rho }}$$

One is the modulus of elasticity of the wall of the blood vessel, h the wall thickness of the blood vessel, r the clear radius of the blood vessel lumen and ρ the density of the blood fluid.

The modulus of elasticity E _inc this results as follows: $$E_{inc}={\mathrm{<figure-callout\; id="0"\; label="E"\; filenames="DE102019125062A1\_0018.png,DE102019125062A1\_0020.png"\; state="\{\{state\}\}">E</figure-callout>}}_0\cdot e^{b\cdot SBP}$$

E ₀ corresponds to E _inc at an initial pressure of zero. SBP corresponds to the systolic blood pressure, which is derived from the higher the amplitude of the respective pulse wave.

All approaches, both the wave decomposition method including the length and the wave analysis, which is independent of the length, can then be considered together, because the PWV is determined several times independently using different approaches (fusion approach). The results can then be weighted against each other so that the final pulse wave speed of the real one PWV best matches.

The method according to the invention was explained using exemplary embodiments. Further exemplary embodiments and modifications are possible within the scope of professional action. Further exemplary embodiments also result from the subclaims.

Reference list

1

Pressure measuring catheter

2nd

Pressure measuring sensor

3rd

Control and storage unit

4th

Access (e.g. A. Radialis)

5

Measurement location (e.g. A. Ascendes)

6

EKG

7

Control and evaluation program

8th

Interface (for updates)

9

Interface for exports and GUI

10th

Display

11

external evaluation unit

12th

external evaluation program

13

Calibration / calibration switch

AP

Ascendens-Pressure curve

BW

Backward wave

ED

Ejection time

FLW

Klow curve

FW

Forward wave

F

Base

I.

Incisor

M

first local minimum

P1

first partial peak

P2

second partial peak

PE

Pulse wave echo signal

PG

Basic pulse wave signal

PW

Pulse wave

PWV

Pulse wave velocity

RP

Reservoir pressure curve

S

Intersection

T

Time offset

TT

Transit time

t _M

horizontal tangent at the minimum

t _W

Tangent at the turning point

Claims (13)

Method for the automated determination of hemodynamic parameters in a blood vessel on invasively recorded pulse waves by an analysis of these pulse waves and / or of pulse waves selected from the entirety, but at least one pulse wave, with the method steps - Inserting a pressure measuring catheter (1) with a pressure measuring sensor (2) up to an intended measuring point within the blood vessel, Recording and storing a temporal pressure curve and detecting at least one pulse wave signal (PW) at the measuring point, Extracting a pulse wave echo component (PE) from the at least one pulse wave signal (PW) and determining a time offset (T) between a pulse wave fundamental signal component (PG) and the pulse wave echo component (PE) in the pulse wave signal (PW), - Converting the time offset (T) into a pulse wave velocity (PWV) using a given reference distance between an echo point of the blood vessel system and the position of the pressure measurement sensor in the blood vessel and / or - Determination of in particular a pulse wave velocity (PWV) without considering a reference distance via a given pressure / flow model and a pressure / volume model of the blood vessel system. Procedure according to Claim 1 , characterized in that for the detection of the pulse wave signal (PW) a signal filtering of the stored temporal pressure curve is carried out using a high and / or a low pass filter. Procedure according to Claim 1 or 2nd , characterized in that when the pulse wave signal (PW) is detected, individual pulse waves are classified, with each individual pulse wave being identified in a first step, in particular by foot point detection, and a signal analysis for the physiological determination of the individual pulse wave in a second step relevant signal points is executed. Method according to one of the preceding claims, characterized in that on the detected pulse waves (PW) for extracting the pulse wave echo component, the classified pulse wave is broken down into a forward signal component (FW) and a reverse signal component (BW), the forward Signal component (FW) is the pulse wave basic signal (PG) and the reverse signal component (BW) is the pulse wave echo signal (PE). Procedure according to Claim 4 , characterized in that a time difference is determined as a time offset between the forward signal component (FW) and the reverse signal component (BW). Procedure according to Claim 5 , characterized in that the time difference is determined via the distance between the base points of the pulse wave basic signal and the pulse wave echo, or via a cross correlation, the displacement of the two signals pulse wave basic signal (PG) and the pulse wave echo signal (PE) describing the time difference from one another, or via the distance the zero crossings of the averaged signals of the pulse wave basic signal (PG) and the pulse wave echo signal (PE) is determined. Procedure according to Claim 1 , characterized in that a reflection distance related to a body size to a point generating a pressure echo within the blood vessel system is used as the given reference distance. Procedure according to Claim 1 , characterized in that for the hemodynamic parameter to be determined, in particular the pulse wave velocity (PWV), or intermediate steps for determining the hemodynamic parameter, which for individual pulse waves from time offset and reference length and / or from pressure / flow model and / or pressure / volume -Model of the blood vessel was determined, an averaging takes place. Procedure according to one of the Claims 1 to 8th , characterized in that the method steps are carried out independently of the platform, with a variation of the sampling rates of the pressure curve and the pulse wave signal (PW) is approximated to a standard resolution, comparability between different device-dependent recording techniques is carried out. Arrangement for carrying out a method according to one of the preceding Claims 1 to 9 comprising a measuring sensor unit that can be inserted and placed in the blood vessel, an evaluation and storage unit (3) for recording, storing and processing the pressure data recorded by the measuring sensor unit, and a control and evaluation program implemented in the evaluation and control unit. Arrangement after Claim 10 , characterized in that the measuring sensor unit is designed as an indirect measuring sensor consisting of a catheter (1) which can be invasively inserted into the blood vessel and a pressure sensor (2). Arrangement after Claim 10 , characterized in that the evaluation and control unit (3) has an interface for connecting an EKG detection device (6). Arrangement according to one of the Claims 10 to 12th , characterized in that the evaluation and control unit (3) has a selection option for triggering a calibration and calibration process and / or for coordination with the measurement sensor unit to be connected.

Images