Blood filtration machine provided with a measuring system including an optical sensor

Abstract

A blood filtration machine has a blood circuit and a measuring system, the blood circuit has a plurality of conduits made of a transparent material, and the measuring system has a plurality of optical sensors connected to the respective conduits. Each optical sensor has a reading window placed at a point of each conduit, a light emitter, and a light receiver. The measuring system includes a single spectrometer, an optical mixer, and a control unit, the optical mixer includes a plurality of input terminals connected to the light receivers of corresponding ones of the optical sensors, and output terminals connected to the input terminals of the spectrometer, and the control unit is configured to activate the light emitters of one optical sensor at a time so as to measure the parameters of one organic liquid at a time.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This patent application claims the priority of Italian Patent Application No. 102020000006706 filed on March 31, 2020, the entire disclosure of which is incorporated herein by reference. Technical Field

[0003] The invention relates to a blood filtration machine provided with a measuring system comprising a plurality of optical sensors. Background Art

[0004] As is known, a blood filtration machine generally comprises a blood circuit, wherein a filter is arranged on the blood circuit and a second filter may be arranged on the plasma circuit, the purpose of arranging the filters depends on the type of treatment to be applied to the patient, in the case of apheresis, the purpose of arranging the second filter is to remove specific molecules in the plasma.

[0005] Most optical sensors on the market for measuring intrinsic parameters of organic liquids operate based on specific wavelength characteristics of the parameter to be measured. For example, hematocrit measurement sensors operate at specific absorption frequencies of hemoglobin, water and references. The light source of this type of optical sensor is a narrowband light emitter that emits light in a narrowband frequency that is a function of the physical parameter of the organic liquid to be measured. This type of sensor also has a light receiver that measures the light intensity at different specific frequencies of the light refracted by the test organic liquid.

[0006] This type of optical sensor has some disadvantages.

[0007] Firstly, the above-mentioned optical sensors are highly sensitive to changes in certain operating parameters of the measuring system, such as the intensity of the light emitted by the light source, the opacity of the reading window or the temperature, which leads to poor measurement reproducibility. Therefore, in order to obtain the required measurement reproducibility, the measuring system including such an optical sensor becomes complicated and expensive. In addition, the above-mentioned optical sensors do not allow the creation of flexible measuring systems, since the narrow-band light emitter must operate at a specific frequency associated with each parameter to be measured, such as pH, saturation, platelet count, hemoglobin concentration in solution or hematocrit. In other words, if it is necessary to measure multiple physical parameters of the organic liquid, the measuring system must include different optical sensors, i.e., multiple optical sensors designed to operate at different wavelengths.

[0008] These disadvantages limit the use of this type of optical sensor in blood filtration machines.

[0009] It is well known that during certain types of hemofiltration treatments, such as continuous renal replacement therapy (CRRT), patients lose weight, which must be measured with extreme accuracy to prevent serious side effects, such as fainting or seizures.

[0010] Some known systems / devices for weight loss measurement installed in hemofiltration machines have problems with overall size and production costs, which have a significant impact on the overall size and overall cost of the hemofiltration machine. This problem is particularly evident in a specific class of hemofiltration machines, namely dialysis machines, where the dialysate comes from a centralized hospital system, so a scale cannot be used to measure weight loss.

[0011] Some solutions for measuring the weight loss of patients in dialysis machines are based on volumetric systems or differential flow meters. However, dialysis machines on the market are not equipped with a secondary and independent system to repeatedly measure the weight loss of patients during treatment to ensure greater safety for the patients themselves, since the measurement is redundant and any failure of the primary measurement system can be easily identified that would put the patient's life at risk, especially for the most vulnerable subjects. Summary of the invention

[0012] The object of the present invention is to provide a blood filtration machine which is able to repetitively measure a plurality of intrinsic parameters of an organic fluid in a blood circuit and which is simple and inexpensive to manufacture.

[0013] According to the present invention there is provided a blood filtration machine as defined in the accompanying claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention will now be described with reference to the accompanying drawings, which show non-limiting embodiments of the invention, in which:

[0015] Figure 1 A blood filtration machine of a type suitable for carrying out apheresis treatment is schematically shown;

[0016] Figure 2 schematically illustrates a blood filtration machine of a type suitable for carrying out a hemodialysis treatment; and

[0017] Figure 3 Schematically showing the Figure 1 or Figure 2 A measuring system for a machine in a vehicle, comprising a plurality of optical sensors. DETAILED DESCRIPTION

[0018] exist Figure 1 and Figure 21 generally denotes a blood filtration machine. The machine 1 comprises a blood circuit 2 comprising at least one filter, the nature of which depends on the type of treatment to be administered to the patient, and a plurality of tubes 5 made of a transparent material and designed to allow the corresponding liquid to flow therethrough.

