JP2004138454A - Method and apparatus for estimating optical scattering characteristics - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and an apparatus, which easily and precisely estimate scattering characteristics of a medium to be measured, i.e., concentration of scatterers or the like, when optical scattering characteristics of the medium to be measured containing the scatterers are varied. <P>SOLUTION: The method which projects light into the medium to be measured containing the scatterers and marker components and estimates the scattering characteristics of the medium to be measured based on the spectra of detected light, is provided with a procedure which obtains a standard function from spectra of light detected from a medium to be measured with known scattering characteristics in order to calculate the concentration of the marker components, a procedure which obtains a relation of the deviation between variation values of the scattering characteristics and estimate values estimated by using the standard function of the marker component concentration, and a procedure which obtains estimate values from the spectra of the light detected from the medium to be measured with unknown scattering characteristics by using the standard function of the marker component concentration and determines the unknown scattering characteristics from the deviation existing in the estimate values. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
BACKGROUND OF THE INVENTION
In the present invention, when the optical scattering characteristic of the medium to be measured changes with the change of the scatterer density in the medium to be measured including the scatterer, the scattering characteristic of the medium to be measured, that is, the scatterer density, etc. The present invention relates to a method and apparatus for estimating with high accuracy.
[0002]
[Prior art]
Conventionally, in order to measure the blood sugar level in the blood in a living body, it is most common to collect blood from the living body and directly measure the concentration of sugar (eg glucose) in the blood sample by chemical analysis or the like. It was the target. However, in order to collect blood from a living body, a person who has acquired specialized techniques for that purpose is required, and it cannot be easily performed by anyone. Moreover, since blood collection is accompanied by pain and anxiety, it is desirable that blood can be measured without collecting blood if possible. For this reason, the realization of a measuring instrument that measures blood glucose level without damaging the living body has been awaited.
[0003]
Therefore, Japanese Patent Application No. 2001-228507 has been filed by the applicant of the present invention. The technique disclosed in this Japanese Patent Application No. 2001-228507 is a method in which light is incident on a sample such as blood and a detection target component of the sample in the sample is analyzed by analyzing the detection spectrum of the light that has passed through the sample or scattered by the sample. Is a measure of the concentration. An optical component measuring method and apparatus capable of accurately calculating the influence of the concentration of the measurement target component on the spectrum of the detection light by a stochastic statistical simulation and obtaining an accurate measurement result are provided. It is.
[0004]
[Problems to be solved by the invention]
The invention of Japanese Patent Application No. 2001-228507 is a technique useful for measuring blood glucose levels without damaging a living body. However, a sample such as blood contains substances of various chemical components in addition to the target sugar (for example, glucose), and also includes various blood cells that become light scatterers. Yes. Other scatterers include cell walls such as blood cells in blood, cell nuclei, and mitochondria. In addition, the light is also scattered by the structure of the medium to be measured and has scattering characteristics. As such a structure, when a living body is considered, the structure of a living tissue such as skin and muscle can be cited. In general, in order to optically measure the concentration of a specific component in a measured medium that includes a scatterer or has scattering characteristics, it is first necessary to accurately determine the optical scattering characteristics of the measured medium. is there. However, obtaining the optical scattering characteristics of the medium to be measured with high accuracy has the following difficulties.
[0005]
First, the Beer-Lambert law cannot be applied as it is to a medium to be measured including a scatterer. In general, the relationship between the intensity of incident light and transmitted light with respect to a medium to be measured is expressed by Beer-Lambert law. FIG. 1 shows the intensity I of a sample for measuring optical scattering characteristics._inIt is a figure which shows the relationship between the incident light at the time of entering incident light, and an emitted light. FIG. 1A shows a case where the sample does not include a scatterer. Part of the incident light is absorbed by the sample, and the intensity I_outIt is detected as emitted light. The absorbance A in this case is represented by the following equation.
A = log (I_in/ I_out) = Εcd ... Formula 1
[0006]
Here, “log” represents a logarithmic function with a base of 10, ε is an extinction coefficient per unit distance and unit concentration, c is a concentration of a component, and d is a thickness of the sample, that is, an optical path length. . Equation 1 is known as Beer-Lambert law. If the sample does not contain a scatterer, the absorbance for each wavelength of the incident light can be obtained from Equation 1, so it is relatively easy to obtain the concentration of the component contained in the sample from the spectrum of the emitted light. It is.
[0007]
FIG. 1B shows a case where a scatterer is included in the sample. In this case, the intensity I of the incident light_inIntensity of emitted light (detection light) against_outCannot be expressed simply as in Equation 1. That is, the Beer-Lambert law cannot be applied as it is. In principle, it follows the Beer-Lambert law, but the concentration c of the target component cannot simply be obtained from Equation 1. This is because there are various paths from the light incident point to the detection point as shown in the figure, and the optical path lengths in the respective paths are also different from each other. As a result, the optical path length d of Equation 1 becomes unknown, and the concentration c of the target component cannot be simply obtained from Equation 1. The intensity of the detection light needs to be obtained by adding the light passing through various paths having different optical path lengths.
[0008]
There are the following methods for measuring scattering characteristics such as a scatterer concentration or a scattering coefficient of a sample by applying the Beer-Lambert law. First, an inverse problem solution based on light propagation analysis is applied to diffuse transmittance and diffuse reflectance (absorbance) measured using a spectrophotometer and an integrating sphere to determine the scattering coefficient and estimate the scatterer concentration. There is a way to do it. Secondly, the time-resolved waveform of diffuse transmitted light or diffuse reflected light is measured by ultra-fast time-resolved measurement in femtosecond and picosecond scales, and the inverse problem solution based on light propagation analysis is applied to the measurement results. There is a method for estimating the scatterer concentration by obtaining the scattering coefficient. These methods can be performed under special conditions in the laboratory (such as the condition that the thickness of the sample to be measured is extremely thin), but measurement is extremely difficult under general measurement conditions. It takes time to analyze measurement data. Further, the measurement accuracy is about 20% error, and the measurement accuracy is poor.
