Concentration calculation device, blood processing system, and concentration calculation method

Abstract

To provide a concentration calculation device that can highly accurately measure concentration of substances such as albumin without increase in the number of groups to be classified pointlessly.SOLUTION: A concentration calculation device 16 comprises: an irradiation unit 80 that irradiates a concentration calculation object with excitation light; a detection unit 81 that detects fluorescence emitting from the concentration calculation object irradiated by the irradiation unit 80; a classification unit 71 that classifies a fluorescence spectrum of the fluorescence detected by the detection unit 81 into a first group not required for a correction and a second group required for the correction on the basis of a first and second amounts of characteristic extracted from the fluorescence spectrum thereof; a correction unit 94 that corrects the fluorescence spectrum classified into the second group; and a concentration calculation unit 72 that calculates concentration of a substance included in the concentration calculation object on the basis of fluorescence intensity of a plurality of wavelengths in a prescribed wavelength range of the fluorescence spectrum classified into the first group, fluorescence intensity of a plurality of wavelengths in a prescribed wavelength range of a fluorescence spectrum after being corrected by the correction unit 94 and a specific calibration model.SELECTED DRAWING: Figure 8

Description

[Technical Field]

[0001] The present invention relates to a concentration calculation device, a blood processing system, and a concentration calculation method, and more particularly to a concentration calculation device, a blood processing system equipped therewith, and a concentration calculation method for calculating the concentration of a target substance (e.g., protein concentration or albumin concentration) in dialysis wastewater, which is a multi-component solution used in hemodialysis therapy and hemodiafiltration therapy. [Background technology]

[0002] Hemodialysis and hemodiafiltration, which are widely used treatments for patients with renal failure, are methods that are currently in use. In these treatments, unwanted components of the blood that have been introduced into the hollow fiber membrane of the dialyzer are flushed out through the hollow fiber membrane into fresh dialysate that has been introduced into the outside of the dialyzer, and are then removed together with the dialysate.

[0003] Incidentally, many patients with renal failure complain of various symptoms such as itching, irritability, and bone and joint pain. To improve such symptoms, aggressive removal of alpha-1 microglobulin (MG) is being performed. However, albumin, which is necessary for the body, is similar in size to α1-MG, so aggressive removal of α1-MG inevitably leads to leakage into the dialysis fluid. As a current method for managing albumin, it is common practice to measure the patient's serum albumin concentration once or twice a month to understand the albumin level, but this is not sufficient for dialysis treatment three times a week. Furthermore, testing for low molecular weight proteins such as α1-MG is essential for understanding the effectiveness of treatment, but the tests are expensive and cannot be performed frequently.

[0004] Given the circumstances described above, there is a need in dialysis treatment settings for a means of monitoring the concentrations of albumin and protein in dialysis effluent in real time, online, through a device connected to the tubing through which the effluent flows. Therefore, in recent years, a method has been proposed to calculate the concentrations of protein and albumin in dialysis effluent online and in real time by classifying predetermined elements that affect the calibration model for concentration calculation into one of several predetermined groups, and then calculating the concentration of the substance using a calibration model corresponding to the classified group of predetermined elements (see Patent Document 1). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] International Publication No. 2020 / 262534 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] However, in order to improve the accuracy of the measured concentration using the conventional method described in Patent Document 1, it was necessary to increase the number of groups to be classified. As a result, the number of groups increased with each additional case, which became apparent as a disadvantage in practical application.

[0007] This invention has been made in view of the above circumstances, and aims to provide a concentration calculation device that can measure (calculate) the concentration of substances such as albumin with high accuracy without unnecessarily increasing the number of groups to be classified. [Means for solving the problem]

[0008] To achieve the above objective, the inventors, after diligent study, discovered that the above problem can be solved by extracting two feature quantities from the fluorescence spectra of each case obtained in advance, designating these feature quantities as the first and second indicators, plotting each data with the first indicator on the vertical axis and the second indicator on the horizontal axis, obtaining a linear correlation from the plotted data, classifying and correcting the fluorescence spectra based on the degree of deviation from the linear correlation, and calculating the concentration of the substance using the uncorrected fluorescence spectrum if no correction is made, or the corrected fluorescence spectrum if correction is made, together with a calibration model, thereby completing the present invention.