[0019] Reference Figure 1 , the blood circuit 2 is configured to perform apheresis treatment and comprises a plasma filter 3, a fractionation filter 4, a selection clamp 6 and a cleaning device 16 interconnected by a plurality of tubes 5. The blood circuit 2 comprises a plurality of pumps, which are generally indicated by 15 and are preferably composed of respective peristaltic pumps to maintain the flow of liquid in some of the tubes 5.

[0020] The task of the plasma filter 3 is to filter the blood taken from the patient in order to separate the particulate portion of the blood consisting of red blood cells, white blood cells and platelets from the liquid portion of the blood formed by plasma. The particulate portion of the blood flows back to the patient, while the plasma is sent to the fractionation filter 4. To this end, the plurality of conduits 5 include a conduit 5a for conveying the blood taken from the patient to the plasma filter 3, a conduit 5b for conveying the particulate portion of the blood from the first output end of the plasma filter 3 to the patient, and a conduit 5c for conveying the plasma from the second output end of the plasma filter 3 to the fractionation filter 4. The first output end of the plasma filter 3 is preferably arranged at the upper end of the plasma filter 3.

[0021] The task of the fractionation filter 4 is to filter the plasma in order to separate the high molecular weight portion, i.e. the plasma portion including cholesterol (LDL), immunoglobulins and cryoglobulins, from the low molecular weight portion (ultrafiltrate), i.e. the plasma portion including albumin, IgG and HDL. The high molecular weight portion is retained in the fractionation filter 4, and the low molecular weight portion must be returned to the patient.

[0022] Then, the high molecular weight portion of the plasma can be removed by a special fractionation filter washing step. To this end, the plurality of conduits 5 include a conduit 5d for transferring the plasma (ultrafiltrate) from the first output end of the fractionation filter 4 to the conduit 5b through the selection clamp 6 so that the plasma can be returned to the patient, and a conduit 5e for connecting the second output end of the fractionation filter 4 to the washing device 16 through the selection clamp 6 so that the fractionation filter 4 can be washed. The conduit 5d is connected to the conduit 5b downstream of the selection clamp 6 by a bifurcation.

[0023] The task of the selection clamp 6 is to select the opening of the duct 5d individually, or the opening of the duct 5e individually, respectively, depending on whether the machine 1 is in a filtering step or a cleaning step of the fractionation filter 4. In other words, during the filtering step, the selection clamp 6 is set to open the duct 5d and close the duct 5e, so that the plasma coming out of the fractionation filter 4 can flow back to the patient along the ducts 5d and 5b, while in the cleaning step, the selection clamp 6 is set to open the duct 5e and close the duct 5d, so that the cleaning device 16 can collect the cleaning solution that has passed through the fractionation filter 4 and is therefore enriched in the high molecular weight fraction of the plasma.

[0024] In particular, the cleaning device 16 comprises a first capsule 16a for containing a clean cleaning solution and a second capsule 16b for collecting a "dirty" cleaning solution, i.e., a high molecular weight fraction rich in plasma, coming from the fractionation filter 4 through the conduit 5e. The plurality of conduits 5 comprises another conduit 5f for conveying the clean cleaning solution from the cleaning device 16 to the fractionation filter 4 through the conduit 5c. In other words, the conduit 5f is connected to the conduit 5c by a bifurcation.

[0025] The plurality of pumps 15 include a pump 15a preferably arranged in the region of the duct 5a, a pump 5c preferably arranged in the region of the duct 5c, and a pump 15f preferably arranged in the region of the duct 5f.

[0026] During the filtration step, the machine 1 is connected to the patient and pumps 15a and 15c are turned on to push blood drawn along conduit 5a towards the plasma filter 3 and plasma along conduit 5c towards the fractionation filter 4, respectively, while pump 15f is turned off so that nothing flows in conduits 5e and 5f.