[0009]
Rather than measuring the scattering characteristics by applying the Beer-Lambert law, it is also possible to estimate the scattering characteristics of the sample by an analytical method. For example, the scatterer concentration within a certain range can be estimated by setting the scatterer concentration of the sample together with the concentration of the specific component to be obtained as a target of multivariate analysis. However, it is difficult to achieve a sufficiently high measurement accuracy with this method. In particular, when the change in scatterer concentration is small, the estimated value includes a relatively large error compared to the amount of change. . And in order to improve the measurement accuracy in this method, it is necessary to measure many types of samples having different scatterer concentrations in advance, and it is necessary to acquire a huge amount of measurement data. Will spend a lot of time and effort.
[0010]
Therefore, the present invention provides a method and apparatus for easily and accurately estimating the scattering characteristics of a medium to be measured, that is, the scatterer density, etc., when the optical scattering characteristics in the medium to be measured including the scatterers change. And it aims at solving the above difficulties.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, the optical scattering characteristic estimation method of the present invention makes light incident on a measured medium including a scatterer and a marker component, passes through the measured medium, and exits from the measured medium. A method of detecting a spectrum of detected light and estimating a scattering characteristic of the medium to be measured based on the spectrum, wherein a concentration of the marker component is determined from a spectrum of the detecting light in a medium to be measured having a known scattering characteristic. A procedure for obtaining a calibration function for calculation, a procedure for obtaining a relationship between a variation in scattering characteristics and a deviation generated in the estimated value of the marker component concentration by the calibration function, and detection light in a measured medium having unknown scattering characteristics A procedure for obtaining an estimated value of the marker component concentration from the spectrum of the calibration function using the calibration function, and obtaining an unknown scattering characteristic from a deviation occurring in the estimated value. It is. In addition, in order to obtain | require the spectrum of a detection light, the spectrum of detection light etc. is calculated | required from the detected value by dividing the detection light inject | emitted from a to-be-measured medium, and detecting the intensity | strength of the light in each wavelength.
[0012]
Further, the optical scattering characteristic estimation method of the present invention makes light incident on a measured medium including a scatterer and a marker component, passes through the measured medium, and emits a spectrum of detected light emitted from the measured medium. The first data indicating the relationship between the concentration of the marker component and the spectrum of the detection light in the measured medium having the first known scattering characteristic A procedure for obtaining a group, a procedure for obtaining a calibration function for calculating the concentration of the marker component from the spectrum of the detection light based on the first data group, and a measured medium having a second known scattering characteristic. Using a procedure for obtaining a second data group indicating a spectrum of the detection light with respect to a known first concentration of the marker component, and using the calibration function, A procedure for obtaining a first estimated value for the concentration of the marker component, a procedure for obtaining a deviation between the first concentration and the first estimated value, and obtaining a relationship between the amount of change in the scattering characteristics and the deviation of the estimated value. Using a measurement function to obtain a third data group indicating a spectrum of the detection light with respect to a known second concentration of the marker component in a measured medium having an unknown scattering characteristic, and using the calibration function , Obtaining a second estimated value for the concentration of the marker component from the third data group, obtaining a deviation between the second concentration and the second estimated value, and determining the amount of change in the scattering characteristics and And a procedure for obtaining the unknown scattering characteristic based on the relationship between the amount of change in the estimated value.
[0013]
In the optical scattering characteristic estimation method, the first data group and the second data group can also be obtained by a statistical simulation of light propagation.
[0014]
In the optical scattering characteristic estimation method, the calibration function is preferably obtained by applying multivariate analysis to the first data group.
[0015]
In the optical scattering characteristic estimation method, the detection light may be transmitted light through which the incident light has passed through the medium to be measured, and transmission absorbance may be obtained as a spectrum of the detection light.
[0016]
In the optical scattering characteristic estimation method described above, diffused reflected light in which the incident light is reflected and scattered by the measured medium can be detected as the detection light.
[0017]
In the optical scattering characteristic estimation method, the light incident on the measured medium is preferably near infrared light having a wavelength of 700 to 2500 nm.
[0018]
In addition, the optical scattering characteristic estimation apparatus of the present invention makes light incident on a measured medium including a scatterer and a marker component, passes through the measured medium, and emits a spectrum of detected light emitted from the measured medium. An apparatus for estimating the scattering characteristics of the medium to be measured by: a light source unit that makes light incident on the medium to be measured; a detection unit that detects a spectrum of detection light that has passed through the medium to be measured; and the light source unit And a control calculation unit that controls the detection unit and performs a calculation process of the spectrum of the detection light, and the control calculation unit is configured to measure the detection light in a measured medium having a known first scattering characteristic. Means for storing a calibration function for calculating the concentration of the marker component from the spectrum, and when the scattering property of the measured medium changes from the first scattering property, the amount of change in the scattering property; Means for storing a relationship between deviations occurring in the estimated value of the concentration of the marker component by a calibration function, and measuring a spectrum of the detection light in a measured medium having an unknown scattering characteristic, By applying a calibration function, an estimated value for the concentration of the marker component is obtained, and the unknown scattering characteristic is obtained from a deviation occurring in the estimated value.
[0019]
In the optical scattering characteristic estimation apparatus, the light incident on the measurement medium is preferably near infrared light having a wavelength of 700 to 2500 nm.