[0009] In other words, the present invention includes the following embodiments. [1] An irradiation unit that irradiates the object to be used for concentration calculation with excitation light, A detection unit that detects fluorescence emitted from an object whose concentration is to be calculated after being irradiated by an irradiation unit, A classification unit that classifies fluorescence spectra into a first group that does not require correction and a second group that requires correction, based on a first and second characteristic quantity extracted from the fluorescence spectrum detected by the detection unit, A correction unit for correcting fluorescence spectra classified into the second group, A concentration calculation unit calculates the concentration of a substance contained in a substance for which concentration calculation is to be performed, based on the fluorescence intensities at multiple wavelengths within a predetermined wavelength range of the fluorescence spectrum classified into the first group, the fluorescence intensities at multiple wavelengths within a predetermined wavelength range of the fluorescence spectrum corrected by the correction unit, and a specific calibration model. A concentration calculation device equipped with the following features. [2] The standardized intensity is obtained by dividing the fluorescence intensity at each wavelength by the maximum fluorescence intensity value of the fluorescence spectrum detected by the detection unit. The spectrum consisting of the standardized intensities of each wavelength is defined as the standardized spectrum. In the standardized spectra obtained multiple times, the standardized intensity near the wavelength where the change in standardized intensity between each standardized spectrum is largest is used as the first feature quantity. The maximum fluorescence intensity value, its reciprocal, or the value obtained by transforming the maximum fluorescence intensity value using a specific formula is used as the second feature quantity. The classification unit calculates a linear correlation equation from each data plotted with the first feature quantity on the vertical axis and the second feature quantity on the horizontal axis, and classifies data that deviates from a predetermined range from the linear correlation equation into a second group. The concentration calculation device according to [1]. [3] Let the maximum fluorescence intensity value be F MAX and let the upper limit intensity value that the spectroscope included in the detection unit can detect at most be F DUL In this case, The second feature quantity I2 is given by the following formula (1) I2 = (F DUL - F MAX ) / F DUL … (1) The concentration calculation device according to [2], which is calculated by. [4] The data that deviates from a predetermined range from the linear correlation equation is data that exceeds ±3% to ±15% from the linear correlation equation. The concentration calculation device according to [2] or [3]. [5] Let the first feature quantity and the second feature quantity be I1 and I2 respectively, the fluorescence intensity before correction be F, the maximum fluorescence intensity value be F MAX and let the upper limit intensity value that the spectroscope included in the detection unit can detect at most be F DUL In this case, when the slope and intercept of the linear correlation equation are a and b respectively, and x is a constant between -0.15 and 0.15, The fluorescence intensity F C in the fluorescence spectrum after correction by the correction unit is given by the following formula (2) F C = F × {F DUL - (I2 - |[I1 - {(I2 × a + b) × (1 + x)}]| / a) × F DUL} / F MAX … (2) The concentration calculation device according to any one of [2] to [4], which is calculated by. [6] Let the first feature quantity and the second feature quantity be I1 and I2 respectively, the fluorescence intensity before correction be F, the maximum fluorescence intensity value be F MAX and let the upper limit intensity value that the spectroscope included in the detection unit can detect at most be F DUL In this case, when the slope and intercept of the linear correlation equation are a and b respectively, and x is a constant between -0.15 and 0.15, Fluorescence intensity F in the fluorescence spectrum after correction by the correction unit C This is given by the following equation (3) F C =F×{F DUL -(I2-[I1―{(I2×a+b)×(1+x)}] / a)×F DUL} / F MAX … (3) A concentration calculation device as described in any of [2] to [4], which is calculated by [the specified method]. A blood processing system equipped with a concentration calculation device as described in any of [7], [1], or [6]. [8] An irradiation step in which excitation light is irradiated onto the object to be used for concentration calculation, A detection step that detects fluorescence emitted from the object whose concentration is to be calculated after irradiation in the irradiation step, A classification step in which fluorescence spectra are classified into a first group that does not require correction and a second group that requires correction, based on a first and second characteristic quantity extracted from the fluorescence spectra of the fluorescence detected in the detection step, A correction step to correct the fluorescence spectra classified into the second group, A calculation step for calculating the concentration of a substance contained in a substance for which concentration calculation is performed, based on the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum classified into the first group, the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum after correction by the correction step, and a specific calibration model. A method for calculating concentration, including the method described above. [Effects of the Invention]

[0010] According to the present invention, it is possible to provide a concentration calculation device that can measure (calculate) the concentration of substances such as albumin with high accuracy without unnecessarily increasing the number of groups to be classified. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing the configuration of a dialysis system according to an embodiment of the present invention. [Figure 2] This is an explanatory diagram showing an example of the configuration of an optical monitor. [Figure 3]This figure shows the excitation wavelengths at which albumin exhibits subpeaks in its fluorescence properties. [Figure 4] This is an explanatory diagram for determining the standardized spectrum. [Figure 5] This is a diagram explaining the first indicator. [Figure 6] This is an example diagram showing the correlation between the first and second indicators. [Figure 7] This is an example diagram showing how to classify data into two categories based on deviations from linear correlation: those exceeding ±10% and those within ±10%. [Figure 8] This is a block diagram showing the functional configuration of a concentration calculation device according to an embodiment of the present invention. [Figure 9] This is a block diagram of the numerical calculation unit. [Figure 10] This is a flowchart illustrating the concentration calculation method according to an embodiment of the present invention. [Figure 11] This figure shows an example of a fluorescence spectrum. [Figure 12] This is a diagram illustrating the correction method. [Figure 13] This diagram illustrates the uncorrected and corrected spectra. [Figure 14] This is a diagram showing the classification of the examples. [Modes for carrying out the invention]

[0012] Embodiments of the present invention will be described below with reference to the figures. Unless otherwise specified, the positional relationships in the drawings, such as top, bottom, left, and right, are based on the positional relationships shown in the drawings. The dimensional ratios in the drawings are not limited to those shown. Furthermore, the following embodiments are illustrative examples for explaining the present invention and are not intended to limit the present invention to these embodiments only. In addition, the present invention can be modified in various ways without departing from its essence.

[0013] <Dialysis System> Figure 1 is a schematic diagram showing the configuration of a dialysis system 1 as a blood processing system equipped with a concentration calculation device according to this embodiment. As shown in Figure 1, the dialysis system 1 according to this embodiment includes a dialyzer 10, a blood circuit 11, a dialysate circuit 12, a drainage circuit 13, a fluid replacement circuit 14, a control device 15, a concentration calculation device 16, etc.

[0014] The dialyzer 10 is, for example, a hollow fiber module incorporating hollow fiber membranes, which can separate unwanted components from blood. The dialyzer 10 has a cylindrical container 20, and inside the cylindrical container 20, a number of hollow fiber membranes 21 are arranged along its longitudinal direction. The hollow fiber membranes 21 can separate unwanted components from blood. The top and bottom of the cylindrical container 20 are provided with inlets 22 and outlets 23 that lead to the internal space (blood side) of the hollow fiber membranes 21, and the sides of the cylindrical container 20 are provided with two inlets and outlets 24 and 25 that lead to the external space (dialysis fluid side) of the hollow fiber membranes 21.

