METHOD FOR ESTABLISHING AND/OR MONITORING THE STATE OF AN EXTRACORPOREAL FLUID OR FLUID FLOW BY MEANS OF ULTRASOUND

Abstract

The invention relates to a method for establishing and/or monitoring foreign structures in an extracorporeal fluid or in a fluid flow, in particular in blood or a bloodstream, wherein the fluid is monitored by means of ultrasound. The method according to the invention is characterized in that the features of the fluid state established by means of the ultrasound monitoring are processed by means of a multi-criteria ultrasonic analysis. The invention furthermore relates to a device for performing this method and the use of this device.

Claims

1. A method for establishing and/or monitoring the state of an extracorporeal fluid or of an extracorporeal fluid flow, wherein the fluid is monitored by means of ultrasound, wherein the extracorporeal fluid or the extracorporeal fluid flow is subjected to time-harmonic mechanical and/or transient mechanical excitation and produced ultrasonic backscattering signal is analyzed in terms of Doppler frequency shifting and elastography, features of the state of the extracorporeal fluid established by means of ultrasonic monitoring are processed with a multi-criteria ultrasonic analysis, in particular an analysis algorithm, by which a change in the distribution of particles contained in the fluid and changes in the viscosity of the fluid are measured by means of ultrasound and the fluid state is assigned to previously defined states.

2. The method of claim 1 wherein the extracorporeal fluid is blood and is monitored by means of ultrasound, wherein extracorporeal bloodstream is subjected to time-harmonic mechanical and/or transient mechanical excitation and the ultrasonic backscattering signal is analyzed in terms of Doppler frequency shifting and elastography, the features of the state of the extracorporeal blood established by means of ultrasonic monitoring are processed with a multi-criteria ultrasonic analysis, in particular an analysis algorithm, by which a change in the distribution of constituents in the blood and changes in the viscosity of the blood are measured by means of ultrasound and the blood state is assigned to previously defined states, in particular clotting states.

3. The method of claim 1, wherein at least one first foreign structure, in particular a solid body such as a blood clot, is detectable and/or distinguishable from at least one second solid/liquid/gaseous foreign structure, e.g., air bubbles, in the extracorporeal fluid.

4. The method of claim 1, wherein a plurality of features of the extracorporeal fluid state are established instantaneously by means of ultrasonic monitoring.

5. The method of claim 1, wherein the analysis algorithm comprises the following signal processing steps: a. Reception of an ultrasonic signal or a plurality of ultrasonic signals, b. Extraction of features from the ultrasonic signal(s), c. Pattern recognition, d. Further processing, and e. Assignment of a score, wherein said score defines the quality (probability) for a given classification result.

6. The method of claim 1, wherein all signal processing steps are optimized offline on the basis of a machine learning algorithm.

7. The method of claim 1, wherein the individual state of the extracorporeal fluid, in particular the blood state, is in addition learnable online by means of the analysis algorithm.

8. The method of claim 1, wherein the extracorporeal fluid is blood and the coagulation state of the blood is assignable, with reference to online-detected changes, to interventions such as the administration and/or concentration increase of an anticoagulant during hemodialysis.

9. The method of claim 1, wherein the extracorporeal fluid is blood and physical parameters that are subject to change by progressing blood coagulation are measured by ultrasonic backscattering.

10. The method of claim 9, wherein physical parameters for detecting incipient blood coagulation are measured by ultrasonic backscattering.

11. The method of claim 9, wherein the physical parameters are selected from the group consisting of sonic speed, diameter of the scattering body, the standard deviation of the backscattering, the maximum deviation of the backscattering, the degree of regularity of the arrangement of the scattering bodies, the turbulence and the velocity distribution of the scattering bodies, the pressure in the tube system.

12. The method of claim 9, wherein the aggregation of thrombocytes and/or erythrocytes is measured.

13. The method of claim 1, wherein errors are detected in an extracorporeal fluid conduit.

14. A device for monitoring state of an extracorporeal fluid flow comprising at least one ultrasonic monitoring means and at least one signal analysis means, wherein the extracorporeal fluid flow, in particular the extracorporeal bloodstream, is subjected to time-harmonic mechanical and/or transient mechanical excitation and the ultrasonic backscattering signal is analyzed in terms of Doppler frequency shifting and elastography, the features of the state of the extracorporeal fluid, particularly of the state of extracorporeal blood, established by means of the ultrasonic monitoring are processed with an analysis algorithm, by which changes in the distribution of particles, in particular blood constituents, contained in the extracorporeal fluid and changes in the viscosity of the extracorporeal fluid, in particular of extracorporeal blood, are measurable by means of ultrasound and the state of the extracorporeal fluid, in particular of the extracorporeal blood, is assignable to previously defined states, in particular to coagulation states.