[0020]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described with reference to the drawings. First, the measurement principle of the optical scattering characteristic measurement method will be described. FIG. 1 shows the intensity I of a sample for measuring optical scattering characteristics._inIt is a figure which shows the relationship between the incident light at the time of entering incident light, and an emitted light. FIG. 1A shows a case where the sample does not include a scatterer. Part of the incident light is absorbed by the sample, and the intensity I_outIt is detected as emitted light. The absorbance A in this case is represented by the above-described formula 1 (Beer-Lambert law). If the sample does not contain a scatterer, the absorbance for each wavelength of the incident light can be obtained by the Beer-Lambert law expressed by Equation 1, so that the component contained in the sample by the spectrum of the emitted light It is relatively easy to determine the concentration of.
[0021]
FIG. 1B shows a case where a scatterer is included in the sample. Examples of the scatterer in the actual measurement sample include tissues such as skin and muscle and / or blood cells in blood, cell nuclei, mitochondria and the like. In this case, the intensity I of the incident light_inIntensity of emitted light (detection light) against_outCannot be expressed simply as in Equation 1. This is because there are various paths from the light incident point to the detection point as shown in the figure, and the optical path lengths in the respective paths are also different from each other. The intensity of the detection light needs to be obtained by adding the light passing through various paths having different optical path lengths.
[0022]
For a sample including a scatterer as shown in FIG._inIntensity of emitted light against_out(Or absorbance A) can be actually measured. Further, when the intensity (or absorbance) of the emitted light cannot be measured, it can be obtained by a probability statistical simulation described later.
[0023]
And the spectrum of detection light is calculated | required from the intensity | strength of the emitted light in each wavelength. The concentration of the target component and other conditions (such as temperature) are changed to various values, the spectrum of the detection light for the plurality of types of samples is obtained, and the calibration function can be obtained from these spectra. The calibration function is a function that outputs the concentration of the target component when a specific measured spectrum is input. The calibration function will be described in detail later.
[0024]
In addition, in the sample containing the scatterer, the incident light is detected on the same surface as the surface on which the incident light is incident, as shown in FIG. can do. Even in this case, as in the case of FIG. 1B, the intensity I of the incident light is obtained by actual measurement or probability statistical simulation._inIntensity of emitted light against_outCan be requested.
[0025]
A case where the Monte Carlo method is used will be described as an example of a stochastic simulation. FIG. 2 is a diagram showing optical path tracking by the Monte Carlo method. First, energy W₀It is assumed that the incident light enters the sample. Consider an orthogonal three-axis xyz coordinate system as shown, with the incident direction being the z-axis. This luminous flux is scattered at the origin of the xyz coordinate system, and the scattered light has energy W₁It is indicated by the luminous flux. The zenith angle θ is an angle formed by the z axis and scattered light, and the azimuth angle φ is an angle formed by projection of the scattered light on the xy plane and the x axis. The stroke length until the scattered light is subjected to the next scattering is L. Each parameter of the scattered light is expressed by the following equation.
[0026]
L = -ln (R₁) / (Μ_s+ Μ_a)
θ = f^-1(G, R₂)
φ = 2πR₃
[0027]
Here, “ln” represents a natural logarithm function with e as the base. R₁, R₂, R₃Are uniform random numbers in the interval [0, 1]. Also, μ_aIs the absorption coefficient of the measurement sample, μ_sIs a scattering coefficient of the measurement sample, and g is a parameter indicating scattering anisotropy. The parameter g is a parameter used for describing a phase angle function (for example, a Henry-Greenstein function) indicating the directivity of scattering. The function f in the above equation is a known function expressed using this phase angle function.
[0028]
Energy W₁The following light beam undergoes the following scattering at a point indicated by polar coordinates (r, θ, φ). This secondary scattered light has energy W₂It is indicated by the luminous flux. Then, the secondary scattered light is subjected to the next scattering at a point indicated by polar coordinates (r ′, θ ′, φ ′) in the x′y′z ′ coordinate system. In this way, the optical path of scattered light is tracked one after another. The energy of the scattered light is expressed by the following equation.
[0029]
W₀= I_in/ N
W₁= W₀{Μ_s/ (Μ_s+ Μ_a)}
W₂= W₁{Μ_s/ (Μ_s+ Μ_a)} = W₀{Μ_s/ (Μ_s+ Μ_a)}²
・ ・ ・ ・
W_m= W_m-1{Μ_s/ (Μ_s+ Μ_a)} = W₀{Μ_s/ (Μ_s+ Μ_a)}^m
i_out= W₀{Μ_s/ (Μ_s+ Μ_a)}^m
I_out= Σ i_out
[0030]
In the above formula, N indicates the number of simulations. That is, W₀Is the energy of the incident light flux in each simulation, and the total incident energy I_inDivided by N. Above W_mIndicates the energy of the luminous flux after being scattered m times. And the energy i of the detected light beam in one simulation_outIs the energy W of the light beam after being scattered m times when the light beam reaches the detection area._mIt is expressed as, and becomes 0 when the light beam does not reach the detection region. Total energy I of detected light I_outIs the energy i of the detected light beam at each of the N simulations._outIt will be the sum of all.
[0031]
In this way, the optical path and energy of the light beam reaching the detection position can be obtained by computer simulation for the light beam of incident light. If this simulation is repeated, for example, 10,000 times, and the energy of the light flux reaching the detection position is added, the light propagation characteristic can be statistically obtained. Although the optical path tracking by the Monte Carlo method is described here, the intensity of the detected light can also be obtained by using another stochastic method such as a random walk method other than the Monte Carlo method.
[0032]
Actually, when determining the concentration of the target component in the sample, various factors such as the concentration of other components other than the target component, the concentration of the scatterer, and the temperature are related to the concentration of the target component. An analysis technique such as multivariate analysis is used to extract the concentration information of the target component from among the many factors that change at the same time. However, when measuring changes in the optical properties (absorption characteristics, scattering characteristics) of the measurement target in the measurement of living organisms, if the analysis method such as multivariate analysis is used as it is, the concentration of other components There is a problem that the influence of the change in the optical characteristic value accompanying the change becomes strong, and the error accompanying the concentration estimation of the target component often becomes very large.