[0015] The blood circuit 11 includes, for example, a blood withdrawal line 31 connecting the blood withdrawal unit 30 to the dialyzer 10, and a blood return line 33 connecting the dialyzer 10 to the blood return unit 32. The blood withdrawal line 31 and the blood return line 33 are mainly made of flexible tubing. The blood withdrawal line 31 is connected to the inlet 22 of the dialyzer 10, and the blood return line 33 is connected to the outlet 23 of the dialyzer 10.

[0016] The blood withdrawal line 31 is equipped with, for example, a blood pump 40. A drip chamber 41 is also connected to the blood withdrawal line 31. However, the drip chamber 41 may not always be present.

[0017] The dialysate circuit 12 is connected from a dialysate supply source (not shown) to the inlet / outlet 25 of the dialyzer 10. The drainage circuit 13 is connected from the inlet / outlet 24 of the dialyzer 10 to the drainage section (not shown). The dialysate circuit 12 and the drainage circuit 13 are equipped with supply and drainage pumps (not shown) that supply dialysate to the dialyzer 10 through the dialysate circuit 12 and drain dialysate from the dialyzer 10 through the drainage circuit 13.

[0018] The fluid replacement circuit 14 is connected, for example, from the dialysate circuit 12 to the drip chamber 41 (blood circuit 11). If there is no drip chamber 41, the fluid replacement circuit 14 is connected directly to the blood withdrawal line 31. The fluid replacement circuit 14 is equipped with a fluid replacement pump 50.

[0019] The control device 15 is, for example, a computer equipped with a memory unit for storing various programs and data, and a CPU for executing various programs. By executing the programs stored in the memory unit with the CPU, the control device can control the operation of the blood pump 40 and the fluid replacement pump 50 to perform dialysis treatment. Communication between the control device 15 and the various devices (blood pump 40 and fluid replacement pump 50) may be performed via a wired connection such as a communication cable, or wirelessly via Bluetooth®.

[0020] In dialysis treatment, in the blood circuit 11, the patient's blood is sent from the blood withdrawal section 30 to the space inside the hollow fiber membrane 21 of the dialyzer 10, passes through the dialyzer 10, and is returned to the patient from the blood return section 32. At this time, the dialysate is sent to the space outside the hollow fiber membrane 21 of the dialyzer 10 through the dialysate circuit 12, and is then drained through the drainage circuit 13. In the dialyzer 10, mainly unwanted components in the blood flowing through the space inside the hollow fiber membrane 21 flow out through the hollow fiber membrane 21 to the space outside the tube (dialysate side) and are discharged together with the dialysate. Replacement fluid (dialysate) from the dialysate circuit 12 is supplied to the blood circuit 11 through the replacement fluid circuit 14, and predetermined components are replenished in the blood. Note that there are hemodiafiltration, in which replacement fluid is replenished, and hemodialysis, in which replacement fluid is not replenished.

[0021] <Concentration calculation device> As shown in Figure 1, the concentration calculation device 16 according to this embodiment includes an optical monitor 70 that irradiates the dialysis effluent, which is a multi-component solution to be used for concentration calculation, with excitation light and detects fluorescence generated from the dialysis effluent; a classification unit 71 that classifies the fluorescence spectrum acquired by the optical monitor 70 into multiple groups; a correction unit 94 (not shown in Figure 1) that corrects the fluorescence spectra of the two groups that require correction; a concentration calculation unit 72 that calculates the concentration of albumin as a substance contained in the dialysis effluent from a specific calibration model based on the fluorescence intensity of multiple wavelengths in a predetermined wavelength range, whether the fluorescence spectrum is uncorrected if no correction is needed or corrected if correction is needed; a display unit 73; and an input unit 74. The correction unit 94 is mounted inside the concentration calculation unit 72 as shown in Figure 8.

[0022] Furthermore, communication between the optical monitor 70, classification unit 71, density calculation unit 72, display unit 73, and input unit 74 may be performed via a wired connection such as a communication cable, or wirelessly such as Bluetooth®.

[0023] The optical monitor 70 includes, for example, an irradiation unit 80 that irradiates the dialysis effluent in the piping of the drainage circuit 13 with excitation light, and a detection unit 81 that includes a spectrometer for detecting fluorescence generated from the dialysis effluent.

[0024] The irradiation unit 80 can irradiate light with a wavelength between 300 nm and 400 nm that excites the fluorescence that forms the albumin subpeak shown in Figure 3. The light source of the irradiation unit 80 is not particularly limited, but examples include halogen lamps, xenon lamps, deuterium lamps, LEDs, etc.

[0025] The detection unit 81 detects fluorescence generated from the dialysis drain fluid using a spectrometer to obtain a fluorescence spectrum. The detection unit 81 only needs to be capable of detecting fluorescence in the wavelength range of 310 nm to 850 nm. Furthermore, the irradiation unit 80 and the detection unit 81 are arranged, for example, perpendicular to each other with respect to the piping of the drain circuit 13, and fluorescence is acquired at an angle perpendicular to the excitation light. However, the arrangement of the irradiation unit 80 and the detection unit 81 and the shape of the drain circuit 13 are not limited to these.

[0026] The classification unit 71 is, for example, a computer equipped with a memory for storing various programs and data, and a CPU for executing various programs, and it classifies the fluorescence spectrum detected by the detection unit 81 into multiple groups. In this embodiment, the classification unit 71 classifies the multiple groups as follows.