[0027] During the cleaning step, the machine is not connected to the patient, and pumps 15a and 15c are turned off, so nothing flows in the conduits 5a-5c, while pump 5f is turned on to take the cleaning solution from capsule 16a and push it through conduits 5f and 5c to the fractionation filter 4. The cleaning solution passes through the fractionation filter 4, thereby removing the high molecular weight portion of the plasma, and the dirty cleaning solution comes out of the fractionation filter 4, passes through conduit 5f, and is collected by capsule 16b.

[0028] Reference Figure 2 , the blood circuit 2 is configured to perform a hemodialysis treatment and comprises a dialyzer filter 23. For example, the dialyzer filter 23 is of a type comprising two compartments, which are separated from each other by a semipermeable membrane and through which the blood to be treated and the dialysate (also called dialysate) flow, respectively, in opposite directions. Blood impurities enter the dialysate through the semipermeable membrane according to a mechanism known per se, and are therefore not described in detail herein.

[0029] The multiple conduits 5 include a conduit 5a for conveying blood taken from a patient to a first input end of the dialyzer filter 23, a conduit 5b for conveying treated blood from a first output end of the dialyzer filter 23 to the patient, a conduit 5g for conveying clean dialysate to a second input end of the dialyzer filter 23, and a conduit 5c for extracting "dirty" dialysate, i.e., rich in impurities extracted from the blood, from a second output end of the dialyzer filter 23.

[0030] The plurality of pumps 15 include a pump 15a preferably arranged in the region of the duct 5a, a pump 15c preferably arranged in the region of the duct 5c, and a pump 15g preferably arranged in the region of the duct 5g.

[0031] Reference Figure 1 and Figure 2 The machine 1 comprises a measuring system 7 designed to measure one or more intrinsic parameters of the organic liquid, such as the amount of hemoglobin, hematocrit or platelets in the blood circuit 2 , at several points in the blood circuit 2 .

[0032] The measuring system 7 comprises a plurality of optical sensors, in particular three optical sensors indicated by 8a, 8b and 8c, arranged in different points of the blood circuit 2, and an acquisition system 9 connected to the optical sensors 8a-8c for acquiring and processing the signals provided by the optical sensors 8a-8c in order to provide a measurement of at least one parameter of at least one organic liquid in the blood circuit 2.

[0033] Reference Figure 1 In the embodiment of the present invention, the optical sensor 8a is preferably arranged upstream of the plasma filter 3 relative to the blood flow direction, that is, arranged at a point of the conduit 5a. The optical sensor 8b is preferably arranged downstream of the plasma filter 3 relative to the blood flow direction, that is, arranged at a point of the conduit 5b. The optical sensor 8c is preferably arranged upstream of the fractionation filter 4 relative to the plasma flow direction, that is, arranged at a point of the conduit 5c.

[0034] Reference Figure 2 In the embodiment in FIG. 5 , the optical sensor 8a is preferably arranged upstream of the dialyzer filter 23 with respect to the blood flow direction, ie, at a point of the conduit 5a.

[0035] The optical sensor 8b is preferably arranged downstream of the dialyzer filter 103 with respect to the blood flow direction, ie at a point of the conduit 5b.

[0036] The sensor 8c is preferably arranged at a point of the conduit 5c.

[0037] Reference Figure 3Each optical sensor 8a, 8b, 8c includes a reading window 10, a light emitter 11, and a light receiver 12. The reading window 10 is arranged at a point of each conduit 5a, 5b, 5c of the blood circuit 2, and in particular, wraps at least a portion of the conduits 5a, 5b, 5c so as to be able to see the organic liquid flowing in the conduits 5a, 5b, and 5c. The light emitter 11 is designed to emit light to the conduits 5a, 5b, and 5c through the reading window 10, and the light receiver 12 is designed to receive the light emitted by the light emitter 11 and has passed through the conduits 5a, 5b, and 5c again through the reading window 10.

[0038] Preferably, the light emitter 11 is of LED type.

[0039] The acquisition system 9 comprises a single spectrometer 14 connected to the light receivers 12 of the optical sensors 8a, 8b and 8c via an optical mixer 13, and a control unit 17 driving the light emitters 11 of the optical sensors 8a, 8b and 8c and connected to the output of the spectrometer 14 so as to read the signal provided by each light receiver 12. In particular, the optical mixer 13 has a plurality of input terminals, each of which is connected to the light receiver 12 of a corresponding one of the optical sensors 8a, 8b and 8c, and an output terminal connected to the input terminal of the spectrometer 14.