[0033]
That is, even if the scatterer concentration is set as the target component and the scatterer concentration is obtained by an analysis technique such as multivariate analysis, the error included in these estimated values may become very large. The present invention does not directly determine the concentration of the scatterer that is the target component, but estimates the concentration of another component as a marker (hereinafter referred to as a marker component), and uses the deviation of the estimated value that occurs at that time, The scatterer concentration, which is difficult to estimate directly, can be estimated with high accuracy and easily. Here, although the amount to be estimated is the scatterer concentration, the scattering coefficient of the sample may be estimated instead of the scatterer concentration. The marker component can be easily measured by other methods such as chemical analysis, and can be dissolved in a solvent so that the component does not cause light scattering. Is desirable.
[0034]
FIG. 3 is a diagram for explaining the difference between the optical scattering characteristic estimation method of the present invention and the existing estimation method and the advantages of the optical scattering characteristic estimation method of the present invention. FIG. 3A shows a simplified procedure of the existing estimation method, and FIG. 3B is a graph schematically showing the relationship between the measured value and the estimated value in the method. In the estimation method of FIG. 3A, first, incident light is incident on a sample, and transmitted light transmitted through the sample or diffusely reflected light scattered by the sample is detected and its spectrum is measured. This is the intensity I of the light emitted from the sample in FIG. 1 (b) or 1 (c)._outWill be measured. In FIG. 3B, the spectrum of the emitted light is measured as absorbance.
[0035]
Next, the scatterer concentration is obtained by an analysis method using the spectrum of the detected light. As an analysis method, the scatterer concentration can be obtained by using multivariate analysis using the scatterer concentration as an objective variable (a variable to be estimated). As another analysis method, assuming the scatterer concentration, the detection light intensity is obtained by the stochastic statistical simulation of light propagation as described above, and such simulation is performed by changing the value of the scatterer concentration in various ways. There is a so-called inverse problem analysis in which a scatterer concentration is obtained such that the detected light intensity by simulation is closest to the actually measured value.
[0036]
Regardless of the analysis method, as shown in FIG. 3B, the amount of change in absorbance when the scatterer concentration changes from the standard value indicated by the dotted line is small. This indicates that, when the scatterer concentration is obtained from the spectrum of the emitted light, that is, the absorbance, a slight measurement error of the absorbance greatly affects the scatterer concentration. That is, with the existing estimation method, it is very difficult to obtain the scatterer concentration with high accuracy.
[0037]
FIG. 3 (c) schematically shows the procedure of the estimation method of the present invention, and FIG. 3 (d) is a graph schematically showing the relationship between the measured value and the estimated value in the method. In the estimation method of FIG. 3C, first, incident light is incident on the sample, and transmitted light that has passed through the sample or diffusely reflected light scattered by the sample is detected and its spectrum is measured. In the sample, a marker component (for example, glucose or the like) is dissolved at a predetermined concentration in addition to the scatterers being uniformly dispersed. The concentration of this marker component is accurately determined by other methods such as chemical analysis.
[0038]
The spectrum measurement is the same as the existing estimation method, and the spectrum of the intensity of the emitted light is measured as shown in FIG. 1B or FIG. For example, the absorbance spectrum is measured as shown in FIG. Then, assuming a standard value as the scatterer concentration, the concentration of the marker component is obtained by an analysis method. As an analysis method, the scatterer concentration can be obtained by using multivariate analysis using the concentration of the marker component as an objective variable. As another analysis method, the light propagation simulation as described above may be performed, and the concentration of the marker component may be obtained by inverse problem analysis.
[0039]
At this time, if the actual scatterer concentration exactly matches the assumed standard value, the concentration of the marker component can be obtained with high accuracy by such an analysis method. However, when the actual scatterer concentration slightly changes from the standard value, the concentration of the marker component obtained by the analysis method greatly changes from the standard concentration. This is illustrated in FIG. 3 (d). When the scatterer concentration is the standard value (indicated by a vertical dotted line), the deviation (error) of the marker component concentration becomes 0, but when the scatterer concentration changes from the standard value, the marker component concentration greatly changes. The deviation increases.
[0040]
In other words, the deviation in the concentration of the marker component obtained by the analysis method is very sensitive to changes in the scatterer concentration. In the present invention, the scatterer concentration is estimated with high accuracy from the deviation of the concentration of the marker component obtained by the analysis method by using such knowledge in reverse. That is, the deviation of the marker component concentration when the scatterer concentration changes from the standard value by a predetermined amount can be obtained in advance by actual measurement or simulation, and the scatterer concentration corresponding to the actual deviation can be obtained. By such a method, it is possible to estimate with high accuracy the scatterer concentration of a sample which is usually difficult to detect with high accuracy.
[0041]
In the above estimation method, the absorbance spectrum used for multivariate analysis usually uses measured values, but can also be obtained by the above-described light propagation simulation. When using the light propagation simulation, it is necessary to know changes in the optical characteristic values (absorption coefficient, scattering coefficient, etc.) of the sample due to changes in the concentration of each component. The change in the optical characteristic value is obtained by actual measurement or is obtained by calculation from the literature value of the optical characteristic of each component. In addition, it is necessary to confirm that the simulation result of the change in absorbance due to the change in these optical characteristic values agrees well with the actual measurement value under actual measurement conditions.
[0042]
Further, in the above estimation method, the spectrum of the detection light is measured as the absorbance by the transmitted light, but the intensity spectrum of the diffuse reflected light may be measured as shown in FIG.