[0027] First, as shown in Figure 4 (left), the standardized intensity is obtained by dividing the fluorescence intensity at each wavelength by the maximum fluorescence intensity of the obtained fluorescence spectrum. The spectrum consisting of the standardized intensities at each wavelength is then standardized as shown in Figure 4 (right). As shown in Figure 5, the standardized intensity near the wavelength with the largest change in standardized intensity between each standardized spectrum (for example, 460 nm) is taken as the first index (first feature) in the multiple standardized spectra obtained. The maximum fluorescence intensity value, or the value obtained by converting the maximum fluorescence intensity value using a specific conversion formula, is taken as the second index (second feature). The reciprocal of the maximum fluorescence intensity of the fluorescence spectrum may also be used as the second index.

[0028] Next, as shown in Figure 6, the first indicator I1 is plotted on the vertical axis and the second indicator I2 (using the transformed value of equation (1) below as an example) is plotted on the horizontal axis to obtain a linear correlation equation. I2=(F DUL -F MAX ) / F DUL … (1) Here, F MAX This is the maximum fluorescence intensity value. Also, F DUL This is the upper limit intensity value that the spectrometer included in the detection unit 81 can detect (spectrometer detection upper limit intensity value), and is given as a constant.

[0029] Furthermore, as shown in Figure 7, the data is classified into two groups based on the linear correlation equation: the first group (data within ±x%) and the second group (data exceeding ±x%). The range of x is preferably 3 to 15. Note that the linear correlation equation obtained from the correlation between the first and second indicators can be the one obtained by excluding points that fall beyond ±y% from the correlation relationship. The range of y is preferably -0.2 to 0.2, and more preferably -0.15 to 0.15. The correction unit 94 (see Figure 8) corrects the data of the second group that falls beyond ±x%% from the linear correlation equation.

[0030] The classification unit 71 does not necessarily have to be built into the main body of the concentration calculation device 16. The detection unit 81 can also transmit the detected fluorescence spectrum data, classify it on the cloud, and then communicate it to the concentration calculation unit 72.

[0031] Figure 8 is a block diagram showing the functional configuration of the concentration calculation unit 72, etc. The concentration calculation unit 72 is, for example, a computer equipped with a memory for storing various programs and data, and a CPU for executing various programs. The concentration calculation unit 72 includes a concentration calculation unit 90, a memory unit 91, a numerical calculation unit 92, a calibration model creation unit 93, and a correction unit 94 for correcting the fluorescence spectra of the second group of data that require correction, which have been classified from the linear correlation equation.

[0032] The concentration calculation unit 90 acquires the fluorescence spectrum of the dialysis fluid from the optical monitor 70 in real time during dialysis treatment and calculates the concentration of albumin as a substance contained in the dialysis fluid from a calibration model based on the fluorescence intensities of multiple wavelengths within a predetermined wavelength range of the fluorescence spectrum of the first group, which does not require correction, from among the multiple groups classified by the classification unit 71. The concentration calculation unit 90 also calculates the concentration of albumin as a substance contained in the dialysis fluid from a calibration model based on the fluorescence intensities of multiple wavelengths within a predetermined wavelength range of the corrected fluorescence spectrum, which is corrected by the correction unit 94 for the second group that requires correction. In other words, for fluorescence spectra that do not require correction, the concentration calculation unit 90 calculates the albumin concentration based on the fluorescence intensity of the uncorrected fluorescence spectrum and the calibration model, while for fluorescence spectra that require correction, it calculates the albumin concentration based on the fluorescence intensity of the corrected fluorescence spectrum and the calibration model.

[0033] The memory unit 91 stores the albumin concentration obtained by the concentration calculation unit 90. The numerical calculation unit 92 takes the albumin concentration value stored in the memory unit 91 and calculates various numerical values ​​in real time during dialysis treatment. As shown in Figure 9, the numerical calculation unit 92 includes, for example, an integral value calculation unit 100 that calculates the integral value of the product of albumin concentration and dialysis drainage flow rate since the start of dialysis treatment, a rate of change calculation unit 101 that calculates the rate of change of albumin concentration, a difference calculation unit 102 that calculates the difference between albumin concentration and a predetermined value, and an estimated value calculation unit 103 that calculates an estimated total amount of albumin in dialysis drainage from the start of dialysis treatment to the end of dialysis treatment.

[0034] The calibration model creation unit 93 creates a calibration model by performing multivariate analysis on a fluorescence spectrum that does not require correction and a corrected fluorescence spectrum. The multivariate analysis for creating the calibration model may be any of the following: partial least squares (PLS) regression analysis, principal component regression analysis, multiple regression analysis, support vector machine regression analysis, or machine learning analysis. An example of a calibration model is shown below. The calibration model can be expressed as an equation that calculates the concentration C by multiplying each fluorescence intensity at each wavelength of the obtained fluorescence spectrum by a coefficient and taking the sum of these coefficients. The calibration model can be expressed as, for example, equation (4). C=a1×F1+a2×F2+…+an×Fn+K…(4) Here, an is a coefficient, Fn is the fluorescence intensity, and K is a constant. The subscript n is a natural number obtained by numbering each wavelength in the fluorescence spectrum from shortest to longest (for example, if the wavelengths are 300 nm, 310 nm, 320 nm, ..., 400 nm, then n = 1, 2, 3, ..., 11).

[0035] The calibration model created in the calibration model creation unit 93 is stored in, for example, the storage unit 91 and used as a parameter in the calculation process in the concentration calculation unit 90 and the numerical calculation unit 92. Alternatively, instead of creating a calibration model in the calibration model creation unit 93, a pre-set calibration model can be stored in the storage unit 91 and used in the calculation process in the concentration calculation unit 90 and the numerical calculation unit 92.