[0040] The light receiver 12 of each optical sensor 8 a , 8 b , 8 c is provided with an optical fiber designed to convey the light received by the light receiver 12 to the optical mixer 13 .

[0041] The spectrometer 14 is a known device that can provide a light intensity distribution as a function of the wavelength of the optical radiation it receives as input, i.e., it can measure the intensity of the optical radiation of the various wavelengths that compose it. Specifically, the spectrometer 14 divides the optical radiation spectrum into a plurality of very narrow bands centered on the respective wavelengths and returns a plurality of signals, each signal representing the light intensity corresponding to a certain band. For simplicity, each signal provided by the spectrometer 14 will be considered below as corresponding to a specific wavelength, i.e., each of the above-mentioned bands will be identified by the respective band center wavelength.

[0042] The control unit 17 is configured to activate the light emitters 11 of one optical sensor 8a, 8b, 8c at a time, disable the other light emitters 11, and read the signals provided by the corresponding light receivers 12 in order to measure the parameters of the organic liquid at one point of the blood circuit 2 at a time, i.e., where the optical sensor 8a, 8b, 8c is located. In other words, the optical sensor 8a, the optical sensor 8b, or the optical sensor 8c is used alone.

[0043] Preferably, the switching frequency of the measuring system 7, i.e. the frequency at which the acquisition system 9 sequentially switches the light emitters 11 of the plurality of optical sensors 8a, 8b and 8c on and off, is low enough to allow the control unit 17 to complete reading a given optical sensor 8a, 8b, 8c before switching off the corresponding light emitter 11 and moving on to the next optical sensor 8a, 8b, 8c.

[0044] It will be appreciated that the blood circuit 2 of the machine 1 may have a Figures 1 to 3 A greater or fewer number of optical sensors than the one shown in FIG. 1 , correspondingly increases or decreases the number of measuring points in the blood circuit 2 .

[0045] Preferably, the light emitter 11 of each optical sensor 8a, 8b, 8c is configured to emit white light, i.e. light having a wavelength throughout the visible spectrum. In other words, the light emitter 11 emits electromagnetic radiation having a plurality of wavelengths distributed throughout the visible spectrum. The spectrometer 14 is of a type suitable for operating within the entire visible spectrum. As previously mentioned, the spectrometer 14 provides a plurality of output signals, each output signal corresponding to the light intensity of a corresponding wavelength.

[0046] Finally, the control unit 17 implements a plurality of machine learning algorithms, each designed to receive as input all the signals from the spectrometer 14 and combine them in order to determine the value of a corresponding parameter of the organic liquid flowing in the ducts 5a, 5b, 5c associated with the optical sensors 8a, 8b, 8c.

[0047] In other words, the control unit 17 implements a plurality of machine learning algorithms specially developed and calibrated by a substantially known training process based on a series of measurements performed by the optical sensors 8a-8c with known results, so that it is possible to measure desired intrinsic parameters of the organic liquid flowing in the ducts 5a, 5b, 5c based on the plurality of output signals provided by the spectrometer 14. In particular, each machine learning algorithm is associated with a parameter of the organic liquid measured with a given optical sensor 8a, 8b, 8c in a given duct 5a, 5b, 5c through which the organic liquid flows, and the training process consists in performing a large number of measurements, for example at least one hundred measurements, by this optical sensor 8a, 8b, 8c providing predetermined values ​​for the parameters, and calibrating the machine learning algorithm so that the control unit 17 provides these predetermined values.

[0048] Preferably, each machine learning algorithm implemented in the control unit 17 comprises at least one artificial neural network 18, i.e. a mathematical model consisting of a plurality of nodes interconnected on one or more layers, which receives as input a plurality of values ​​and outputs a combination of all the input values ​​received. In more detail, each neural network 18 comprises a linear combiner which adds, based on a series of measurements with known results, a plurality of contribution values ​​obtained by multiplying the signal output from the spectrometer 14 with the respective weights previously obtained through an initial training process of the neural network 18.

[0049] The technical advantage of neural networks is that they can be used to simulate complex relationships between input and output that cannot be represented by analytical functions.

[0050] According to another embodiment of the control unit 17, the machine learning algorithm is, for example, a statistical calculation method or an adaptive data filtering method.