15. The device of claim 14, wherein the signal analysis means detects at least one first foreign structure, in particular one or a plurality of solid bodies, in particular blood clots, and distinguishes the first foreign structure from at least one second foreign structure in the extracorporeal fluid.

16. The device of claim 14, wherein the signal analysis means detects changes in the distribution of particles, in particular blood constituents, contained in the extracorporeal fluid and changes in the viscosity of the extracorporeal fluid, in particular extracorporeal blood, records signal-triggering events, and emits a signal.

17. The device of claim 14, wherein the device has a receptacle adapted to receive a fluid conduit means, preferably a tube system or a portion of a tube system or a cartridge, in particular a disposable cartridge, or a measurement channel.

18. The device of claim 14, wherein the device has at least one alarm means and/or is hooked up to at least one alarm means for alerting changes in the distribution of particles, in particular blood constituents, contained in the extracorporeal fluid, and to changes in the viscosity of the extracorporeal fluid, in particular of extracorporeal blood.

19. The device of claim 14, wherein the device has at least one protection means and/or is hooked up to at least one protection means for stopping extracorporeal fluid flow and/or at least one means for correcting the state of the extracorporeal fluid, in particular the clotting state of extracorporeal blood by addition of anticoagulants.

20-22. (canceled)

23. A device for monitoring state of an extracorporeal fluid which comprises a conduit for extracorporeal fluid; an ultrasonic transducer operably associated with the conduit; an ultrasonic back scattering detector operably associated with the conduit; and a multi-criteria signal analyzer operably associated with the ultrasonic back scattering detector.

Description

[0134] In the following, further details of the invention will be explained more fully in the drawings and in an exemplary embodiment.

[0135] Shown are:

[0136] FIG. 1: the typical structure of the analysis algorithm according to the invention.

[0137] FIG. 2: the classification results from the differentiation of particle sizes in varying volume concentrations. The sizes are expressed in m on the axes.

[0138] FIG. 3: the typical structure of extracorporeal blood circulation system with the corresponding ultrasonic monitoring means according to the present invention.

EXAMPLE

[0139] The aim is to develop bases for the scattering theory of blood cell aggregation, impact of the coagulation cascade on the physical characteristics of the blood, and medically relevant parameters.

[0140] A first population of blood mimicking fluids (suspensions) with slightly varying volume concentrations at different mean particle sizes (5, 20 and 40 m) was used for training an analysis algorithm. A second population with the same volume concentrations and particle sizes was then used for testing the quality of the analysis algorithm.

[0141] 1. Technical design

[0142] An ultrasound device possessing a high signal-noise ratio was used for the experimental verification. The device was optimized for Doppler analysis by transmitting an ultrasonic burst (several oscillation periods). In addition, an analysis software permitting a high scanning speed of the pulse repetition frequency was implemented. This was required in order to satisfy Shannon's sampling theorem for the elastographic analysis. The classification was carried out with Matlab (The MathWorks Inc.).

[0143] The mechanical assembly was performed on a vibration couch with two electromechanical transducers. A holder for the cuvettes was placed on the two transducers. The suspension was made to vibrate by operating the transducers in counter-phase mode. The assembly comprised an analysis computer, units for supplying power to the stirrer and the ultrasound interface, as well as a cuvette that was mounted on a cuvette holder, i.e., on the transducers.

[0144] The plane-parallel walls made it possible to generate a complex network of modes, which interacted with the viscous properties of a suspension. For the Doppler measurements, a stirrer was immersed in the suspension and ultrasonic measurements were performed at different stirring speeds. For pure backscattering measurements, the suspension was stirred and measurements were taken with the stirrer operating at a very low rotational speed.

[0145] 2. Preparation of a blood-like fluid

[0146] Human blood is not used directly in many areas of medical research, because among other things it does not store well. Because often only a few specific characteristics are used, which only in these characteristics behave like true blood. In addition to animal blood, partially and fully synthetic fluids, so-called blood-mimicking fluids (BMF), are also used. BMFs suitable for examinations to detect blood coagulation (hemostasis) by means of ultrasonic backscattering are used in this example.

[0147] BMF preliminary examinations for the backscattering method were performed on different

[0148] BMFs, which had to fulfill the following requirements: [0149] different particle sizes, [0150] suspensions, i.e., separable particle-solvent mixtures, [0151] chemical and physical stability of the BMF.

[0152] BMFs were produced using various substances. The dependency of the BMF properties on the properties of the substances was investigated.

[0153] Materials: Two different substances were used to prepare BMFs: ORGASOL (Arkema Inc.) and megaCRYL (megadental GmbH).

TABLE-US-00003 TABLE 3 Properties of the substances used megaCRYL ORGASOL Material PMMA PA12 Diameter 10-60 m 5 m; 20 m; 40 m Density 1.16 g/cm.sup.3 1.03 g/cm.sup.3 Original intended use Cold-cure denture Paint industry additive polymerizer PMMA Polymethyl methacrylate PA12 Polyamide 12

[0154] Both substances exhibited different behavior when stirred and in terms of the storability of the suspensions, as shown in Table 4.