[0036] The display unit 73 shown in Figure 8 is, for example, a panel display, which displays the albumin concentration calculated by the concentration calculation unit 90, various numerical values ​​obtained by the numerical calculation unit 92, and various warnings. Warnings are issued, for example, when the albumin concentration calculated by the concentration calculation unit 90 exceeds a predetermined range, when the integral value of the albumin leakage amount calculated by the integral value calculation unit 100 exceeds a predetermined range, when the rate of change of the albumin concentration calculated by the rate of change calculation unit 101 exceeds a predetermined range, when the difference between the albumin concentration calculated by the difference calculation unit 102 and its threshold exceeds a predetermined range, when the estimated total amount of albumin per treatment calculated by the estimated value calculation unit 103 exceeds a predetermined range, etc.

[0037] The input unit 74 shown in Figure 8 has the function of receiving information from an external source, in addition to the information necessary for classification by the classification unit 71, as well as information necessary for albumin concentration calculation by the concentration calculation unit 72, creation of calibration models, and various numerical calculations. Note that the control device 15, classification unit 71, concentration calculation unit 72, display unit 73, and input unit 74 may be implemented by the same computer.

[0038] <Operation of the concentration calculation device> Next, the operation of the concentration calculation device 16 according to this embodiment will be described. This operation will perform the concentration calculation method shown in the flowchart of Figure 10.

[0039] The concentration calculation device 16 calculates the albumin concentration in the dialysis effluent of the effluent circuit 13 during dialysis treatment in real time, either continuously or intermittently. The dialysis effluent is a continuous flow with a flow rate of 10 mL / min to 1000 mL / min.

[0040] Specifically, first, the irradiation unit 80 of the optical monitor 70 irradiates the drainage circuit 13 with excitation light (irradiation step: S1), and the detection unit 81 detects the fluorescence generated from the dialysis drainage using a spectrometer (detection step: S2). The excitation light irradiated from the irradiation unit 80 includes light with a wavelength between 300 nm and 400 nm that excites the fluorescence that becomes the subpeak of albumin.

[0041] As shown in Figure 11, the fluorescence spectrum detected by the detection unit 81 is, for example, in the wavelength range of 310 nm to 850 nm. The detection unit 81 adjusts the parameters related to the measurement based on the value of the maximum intensity between 380 nm and 480 nm in the wavelength range of the obtained fluorescence spectrum, so that the measurement is performed within the upper limit of the spectrometer of the detection unit 81.

[0042] Next, the classification unit 71 classifies the fluorescence spectra in the following procedure (classification step: S3).

[0043] The fluorescence spectrum detected by the detection unit 81 (see Figure 4 (left)) is standardized by dividing the fluorescence intensity at each wavelength by the maximum fluorescence intensity value of the spectrum, and the spectrum consisting of the standardized intensities at each wavelength is defined as the standardized spectrum (see Figure 4 (right)). Then, among the standardized spectra acquired multiple times, the standardized intensity near the wavelength with the largest change in standardized intensity between each standardized spectrum (in this case, 460 nm) is defined as the first index (first feature) (see Figure 5). In addition, the maximum fluorescence intensity value, or the value obtained by converting the maximum fluorescence intensity value using a specific conversion formula, is defined as the second index (second feature).

[0044] Next, the first indicator I1 is plotted on the vertical axis and the second indicator I2 on the horizontal axis to obtain a linear correlation equation (see Figure 6). The second indicator I2 uses the transformed value from equation (1) already explained. The data is then classified into two groups: the first group, which is within ±x% of the linear correlation equation, and the second group, which is outside the linear correlation equation by more than ±x%. The correction unit 94 (see Figure 8) corrects the data in the second group, which is outside the linear correlation equation by more than ±x%%.

[0045] The concentration calculation unit 90 of the concentration calculation unit 72 acquires the fluorescence spectrum of the dialysis effluent from the optical monitor 70 in real time during dialysis treatment. Among the multiple groups classified by the classification unit 71, the concentration calculation unit 90 calculates the albumin concentration for fluorescence spectra that do not require correction based on the fluorescence intensity of the uncorrected fluorescence spectrum and the calibration model, while for fluorescence spectra that require correction, it calculates the albumin concentration based on the fluorescence intensity of the corrected fluorescence spectrum and the calibration model.

[0046] For example, if the fluorescence spectrum detected by the detection unit 81 is classified into the second group requiring correction, the correction unit 94 corrects the fluorescence spectrum (correction step: S4), and the concentration calculation unit 90 calculates the albumin concentration from the corrected fluorescence spectrum using the calibration model stored in the storage unit 91. If the fluorescence spectrum detected by the detection unit 81 is classified into the first group that does not require correction, the concentration calculation unit 90 calculates the albumin concentration from the uncorrected fluorescence spectrum using the calibration model stored in the storage unit 91 (calculation step: S5).

[0047] The concentration calculation unit 90 calculates the albumin concentration in the dialysis drain fluid in real time, either continuously or intermittently, during dialysis treatment. The albumin concentration calculated by the concentration calculation unit 90 is stored in the storage unit 91 each time. This albumin concentration is displayed, for example, in real time on the display unit 73.