[0051] refer to Figure 1 In the embodiment of the invention, the control unit 17 is configured to measure the volume percentage of plasma extracted by the plasma filter 3 relative to the blood taken from the patient as a function of the signals provided by the optical sensors 8a and 8b. In other words, the control unit 17 is configured to implement a specially trained machine learning algorithm to calculate the blood concentration difference caused by the plasma extraction performed by the pump 15c based on the output signal provided by the spectrometer 14 when the light emitters 11 in the optical sensors 8a and 8b are turned on.

[0052] refer to Figure 1 In the embodiment of the invention, the control unit 17 is configured to measure the hemoglobin concentration in the plasma upstream of the fractionation filter 4, i.e. the hemoglobin concentration in the organic liquid in the conduit 5c, as a function of the signal provided by the optical sensor 8c. The control unit 17 is further configured to check whether the hemoglobin concentration measured in the plasma upstream of the fractionation filter 4 is above a predetermined threshold value, for example, above 1.5%.

[0053] A hemoglobin measurement value above the above exemplary threshold value indicates the presence of hemoglobin in the plasma upstream of the fractionation filter 4, which in turn means that hemolysis, i.e. the process of lysis of red blood cells accompanied by hemoglobin leakage, may be occurring. In other words, the control unit 17 is configured to implement a specially trained machine learning algorithm to check whether hemolysis is present in the blood circuit 2 based on the output signal provided by the spectrometer 14 when the light emitter 11 in the optical sensor 8c is turned on.

[0054] refer to Figure 1In the embodiment of the invention, the control unit 17 is configured to measure the number of platelets in the plasma upstream of the fractionation filter 4, i.e. the number of platelets in the organic liquid in the conduit 5c, as a function of the signal provided by the optical sensor 8c. In other words, the control unit 17 is configured to implement a specially trained machine learning algorithm 18 to measure the number of platelets in the organic liquid in the conduit 5c based on the output signal provided by the spectrometer 14 when the light emitter 11 in the optical sensor 8c is turned on.

[0055] The patient's blood filtration treatment ends with a phase in which the blood contained in the plasma filter 3 flows back to the patient. To implement the blood return phase, the catheter 5a is removed from the patient's corresponding venous access and connected to a saline solution capsule (not shown), the pump 15c remains closed and the pump 15a is started to feed saline solution to the plasma filter 3, so that the blood comes out of the plasma filter 3 and flows back to the patient through the catheter 5b. However, when all the blood has flowed out of the plasma filter 3 and before the saline solution enters the patient's venous access, the pump 15a must be stopped.

[0056] In order to automatically manage the phase of blood return to the patient, the control unit 17 is configured to measure the volume percentage of blood flowing out of the plasma filter 3 relative to the saline solution flowing into the plasma filter 3 as a function of the signals provided by the optical sensors 8a and 8b, and to generate a pump 15a stop event when the volume percentage of blood relative to the saline solution is below a predetermined threshold value (e.g., below 10%). In other words, the control unit 17 is configured to implement a specially trained machine learning algorithm to generate a pump 15a stop event when the volume percentage of blood relative to the saline solution measured based on the output signal provided by the spectrometer 14 is below a predetermined threshold value when the light emitters 11 in the optical sensors 8a and 8b are turned on.

[0057] Reference now Figure 2 In the embodiment of the invention, the control unit 17 is configured to: measure the hematocrit in the blood upstream of the dialyzer filter 23 at a first moment, i.e. the hematocrit in the organic liquid in the conduit 5a, as a function of the signal provided by the optical sensor 8a; measure the hematocrit in the blood downstream of the dialyzer filter 23 at a second moment other than the first moment, in particular a second moment after the first moment, i.e. the hematocrit in the organic liquid in the conduit 5b, as a function of the signal provided by the optical sensor 8b; determine the hematocrit reduction as a function of the hematocrit measured at the first moment and the second moment; and determine the liquid volume reduced by the dialyzer filter 23, i.e. the weight loss of the patient during the hemodialysis treatment, as a function of the hematocrit reduction. The time distance between the first moment and the second moment is at least equal to the time it takes for the blood to pass through the dialyzer filter 23, which is a few seconds.

[0058] Note that the measuring system 7 is also suitable for installation on any other type of machine or device for measuring parameters of other organic liquids (e.g., dialysate, fresh plasma or urine). To this end, the measuring system 7 comprises one or more optical sensors 8a-8c depending on the type of machine or device on which it is installed.