[0043]
FIG. 4 is a diagram showing each procedure of the optical scattering characteristic estimation method of the present invention in more detail. First, in the procedure 1, sample data for estimating the marker component concentration is created with the scatterer concentration of the sample as a standard value. The sample data is created by actual measurement or light propagation simulation. The sample scatterer concentration is fixed as a standard value, the marker component concentration is changed to a plurality of types, and sample data of an absorbance spectrum is created for each marker component concentration.
[0044]
Marker component concentration is n types of concentration C_iAnd each concentration C_iAbsorbance A for sample_λSuppose that (i) is observed. Here, i = 1 to n, and the wavelength λ of light means a number of discrete wavelengths necessary for spectrum measurement. For example, the marker component is glucose, and three concentrations C₁= 1000, C₂= 2000, C₃= 3000 (unit: [mg / dL]), three types of absorbance spectra A corresponding to each concentration sample_λ(1), A_λ(2), A_λ(3) is obtained.
[0045]
Although only the sample data when the marker component concentration is changed is shown here, if it is necessary to consider disturbance factors such as other component concentrations and temperatures, the sample data with those values also changed is shown. Need to create.
[0046]
Next, in step 2, a calibration function having the marker component concentration as an objective variable when the scatterer concentration is a standard value is obtained. This calibration function is, for example, the three types of concentrations C described in Procedure 1.₁In the example of, it can be obtained as follows. Coefficient β constituting the substance of the calibration function_λAre used to express the three types of concentration as follows:
[0047]
C₁= Σ_λ{Β_λ・ A_λ(1)}
C₂= Σ_λ{Β_λ・ A_λ(2)}
C₃= Σ_λ{Β_λ・ A_λ(3)}
[0048]
Where Σ_λIndicates that the sum is taken over all of the discrete wavelengths used in the spectrum measurement. That is, the coefficient β at each wavelength such that the above equations representing the three known concentrations hold respectively._λ, The coefficient β_λThe unknown marker component concentration C is obtained by the following equation 2 using However, the absorbance spectrum for the unknown marker component concentration C is A_λIt is said.
C = Σ_λ{Β_λ・ A_λ} ... Equation 2
[0049]
That is, Equation 2 is a calibration function with the marker component concentration as an objective variable. Therefore, obtaining the calibration function means that the coefficient β for each wavelength used in the spectrum measurement._λIt is none other than asking for. Coefficient β_λFor example, for a known marker component concentration, a coefficient β that minimizes the error between the concentration obtained by Equation 2 and the actual concentration is as follows._λTo decide. For this, a known multivariate analysis (for example, PLS (partial least square regression analysis), PCR (principal component regression analysis), MLR (multiple regression analysis), etc.) is used. It should be noted that the calibration function of Equation 2 is calculated by using each coefficient β_λTherefore, if the scatterer concentration is a standard value, the marker component concentration can be obtained with high accuracy.
[0050]
Next, in procedure 3, sample data is created from a sample in which the scatterer concentration is changed from the standard value by a predetermined change amount. This is to confirm what deviation (error) occurs in the marker component concentration obtained from the calibration function of Equation 2 when the scatterer concentration changes from the standard value. That is, it is necessary to accurately know the scatterer concentration in this sample. The sample data is created in the same manner as in Procedure 1.
[0051]
Next, in step 4, the marker component concentration (estimated value) is obtained from the sample data obtained in step 3, using the calibration function of equation 2 created in step 2. At this time, if the scatterer concentration is a standard value, there is almost no error in the marker component concentration. However, since the sample data of procedure 3 changes the scatterer concentration from the standard value, the marker component concentration is changed according to the amount of change. There is also a structural deviation in concentration.
[0052]
Next, in step 5, the actual measurement value is obtained by actually measuring the marker component concentration of the sample for which the sample data has been created in step 3. The actual measurement value of the marker component concentration is measured by other methods such as chemical analysis.
[0053]
Next, in step 6, the deviation occurring in the estimated value is obtained by taking the difference between the estimated value of the marker component concentration obtained in step 4 and the actual measured value of the marker component concentration obtained in step 5. At this time, a large deviation occurs in the marker component concentration compared to the amount of change in scatterer concentration, and the change in scatterer concentration is reflected with high sensitivity.
[0054]
Moreover, the change in the scatterer concentration changes the scattering characteristics such as the scattering coefficient, but the wavelength dependence of the change in the scattering characteristics is small. That is, the scattering characteristics change to the same extent at any wavelength. Therefore, the change in the absorbance spectrum accompanying the change in the scatterer concentration is almost the same over the entire wavelength range being measured. As a result, the deviation of the marker component concentration due to the change in the scatterer concentration becomes the same value at any marker component concentration. That is, the deviation is shifted by the deviation from the true marker component concentration. Since such a structural deviation occurs, the amount of change in the scatterer concentration can be obtained from the deviation.
[0055]
Next, in step 7, the standard value ε of the scatterer concentration in the sample of step 3₀The relationship between the amount of change Δε and the deviation E is obtained from the amount of change Δε from the above and the deviation E generated in the marker component concentration C (obtained in step 6). A proportional relationship was recognized between them, and it was found that there was a relationship as shown in the following equation.
E = α · Δε (3)
[0056]
From the change amount Δε and the deviation E, the constant α in Equation 3 can be obtained. Equation 3 also represents the relationship between the variation of the scatterer concentration and the deviation of the marker component concentration for any generalized sample. According to Equation 3, the scatterer concentration can be obtained from the deviation occurring in the marker component concentration with respect to the sample having an unknown scatterer concentration.
[0057]
Next, in step 8, sample data is created by actual measurement for a sample having an unknown scatterer concentration, and the marker component concentration is obtained using the calibration function of equation 2 created in step 2. However, the marker component concentration of this sample must be known or accurately measurable by another method. At this time, the scatterer concentration of the sample is the standard value ε₀If there is no deviation in the marker component concentration, the scatterer concentration is the standard value ε₀If there is a change, the structural deviation E also occurs in the marker component concentration C according to the change amount Δε. Deviation E occurring in the marker component concentration is determined.