[0048] Here, the first and second indicators are I1 and I2, respectively, the uncorrected fluorescence intensity is F, and the maximum fluorescence intensity value is F. MAX The upper limit intensity value that the spectrometer included in the detection unit 81 can detect is F DUL Assuming that the slope and intercept of the linear correlation equation are a and b, respectively, and x is a constant between -0.15 and 0.15, the fluorescence intensity F in the fluorescence spectrum corrected by the correction unit 94 is... C This is calculated by equation (2) or (3) below. F C =F×{F DUL-(I2-|[I1―{(I2×a+b)×(1+x)}]| / a)×F DUL} / F MAX … (2) F C =F×{F DUL -(I2-[I1―{(I2×a+b)×(1+x)}] / a)×F DUL} / F MAX … (3)

[0049] The significance of the above correction formulas ((2) and (3)) will be explained. As shown in Figure 12, when data 1 is corrected, its coordinates become as follows. (Coordinates of Data 1) = (Second indicator of Data 1, First indicator of Data 1) ... (5) Next, the coordinates of point P on the linear correlation equation for the second indicator, which is the same as in Data 1, are expressed as follows. (Coordinates of point P) = (Data 1 second index, Corrected Data 1 first index) ... (6)

[0050] If the value of the first index of Data 1 deviates from the linear correlation equation, it is assumed that proteins such as albumin were oxidized more than the first index on the linear correlation equation (= corrected Data 1 first index). If we assume that the amount by which the first index of Data 1 deviates from the linear correlation equation is due to albumin not being oxidized and small molecular weight substances such as uric acid being oxidized instead, then the second index at the intersection of the line represented by "y = corrected Data 1 first index" and the line passing through the points of Data 1 with the same slope as the linear correlation equation (dotted line in Figure 12) becomes the corrected Data 1 second index.

[0051] When formula (1) is used as the second indicator, the corrected maximum fluorescence intensity value is obtained from the corrected second indicator. As shown in Figure 13, the corrected maximum fluorescence intensity value is multiplied by the fluorescence intensity of each wavelength obtained for each dialysis time (left) by the ratio of the corrected maximum fluorescence intensity value to the uncorrected maximum fluorescence intensity value obtained at the beginning of dialysis treatment = correction ratio R (= corrected maximum fluorescence intensity value / uncorrected maximum fluorescence intensity value), thereby obtaining the corrected fluorescence intensity and the corrected spectrum (right).

[0052] The reason there are two correction formulas, (2) and (3), is that the deviation from the linear correlation formula is either taken as an absolute value or as a difference (including negative values). Here, the fluorescence spectrum obtained during the remaining dialysis time was corrected with the correction ratio R determined at the beginning of the dialysis treatment, but it is also acceptable to determine the correction ratio R each time a fluorescence spectrum is obtained.

[0053] The calibration model is created in the calibration model creation unit 93 before the start of dialysis treatment for the case in which the concentration of the target substance is calculated, and is stored in the memory unit 91.

[0054] The numerical calculation unit 92 retrieves multiple albumin concentrations for each time point calculated by the concentration calculation unit 90 from the storage unit 91 and calculates various numerical values. The integral value calculation unit 100 calculates the integral value of the product of albumin concentration and dialysis drainage flow rate over time since the start of dialysis treatment, the rate of change calculation unit 101 calculates the rate of change of albumin concentration, the difference calculation unit 102 calculates the difference between albumin concentration and a predetermined value, and the estimated value calculation unit 103 calculates an estimated total amount of albumin in the dialysis drainage from the start to the end of dialysis treatment. These values ​​are displayed, for example, on the display unit 73. In addition, if the albumin concentration, the integral value of albumin leakage, the rate of change of albumin concentration, the difference between albumin concentration and the threshold, or the estimated amount of albumin leakage for one treatment is not within the appropriate range, a warning will be displayed on the display unit 73.

[0055] <Effects and Effects> In the concentration calculation device 16 according to the embodiment described above, the fluorescence spectrum of the fluorescence emitted from the object to be calculated is classified into a first group that does not require correction and a second group that requires correction, based on specific characteristics. For the first group, the concentration of the substance contained in the object to be calculated is calculated based on the fluorescence intensity at multiple wavelengths within a predetermined wavelength range of the fluorescence spectrum and a specific calibration model. For the second group, the concentration of the substance contained in the object to be calculated is calculated based on the fluorescence intensity at multiple wavelengths within a predetermined wavelength range of the corrected fluorescence spectrum and the calibration model. In other words, with this device, it is only necessary to classify the fluorescence spectrum of the fluorescence emitted from the object to be calculated into two groups depending on whether correction is required, and for the group that requires correction, the concentration of the substance contained in the object to be calculated can be calculated with high accuracy using the corrected fluorescence spectrum. Therefore, it is possible to measure (calculate) the concentration of substances such as albumin with high accuracy without unnecessarily increasing the number of groups to be classified.

[0056] Furthermore, the group classification in the concentration calculation device 16 according to the embodiment described above is based on the characteristic quantities that cause fluorescence in the fluorescence spectrum, so it is possible to calculate the albumin concentration with high accuracy. In other words, the first indicator (first characteristic quantity) is based on characteristics mainly related to proteins, and the second indicator (second characteristic quantity) is based on characteristics mainly related to low molecular weight substances other than proteins, such as uric acid.

[0057] Furthermore, the concentration calculation device 16 according to the embodiment described above includes a calibration model creation unit 93 that creates a calibration model by performing multivariate analysis on a fluorescence spectrum that does not require correction and a corrected fluorescence spectrum, thereby enabling the easy creation of a highly reliable calibration model. The multivariate analysis can be any of the following: partial least squares (PLS) regression analysis, principal component regression analysis, multiple regression analysis, support vector machine regression analysis, or machine learning analysis, allowing for the easy and accurate creation of a calibration model.

[0058] Furthermore, the concentration calculation device 16 according to the embodiment described above includes a storage unit 91 that stores multiple albumin concentrations calculated at multiple time points after the start of dialysis treatment. For example, various numerical calculations can be performed using the calculated albumin concentrations to analyze the status of dialysis treatment.

[0059] Furthermore, the concentration calculation device 16 according to the embodiment described above is equipped with a display unit 73 that displays information regarding the albumin concentration calculated by the concentration calculation unit 72, so that the user can understand the albumin concentration in real time during dialysis treatment.