[0059] According to another embodiment of the measuring system 7 that is not shown, the measuring system 7 comprises a single optical sensor 8 a , the acquisition system 9 is devoid of the light mixer 13 , and the light receiver 12 is directly connected to the input of the spectrometer 14 .

[0060] The above-described measuring system 7 and the corresponding blood filtration machine 1 have numerous advantages.

[0061] Firstly, the measuring system 7 is simpler and cheaper if compared to known types of measuring systems in which each light emitter operates with a specific wavelength instead of with broad spectrum light (i.e. white light) like the light emitter 11. In other words, the measuring system 7 has only one type of optical sensor (i.e. a white light sensor) that can be used to measure any parameter, and the spectrometer 14 operating in the visible spectrum is cheaper than other spectrometers operating in other wavelengths, such as in the infrared range.

[0062] Secondly, the quality of the measurements obtained by the measuring system 7 does not depend on the quality of the reading window 10, so a greater opacity of the material of the reading window or of the ducts 5a, 5b, 5c does not degrade the quality of the measurements performed.

[0063] Furthermore, the combined use of a white light optical sensor and a machine learning algorithm implemented in the control unit 17 allows reproducible measurements of even very small values ​​of the parameter of interest, for which the prior art does not provide reliable measurements. In fact, the use of white light increases the information available for training the machine learning algorithm. This allows the measuring system 7 to detect small amounts of blood in the conduits of the blood circuit 2, thereby allowing the system to be further used, for example, to check for damage in the blood circuit 2.

[0064] Furthermore, even if the measuring system 7 of the above-described blood filtration machine 1 comprises a single acquisition system 9, ie a single spectrometer 14 and a single control unit 17, it can be provided with a large number of optical sensors, since the optical sensors are of the same type.

[0065] Thus, several parameters can be measured at the same measuring point thanks to the white light optical sensor and / or the same parameter can be measured at several measuring points within the blood circuit 2 without increasing the costs of the measuring system 7 .

[0066] Finally, the measuring system 7 can be used in any circuit comprising a conduit made of transparent material in which an organic liquid flows, to measure physical parameters of this organic liquid. To this end, it is sufficient to appropriately train the machine learning algorithm implemented in the control unit 17 .

Claims

1. A blood filtration machine, comprising a blood circuit (2) and a measuring system (7), wherein the blood circuit (2) comprises at least one filter (3, 4; 23) and a plurality of conduits (5a-5c) made of a transparent material and designed to flow organic liquids, and wherein the measuring system (7) comprises: At least one optical sensor (8a-8c), the at least one optical sensor (8a-8c) comprising a reading window (10) placed at a point of each of the plurality of conduits (5a-5c) so as to be able to see the organic liquid flowing in the conduits (5a-5c), a light transmitter (11) emitting light to the conduits (5a-5c) through the reading window (10), and a light receiver (12) receiving the light after the light passes through the conduits (5a-5c); and a signal acquisition and processing device (9), the signal acquisition and processing device (9) being used to read the signal provided by the light receiver (12) and measure at least one intrinsic parameter of the organic liquid in the conduits (5a-5c) as a function of the signal; The machine (1) is characterized in that the at least one optical sensor comprises a plurality of optical sensors (8a-8c), each optical sensor being arranged on a respective one of the plurality of conduits (5a-5c), and the signal acquisition and processing device (9) comprises a single spectrometer (14), an optical mixer (13) and a control unit (17), the optical mixer (13) comprising a plurality of input ends of the optical receiver (12) respectively connected to a corresponding one of the optical sensors (8a-8c) and an output end connected to an input end of the spectrometer (14), the control unit (17) being configured to activate the light emitter (11) of one optical sensor (8a-8c) at a time so as to measure one parameter of the organic liquid at a time.

2. A machine according to claim 1, wherein the at least one filter (3, 4) comprises a plasma filter (3), the plasma filter (3) being designed to separate blood taken from a patient into plasma and particulate fractions, and the plurality of conduits (5a-5c) comprises a first conduit (5a) for conveying blood taken from a patient to the plasma filter (3), and the plurality of optical sensors (8a-8c) comprises a first optical sensor (8a) placed at a point of the first conduit (5a).

3. A machine according to claim 1 or 2, wherein the at least one filter (3, 4) comprises a plasma filter (3), the plasma filter (3) being designed to separate blood taken from a patient into plasma and a particulate fraction, and the plurality of conduits (5a-5c) comprises a second conduit (5b) for receiving the particulate fraction from the plasma filter (3), and the plurality of optical sensors (8a-8c) comprises a second optical sensor (8b) placed at a point of the second conduit (5b).