[0058]
Next, in step 9, the scatterer concentration of the unknown sample is obtained from the deviation E obtained in step 8. The unknown scatterer concentration is ε (where ε = ε₀(+ Δε), the change amount Δε of the scatterer concentration is obtained from the deviation E according to Equation 3, and thus the unknown scatterer concentration ε can be obtained. Eventually, the scatterer concentration ε is obtained by the following equation 4. That is, Equation 4 is a calibration curve for estimating the scatterer concentration.
ε = ε₀+ E / α S ... Formula 4
[0059]
By estimating the scatterer concentration of an unknown sample according to the procedure 1 to the procedure 9 as described above, the scatterer concentration can be obtained with high accuracy with a relatively small amount of calculation and in a short time. Here, the scatterer concentration is described as an example of the scattering property, but the scattering property can be estimated by the same procedure even if the characteristic value is other than the scatterer concentration (for example, the scattering coefficient). For example, when it is desired to obtain the scattering coefficient of the medium instead of the scatterer concentration, the scattering coefficient of the medium can be obtained by the same procedure by replacing “scattering element concentration ε” with “scattering coefficient μ of medium” in the above procedure. it can.
[0060]
Next, the result of actually estimating the scatterer concentration by the optical scattering characteristic estimation method of the present invention will be described. As the sample, glucose was used as a marker component, and intralipid (lipid particles) as a scatterer was suspended in an aqueous solution of glucose. In this sample, the absorbance spectrum of the sample was measured while changing the marker component concentration, scatterer concentration, and sample temperature, and the effectiveness of the optical scattering characteristic estimation method of the present invention was verified.
[0061]
The measurement conditions of the sample are as follows. The scatterer concentration was changed in the range of 1.0 to 3.0% by volume percentage, the marker component concentration was changed in the range of 1000 to 3000 mg / dL, and the temperature of the sample was changed in the range of 25 to 45 ° C. . Thus, the sample data of the absorbance spectrum of the transmitted light was measured for the sample in which each parameter was changed. The wavelength range of the absorbance spectrum was 1250 to 1800 nm, and measurement was performed at 318 points within this wavelength range.
[0062]
FIG. 5 shows an example of sample data of the absorbance spectrum. However, FIG. 5 shows an outline of the entire absorbance spectrum and does not represent actual measurement data itself. The absorbance spectrum in each case was measured with a scatterer concentration of 1, 2, and 3%. In FIG. 5, three spectral curves are shown for each scatterer concentration for simplification, but actually nine spectral curves exist.
[0063]
First, according to the procedure 1 of FIG. 4, sample data of the absorbance spectrum was measured in order to create a calibration function. The scatterer concentration is fixed at 1%, the marker component concentration is changed to 3 types of 1000, 2000, 3000 [mg / dL], the sample temperature is changed to 3 types of 25, 35, 45 [° C.], Absorbance spectrum sample data were measured for a total of nine samples.
[0064]
Next, according to the procedure 2, a calibration function having the marker component concentration as an objective variable when the scatterer concentration is a standard value (1%) was obtained. A calibration function as shown in Equation 2 was obtained by applying PLS, which is one method of multivariate analysis, to nine types of sample data in Procedure 1. Specifically, a calibration function using regression coefficients up to the first to fifth principal components was obtained. For specific application of PLS, commercially available analysis software was used. In PLS, the number of parameters can be greatly reduced by principal component analysis, so the solution can be obtained even if the number of marker component concentrations is smaller than the number of wavelengths used, compared to the normal least square method. Benefits.
[0065]
Next, according to the procedure 3, sample data of the sample was measured when the scatterer concentration was changed from the standard value (1%) to 3%. Here, the marker component concentration is changed to three types of 1000, 2000, and 3000 [mg / dL], and the sample temperature is changed to three types of 25, 35, and 45 [° C.], for a total of nine types of samples. Absorbance spectrum sample data was prepared. Here, nine types of sample data are created for verification of the deviation, but it is not always necessary to create a plurality of types of sample data, and only one type of sample data may be created.
[0066]
Next, according to the procedure 4, the calibration function created in the procedure 2 is applied to the sample data obtained in the procedure 3 to obtain the marker component concentration. The calibration function created in step 2 is for the case where the scatterer concentration is the standard value (1%), so there is almost no error when applied to the sample data for the scatterer concentration of 1% in step 1. The marker component concentration can be estimated with high accuracy. FIG. 6 shows the results when the calibration function is applied to sample data having a scatterer concentration of 1%. A black dot indicates each estimated value of the marker component concentration. A dotted line indicates an estimated value when the error is zero. As shown in the figure, the estimated value is obtained in a range that substantially matches the true value.
[0067]
However, since the sample data of procedure 3 has a scatterer concentration of 3%, when the calibration function of procedure 2 is applied to this sample data, the marker component concentration depends on the amount of change from the standard value of the scatterer concentration. There are also structural deviations. A deviation occurring in the marker component concentration is obtained. The result at this time is shown in FIG. The lower black dot indicates each estimated value of the marker component concentration. The lower dotted line indicates the average value of the estimated values. The upper dotted line indicates an estimated value, that is, a true value when the error is zero. As shown in the figure, a large deviation occurs in the estimated value of the marker component concentration, and the value of the deviation E is -7871 mg / dL. Here, obtaining the true value of the marker component concentration corresponds to the procedure 5, and obtaining the deviation E corresponds to the procedure 6.
[0068]
Next, according to the procedure 7, the relationship between the change amount of the scatterer concentration and the generated deviation is obtained. Since the amount of change Δε from the standard value of the scatterer concentration is 2% and the value of deviation E from procedure 4 is −7871 mg / dL, in Equation 3, E = −7871 and Δε = 2, and the constant α is When calculated, α = −3935.5.