[0060] Preferred embodiments of the present invention have been described above with reference to the attached drawings, but the present invention is not limited to these examples. It will be clear to those skilled in the art that various modifications or alterations can be conceived within the scope of the ideas described in the claims, and these will naturally also fall within the technical scope of the present invention.

[0061] For example, in the above embodiment, the substance to be irradiated with excitation light for concentration calculation was dialysis effluent (protein solution) containing albumin, and the concentration calculation device 16 calculated the concentration of albumin contained in the dialysis effluent. However, the present invention can also be applied to calculating the concentration of substances other than albumin (e.g., proteins) in dialysis effluent. Furthermore, the present invention may also be applied to irradiating a substance other than dialysis effluent with excitation light and calculating the substances contained in that substance.

[0062] Furthermore, the configuration of the dialysis system 1 is not limited to one that performs hemodialysis filtration as in the above embodiment. For example, when performing hemodialysis in the dialysis system 1, the fluid replacement pump 50 and the fluid replacement circuit 14 do not need to be used. The fluid replacement circuit 14 may be connected to the blood return line 33 instead of the blood withdrawal line 31. Moreover, the present invention is not limited to dialysis systems but can also be applied to other blood processing systems. For example, the present invention can be applied to blood processing systems that perform plasma exchange therapy, leukocyte apheresis, continuous slow hemofiltration therapy, and the like.

[0063] Next, embodiments of the present invention will be described.

[0064] <Example 1> Online dialysis filtration therapy was performed on dialysis patient A, who had a 13-year history of dialysis, using an ABH-22PA filter (manufactured by Asahi Kasei Medical Co., Ltd.) with a pre-replacement fluid volume of 60 L. Figure 14 shows the group classification used to calculate the albumin concentration. This group classification was obtained by determining the first index (first characteristic) and second index (second characteristic) from the fluorescence spectrum obtained in advance at 3 minutes after the start of dialysis, and plotting the first index on the vertical axis and the second index on the horizontal axis.

[0065] For linear correlation, a provisional linear correlation equation was obtained by removing data that deviated significantly from the correlation. Furthermore, data exceeding ±10% from this provisional linear correlation equation were removed, and data within ±10% were included to obtain the final linear correlation equation. Data were then classified into two groups: those deviating within ±10% of this linear correlation equation (first group) and those deviating more than ±10% (second group), resulting in Figure 14.

[0066] Then, the group with a deviation of more than ±10% from the linear correlation equation (the second group) was corrected, and combined with the group with a deviation of within ±10% from the linear correlation equation (the first group), the number of groups was reduced to one. The corresponding calibration model was then obtained by performing PLS regression analysis. In other words, in this embodiment, there is one group and one calibration model.

[0067] The data in the square plots in Figure 14 represent the first and second indicators obtained from the 3-minute fluorescence spectra of dialysis patient A. Since the deviation from the linear correlation equation was within ±10%, the albumin concentration for each dialysis time was calculated from the uncorrected spectrum and calibration model. The leakage amount was predicted as the total value obtained by integrating the product of the concentration and the dialysis drainage flow rate over the dialysis treatment time and compared with the measured value.

[0068] Table 1 shows the results of comparing the albumin leakage amount verified with the actual value using data from the same case that was not included in the data used to create the calibration model (i.e., validation data from the same case as the data used to create the calibration model), and the results of comparing the predicted albumin leakage amount with the actual value for dialysis patient A.

[0069] <Example 2> In this example, the group with a deviation of more than ±5% from the linear correlation formula (the second group) was corrected, and the number of groups was reduced to one by combining it with the group with a deviation of within ±5% from the linear correlation formula (the first group). One calibration model corresponding to this group was adopted, and the comparison results between leakage amount and measured value were obtained in the same manner as in Example 1. The comparison results are shown in Table 1.

[0070] <Example 3> In this example, the group with a deviation of more than ±3% from the linear correlation formula (the second group) was corrected, and the number of groups was reduced to one by combining it with the group with a deviation of within ±3% from the linear correlation formula (the first group). One calibration model corresponding to this group was adopted, and the comparison results between leakage amount and measured value were obtained in the same manner as in Example 1. The comparison results are shown in Table 1.

[0071] <Example 4> In this example, the group with a deviation of more than ±0% from the linear correlation formula (the second group) was corrected, and the number of groups was reduced to one by combining it with the group with a deviation of within ±0% from the linear correlation formula (the first group). One calibration model was adopted to correspond to this group, and the comparison results between leakage amount and measured value were obtained in the same manner as in Example 1 (in this example, virtually all data belong to the "second group"). The comparison results are shown in Table 1.

[0072] <Comparative Example 1> Calibration models were pre-determined by performing PLS regression analysis for both the group with a deviation of ±10% from the linear correlation equation (first group) and the group with a deviation exceeding ±10% from the linear correlation equation (second group). In this comparative example, calibration models corresponding to the first and second groups were adopted respectively (i.e., there were two groups and two calibration models), and no correction was made for the second group. The comparison results between leakage amount and measured value were obtained in the same manner as in Example 1. The comparison results are shown in Table 1.

[0073] <Comparative Example 2> Without performing the classification of whether correction was necessary using feature quantities as in the above example, the data was classified into five types based on the type of filter (hollow fiber module) of the dialyzer 10 of the dialysis system 1, and a corresponding calibration model was created. Then, the leakage amount was compared with the measured value. In other words, in this comparative example, both the number of groups and the calibration model are five.