TABLE-US-00004 TABLE 4 Properties of the BMFs produced megaCRYL ORGASOL ORGASOL ORGASOL Scattering body 10-60 m 5 m 20 m 40 m diameter Dispersity polydisperse monodisperse monodisperse monodisperse Suspendability in very good poor good good water Viscosity yes no no no influenced by concentration? Resuspendability moderate moderate good good Chem. stability Polymerization 1) 1) 1) Phys. stability 1) 1) 1) 1) Biol. stability 2) 2) 2) 2) Highest 33 M. % 25 M. % 18 M. % 18 M. % concentration produced 1) no change observed thus far, i.e., the suspensions remained unchanged for months. 2) the suspensions do not contain any biological components, and 3) maximum possible concentration with additives used thus far.

[0155] Viscosity Measurement

[0156] A dependency of the viscosity on the concentration was established for the megaCRYL suspensions. The viscosity of the ORGASOL suspensions is similar to that of water.

[0157] Database

[0158] For verifying the classification method, the suspensions in Table 5 were created. As in the eventual dialysis machine application, a reliable detection of the particle diameter of individual blood cells and their microaggregates should be possible in spite of a varying volume concentration (hematocrit).

TABLE-US-00005 TABLE 5 Database for verifying the classification of suspensions differing in terms of both particle diameter and particle number. Suspension (Set/Diameter) Percent by volume 1/5 10.1 0.23 1/20 9.8 0.24 1/40 9.3 0.25 2/5 11.7 0.24 2/20 11.7 0.24 2/40 11.7 0.24 3/5 13.8 0.25 3/20 13.8 0.25 3/40 13.6 0.24 4/5 14.4 0.25 4/20 14.6 0.25 4/40 14.5 0.25 5/5 15.7 0.25 5/20 15.8 0.26 5/40 15.8 0.26 6/5 17.6 0.26 6/20 17.8 0.26 6/40 17.5 0.26

[0159] 3. Performance of the measurements

[0160] The measurements were performed using the ft2232 (GAMPT) software with the following settings: prt 1999 (plus-repetition rate), data num. 2048 (number of data points of the A-scan), delay 5 (start of the recording of data points), gain 50 (analog amplification in dB re 1V), transmitter 30 (analog amplification transducer dB re 1 V), filter 1 100 (no filtering), filter 2 100 (no filtering), spl 4000 (recording 4000 A-scan lines), fs 50 MHz (scanning frequency), burst f 3,125 Mhz (signal frequency), TGC max. (full digital amplification) in the measurement range using a panametrics transducer with a center frequency of 3.5 MHz.

[0161] From the measurements of sets 4 to 6, the features backscattering, Doppler, and elastography were determined on samples taken at different times, over a recording time of 0.6 seconds in each case. Accordingly, there were 9 suspensions7 excitation forms (stirring speed, etc.)6 different time samples for the training. The measurements and feature extractions of the 7 excitation forms were arranged in a vector, which ultimately resulted in 54 samples. The same number was used for the evaluation of the analysis algorithm. Both populations are temporally independent of one another. The analysis of backscattering, Doppler, and elastography was performed for each measurement, regardless of the excitation form. The depth range was limited to 300 data points in the backscattering range in order to avoid effects due to different fill levels in the cuvette. Each sample resulted in a vector of 7 excitation types1068 features in each case. After they were calculated, all features were logarithmized (base e).

[0162] The seven excitation forms comprised supplying the stirrer with 0.7, 1, 1.5, and 2V (Hewlett Packard E3611A, at full amperage) of power as well as harmonic excitation with 5, 8, and 15 Hz (tamp TSA 1400 amplifier at 50% output, PC sound card type: Realtek Audio High Definition sound card, maximum amplification, Matlab with an amplitude factor of 0.5), with the stirrer turned on. Before each measurement, the suspension was briefly stirred in order to avoid sedimentation. The stirring speed at 0.7 V corresponds to a stationary state of the suspension per A-scan, which was a required criterion for the backscattering analysis performed here.

[0163] The evaluation results of the trained analysis algorithm are presented in FIG. 2 in the form of a so-called confusion matrix. The ordinate gives the actual label, in this case the diameter. The abscissa designates the classification results. The results are expressed as percents. A perfect classification would lead to a principal diagonal with values of 100% in each case. A very high classification quality with an error of 2% was attained in actuality.

[0164] The results of the described measurement verify the validity of the approach presented here of the multi-criteria examination of fluids for establishing particle size differences with varying volume concentrations far below the resolution of the ultrasonic signal (the wavelength of the ultrasonic signal used in water was 480 m).