[0069]
Next, according to Procedure 8, sample data is measured for a sample having an unknown scatterer concentration, and the marker component concentration is obtained using the calibration function of Equation 2 created in Procedure 2. Here, sample data of a sample having a scatterer concentration of 2%, a marker component concentration of 2000 mg / dL, and a sample temperature of 35 ° C. was measured as a sample having an unknown scatterer concentration. The scatterer concentration of this sample was estimated according to the estimation method of the present invention, and the effectiveness of the estimation method of the present invention was verified.
[0070]
When the calibration function of the procedure 2 is applied to the sample data measured in the procedure 8, the estimated value of the marker component concentration is −1866 mg / dL, and the deviation E is −3866 mg / dL with respect to the true value of the marker component concentration of 2000 mg / dL. It became.
[0071]
Next, according to procedure 9, the unknown scatterer concentration was estimated. That is, in Equation 4, ε₀= 1.0, E = -3866, α = −3935.5 were substituted, and an estimated value of 1.98% with respect to the scatterer concentration ε was obtained. From this result, it was verified that the scatterer concentration can be estimated easily and with high accuracy according to the present invention.
[0072]
Further, as a sample having an unknown scatterer concentration, the scatterer concentration is 2%, the marker component concentration is changed to three types of 1000, 2000, 3000 [mg / dL], and the sample temperature is 25, 35, 45. When the scatterer concentration is similarly estimated by the estimation method of the present invention for the other samples changed to [° C.], the estimated value of the scatterer concentration is 1.962 for any sample. It was within the range of 2.019%.
[0073]
On the other hand, as an existing estimation method, the result when the final target scatterer concentration is obtained directly from the absorbance spectrum is shown below. Sample with scatterer concentration of 2%, marker component concentration changed to 3 types of 1000, 2000, 3000 [mg / dL], and sample temperature changed to 3 types of 25, 35, 45 [° C.] On the other hand, when the scatterer concentration was estimated by the existing estimation method, the estimated value of the scatterer concentration was in the range of 1.908 to 2.053%. Therefore, according to the estimation method of the present invention, the variation in the estimated value of the scatterer concentration is reduced to about one third compared with the existing estimation method.
[0074]
Further, when obtaining the final scatterer concentration directly from the absorbance spectrum by ordinary multivariate analysis, it is necessary to remarkably increase the number of sample data in order to improve the accuracy of the scatterer concentration. In the present invention, since the linear relationship of Equation 3 between the variation of the scatterer concentration and the deviation of the marker component concentration is used, the number of sample data can be remarkably reduced, and the scattering characteristics can be estimated. Calculation time and cost can be greatly reduced.
[0075]
FIG. 8 is a diagram showing an overall configuration of the optical scattering characteristic estimation apparatus 1 of the present invention. The optical scattering characteristic estimation apparatus 1 includes a light source unit 5 for irradiating measurement light to a sample 7 and a detection unit 6 for detecting a spectrum of detection light that has passed through the sample 7. As the light emitted from the light source unit 5, near infrared light in a predetermined wavelength range, for example, light having a wavelength of 700 to 2500 nm is used. Light in this wavelength range has relatively good biological permeability, and is suitable for measurement on a living body. In addition, there are many absorption bands of in-vivo components as marker components in this wavelength range, and it is easy to create a calibration function corresponding to the marker components.
[0076]
The light emitted from the light source unit 5 passes through the sample 7 and its spectrum is detected by the detection unit 6. The control calculation unit 2 controls the light source unit 5 and the detection unit 6 to perform measurement, performs various processes on the spectrum detected by the detection unit 6, and estimates the scattering characteristics of the sample 7. The display unit 3 displays various calculation results and estimated values. The input unit 4 is for an operator to input various operation instructions and data to the control calculation unit 2. A CRT, a liquid crystal display panel or the like can be used as the display unit 3, and a keyboard or a pointing device can be used as the input unit 4.
[0077]
FIG. 9 is a block diagram illustrating a configuration of the control calculation unit 2 of the optical scattering characteristic estimation apparatus 1. The control arithmetic unit 2 is provided with a CPU 20 for performing various data processing, and a memory 22 such as a ROM or a RAM is connected to the CPU 20 via a bus 21. The CPU 20 operates according to programs and data stored in the memory 22. The memory 22 includes an OS (operating system) that is a basic program, a scattering characteristic estimation program 221 for estimating the scattering characteristic of the sample from the spectrum detected by the detection unit 6, and a multiplicity for performing multivariate analysis. A variable analysis program 222 and the like are stored.
[0078]
The calibration function storage area 223 in the memory 22 stores a calibration function having the marker component concentration as an objective variable when the scattering characteristic is a standard value. The same calibration function can be used as long as the component configuration and measurement environment of the measurement sample are not significantly changed. It is also possible to store a plurality of calibration functions and switch the calibration functions according to the type of sample.
[0079]
Further, the constant storage area 224 in the memory 22 stores the constant α in Expression 3. With this constant α, the relationship between the amount of change Δε from the standard value of the scattering characteristic (scatterer concentration) and the deviation E generated in the estimated value of the marker component concentration can be understood. As with the calibration function, the constant α can be the same as long as the composition of the measurement sample and the measurement environment do not change significantly. It is also possible to store a plurality of types of constant α and switch between the calibration function and the constant α according to the type of sample.
[0080]
Further, the control calculation unit 2 is provided with a fixed disk device 23. The fixed disk device 23 stores various programs and various data to be loaded into the memory 22. Further, measured spectrum data, various calculation results, estimated values of scattering characteristics, and the like are stored in the fixed disk device 23.