[0074] [Table 1]

[0075] In Comparative Example 1, where no correction was applied, the maximum error in the predicted leakage amount was relatively large. In Comparative Example 2, the maximum error in the predicted leakage amount was relatively small, but the number of groups was extremely large, with 5 groups for the fluorescence spectra of 24 cases. By applying correction as in Examples 1, 2, and 3, the number of groups decreased compared to Comparative Examples 1 and 2, the maximum error in the predicted leakage amount decreased, and an improvement in accuracy was observed. In Example 4, although the maximum error in the predicted leakage amount increased, the number of groups decreased. By applying correction, the number of groups decreased, and an improvement in prediction accuracy was observed with a correction process that exceeded a percentage deviation from linear correlation of 3-15%. [Industrial applicability]

[0076] The present invention is useful in providing a concentration calculation device that can measure (calculate) the concentration of substances such as albumin with high accuracy without unnecessarily increasing the number of groups to be classified. [Explanation of symbols]

[0077] 1…Dialysis system (blood processing system) 16...Concentration calculation device 71...Classification section 72...Concentration calculation section 80... Irradiation area 81...Detection unit 94...Correction section S1…Irradiation process S2...Detection process S3…Classification process S4…Correction process S5…Calculation process

Claims

1. An irradiation unit that irradiates the object to be used for concentration calculation with excitation light, A detection unit for detecting fluorescence emitted from the object whose concentration is to be calculated, which has been irradiated by the irradiation unit, A classification unit that classifies fluorescence spectra into a first group that does not require correction and a second group that requires correction, based on a first and second characteristic quantity extracted from the fluorescence spectrum of the fluorescence detected by the detection unit, A correction unit for correcting the fluorescence spectra classified into the second group, The system comprises a concentration calculation unit that calculates the concentration of a substance contained in the substance to be calculated based on either or both of the following: the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum classified into the first group and a specific calibration model, and the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum corrected by the correction unit and the specific calibration model, The standardized intensity is obtained by dividing the fluorescence intensity at each wavelength by the maximum fluorescence intensity value of the fluorescence spectrum detected by the detection unit. The spectrum consisting of the standardized intensities for each wavelength is defined as the standardized spectrum. In the standardized spectra obtained multiple times, the standardized intensity near the wavelength where the change in standardized intensity between each standardized spectrum is largest is defined as the first feature quantity. The second feature quantity is defined as the maximum fluorescence intensity value, its reciprocal, or the value obtained by transforming the maximum fluorescence intensity value using a specific formula. Concentration calculation device.

2. The concentration calculation device according to Claim 1, wherein the classification unit calculates a linear correlation equation from each data plotted with the first feature quantity on the vertical axis and the second feature quantity on the horizontal axis, and classifies data that falls beyond a predetermined range from the linear correlation equation into the second group.

3. The maximum fluorescence intensity value F MAX The upper limit intensity value that the spectrometer included in the detection unit can detect is F. DUL In that case, The second feature quantity I2 is given by the following equation (1) I2=(F DUL -F MAX ) / F DUL …(1) A concentration calculation device according to claim 2, which is calculated by the method described above.

4. The concentration calculation device according to claim 2 or 3, wherein the data that deviates beyond a predetermined range from the linear correlation formula is data that is more than ±3% to more than ±15% from the linear correlation formula.

5. Let the first and second feature quantities be I1 and I2, respectively, let the uncorrected fluorescence intensity be F, and let the maximum fluorescence intensity value be F. MAX The upper limit intensity value that the spectrometer included in the detection unit can detect is F. DUL Let the slope and intercept of the linear correlation equation be a and b, respectively, and let x be a constant between -0.15 and 0.

15. Fluorescence intensity F in the corrected fluorescence spectrum by the correction unit. C This is given by the following equation (2) F C =F×{F DUL -(I2-|[I1-{(I2×a+b)×(1+x)}]| / a)×F DUL } / F MAX …(2) A concentration calculating device according to any one of claims 2 to 4, which is calculated by [the specified method].

6. Let the first and second feature quantities be I1 and I2, respectively, let the uncorrected fluorescence intensity be F, and let the maximum fluorescence intensity value be F. MAX The upper limit intensity value that the spectrometer included in the detection unit can detect is F. DUL Let the slope and intercept of the linear correlation equation be a and b, respectively, and let x be a constant between -0.15 and 0.

15. Fluorescence intensity F in the corrected fluorescence spectrum by the correction unit. C This is given by the following equation (3) F C =F×{F DUL -(I2-[I1-{(I2×a+b)×(1+x)}] / a)×F DUL } / F MAX …(3) A concentration calculating device according to any one of claims 2 to 4, which is calculated by [the specified method].

7. A blood processing system comprising a concentration calculation device according to any one of claims 1 to 6.

8. An irradiation step in which excitation light is irradiated onto the object to be used for concentration calculation, A detection step for detecting fluorescence emitted from the object to be used for concentration calculation that was irradiated in the irradiation step, A classification step in which fluorescence spectra are classified into a first group that does not require correction and a second group that requires correction, based on a first and second characteristic quantity extracted from the fluorescence spectrum of the fluorescence detected in the above detection step, A correction step for correcting the fluorescence spectra classified into the second group, The calculation step includes calculating the concentration of a substance contained in a substance for which concentration calculation is performed based on either or both of the following: the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum classified into the first group and a specific calibration model, and the fluorescence intensities at multiple wavelengths in a predetermined wavelength range of the fluorescence spectrum corrected by the correction step and the specific calibration model, The standardized intensity is obtained by dividing the fluorescence intensity at each wavelength by the maximum fluorescence intensity value of the fluorescence spectrum detected in the above detection step. The spectrum consisting of the standardized intensities for each wavelength is defined as the standardized spectrum. In the standardized spectra obtained multiple times, the standardized intensity near the wavelength where the change in standardized intensity between each standardized spectrum is largest is defined as the first feature quantity. The second feature quantity is defined as the maximum fluorescence intensity value, its reciprocal, or the value obtained by transforming the maximum fluorescence intensity value using a specific formula. Concentration calculation method.

Images