4. A machine according to claim 2, wherein the plurality of ducts (5a-5c) includes a second duct (5b) for receiving the particulate fraction from the plasma filter (3), the plurality of optical sensors (8a-8c) includes a second optical sensor (8b) placed at a point of the second duct (5b), and the control unit (17) is configured to measure the volume percentage of plasma extracted through the plasma filter (3) relative to the blood taken from the patient as a function of the signals provided by the first optical sensor (8a) and the second optical sensor (8b).

5. The machine according to claim 2, wherein the plurality of conduits (5a-5c) comprises a second conduit (5b) for receiving the particulate fraction from the plasma filter (3), the plurality of optical sensors (8a-8c) comprises a second optical sensor (8b) placed at a point of the second conduit (5b), the first conduit (5a) is connectable to a saline solution container during a phase of blood return to the patient, and the control unit (17) is configured to measure the volume percentage of blood flowing out of the plasma filter (3) relative to the saline solution flowing into the plasma filter (3) during a phase of blood return to the patient as a function of the signals provided by the first optical sensor (8a) and the second optical sensor (8b), and to generate a blood return stop event when the volume percentage of blood relative to the saline solution is below a predetermined threshold.

6. The machine according to claim 1, wherein the at least one filter (3, 4) comprises a plasma filter (3) and a fractionation filter (4), the plasma filter (3) being designed to separate blood taken from a patient into plasma and a particulate fraction, the fractionation filter (4) being designed to separate the plasma into a high molecular weight fraction and a low molecular weight fraction, the plurality of conduits (5a-5c) comprising a third conduit (5c) for conveying the plasma from the plasma filter (3) to the fractionation filter (4), and the plurality of optical sensors (8a-8c) comprising a third optical sensor (8c) placed at a point of the third conduit (5c).

7. Machine according to claim 6, wherein said control unit (17) is configured to measure the hemoglobin concentration in the plasma upstream of said fractionation filter (4) as a function of the signal provided by said third optical sensor (8c).

8. Machine according to claim 6 or 7, wherein said control unit (17) is configured to measure the platelet count of the organic liquid upstream of said fractionation filter (4) as a function of the signal provided by said third optical sensor (8c).

9. A machine according to claim 1, wherein the at least one filter includes a dialyzer filter (23), the multiple conduits (5a-5c) include a first conduit (5a) for conveying blood taken from a patient to the dialyzer filter (23) and a second conduit (5b) for conveying blood processed by the dialyzer filter (23) to the patient, and the multiple optical sensors (8a-8c) include a first optical sensor (8a) placed at a point on the first conduit (5a) and a second optical sensor (8b) placed at a point on the second conduit (5b).

10. A machine according to claim 9, wherein the control unit (17) is configured to: measure the hematocrit in the blood upstream of the dialyzer filter (23) at a first moment in time as a function of the signal provided by the first optical sensor (8a); measure the hematocrit in the blood downstream of the dialyzer filter (23) at a second moment in time other than the first moment in time as a function of the signal provided by the second optical sensor (8b); determine a reduction in hematocrit as a function of the hematocrit measured at the first moment in time and at the second moment in time; and determine the volume of liquid reduced by the dialyzer filter (23) as a function of the reduction in hematocrit.

11. The machine according to claim 1, wherein the light emitter (11) is configured to emit white light and the spectrometer (14) operates in the entire visible spectrum.

12. A machine according to claim 1, wherein the spectrometer (14) is configured to provide a plurality of output signals, each output signal corresponding to the light intensity of a corresponding wavelength, and the control unit (17) is configured to implement a plurality of machine learning algorithms (18), each machine learning algorithm being designed to receive as input all the output signals of the spectrometer (14) and to combine them in order to determine the value of a corresponding parameter of the organic liquid in a given conduit (5a-5c).

13. A machine according to claim 12, wherein each machine learning algorithm comprises an artificial neural network (18) trained based on a series of measurements with known outcomes.

14. A machine according to claim 13, wherein each artificial neural network (18) includes a linear combiner designed to add a plurality of contribution values ​​obtained by multiplying the signal output from the spectrometer (14) by respective weights previously obtained through an initial training process of the neural network (18).

Images