[0081]
In addition, a display unit 3 for displaying characters and images and an input unit 4 for an operator to input data and the like are connected to the control calculation unit 2 via an interface circuit 24. The control calculation unit 2 is connected to the light source unit 5 and the detection unit 6 through the interface circuit 25, and controls them to measure the absorbance spectrum of the sample 7.
[0082]
A calibration function to be created in steps 1 to 2 in FIG. 4 is created in advance and stored in the control calculation unit 2, and from the standard value of the scattering characteristic to be obtained in steps 3 to 7 in FIG. 4. The relationship (ie constant α) between the amount of change and the deviation occurring in the estimated value of the marker component concentration is obtained and stored in advance. Therefore, steps 8 and 9 in FIG. 4 are performed by the optical scattering characteristic estimation apparatus 1 during measurement.
[0083]
That is, sample data of a sample having an unknown scattering characteristic is obtained by actual measurement, a calibration function is applied to the sample data, a marker component concentration is obtained, and a deviation occurring in the estimated value of the marker component concentration is obtained. Note that the actual concentration of the marker component is determined in advance by another measurement method. Then, according to the procedure 9, the scattering characteristic of the sample is estimated from the deviation value. As described above, since steps 8 and 9 are executed at the time of measurement, the amount of calculation is small and immediate measurement is possible.
[0084]
As described above, according to the optical scattering characteristic estimation method and apparatus of the present invention, an estimated value for the scattering characteristic of a sample can be obtained with high accuracy and in a short time. In addition, since the estimated value can be calculated at high speed, it is possible to measure the sample in real time. For example, it is possible to measure in real time blood flowing in an artificial kidney or heart-lung machine outside the body. . Then, by monitoring the change in the scattering characteristic of the sample and incorporating it in the measurement, it is possible to eliminate the measurement error of the target component concentration due to the change in the scattering characteristic.
[0085]
In addition, when measuring a medium in a living body, if a component whose concentration is kept almost constant in the living body is selected as a marker component, the marker component concentration is constant, so that it is obtained by measurement using another measurement method. This eliminates the need and makes it easier to estimate the scattering characteristics. A calibration function for such a marker component having a constant concentration can be obtained by creating sample data of a detected light spectrum by light propagation simulation.
[0086]
In the embodiment described above, in the verification examples of FIGS. 5 to 7, the wavelength range of spectrum measurement is 1200 to 1800 nm. Not only this wavelength range, but also visible light having a wavelength of about 400 to 700 nm, wavelength 700 Near infrared light of ˜4000 nm, whole infrared light having a wavelength of 700 nm or more, and the like can also be used. In addition, although the scattering characteristic is mainly described as an example of the scattering characteristic, it is possible to estimate the scattering characteristic by the same procedure even if the characteristic value (for example, the scattering coefficient) other than the scatterer density is used. .
[0087]
【The invention's effect】
Since the present invention is configured as described above, the following effects can be obtained.
[0088]
The scattering characteristics are estimated using the relationship between the amount of change in the scattering characteristics and the deviation generated in the estimated value by the calibration function of the marker component concentration. Can be sought. In addition, since the estimated value can be calculated at high speed, it is possible to measure the sample in real time. For example, it is possible to measure in real time blood flowing in an artificial kidney or heart-lung machine outside the body. . Then, by monitoring the change in the scattering characteristic of the sample and incorporating it into the measurement, it is possible to eliminate the measurement error of the target component concentration due to the change in the scattering characteristic.
[0089]
If the sample data of the detected light spectrum is obtained by a stochastic statistical simulation of light propagation, a calibration function can be created even for samples that are difficult to measure, and the estimated value for the scattering characteristics of the sample can be increased. The accuracy can be obtained in a short time.
[0090]
When the spectrum is measured with near-infrared light having a wavelength of 700 to 2500 nm, the permeability of the living body is relatively good, and it is suitable for measurement of a living body. Specifically, it is possible to monitor a biological change accompanied by a change in scattering characteristics, for example, a change in a biological structure such as skin, a change in a brain activity site, or the like. In addition, there are many absorption bands of in-vivo components as marker components in this wavelength range, and it is easy to create a calibration function corresponding to the marker components. Furthermore, it is possible to perform measurement without being invasive to the living body, that is, without damaging the living body.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a relationship between incident light and emitted light on a sample for which optical scattering characteristic measurement is performed.
FIG. 2 is a diagram illustrating optical path tracking by a Monte Carlo method.
FIG. 3 is a diagram for explaining the difference between the optical scattering characteristic estimation method of the present invention and the existing estimation method and the advantages of the present invention.
FIG. 4 is a diagram showing in more detail each procedure of the optical scattering characteristic estimation method of the present invention.
FIG. 5 is a diagram showing an example of sample data of an absorbance spectrum.
FIG. 6 shows the results when a calibration function is applied to sample data.
FIG. 7 shows the results when the calibration function is applied to sample data with different scatterer concentrations.
FIG. 8 is a diagram showing an overall configuration of an optical scattering characteristic estimation apparatus of the present invention.
FIG. 9 is a block diagram illustrating a configuration of a control calculation unit of the optical scattering characteristic estimation apparatus.
[Explanation of symbols]
1 ... Optical scattering characteristic estimation device
2 ... Control operation part
3 ... Display section
4 ... Input section
5 ... Light source
6 ... Detection unit
7 ... Sample
20 ... CPU
21 ... Bus
22 ... Memory
23. Fixed disk device
24, 25 ... interface circuit
35 ... Sample temperature
221 ... Scattering characteristic estimation program
222 ... Multivariate analysis program
223 ... Calibration function storage area
224 ... Constant storage area

Claims (9)

A method in which light is incident on a medium to be measured including a scatterer and a marker component, and a scattering characteristic of the medium to be measured is estimated from a spectrum of detection light that passes through the medium to be measured and is emitted from the medium to be measured. There,
A procedure for obtaining a calibration function for calculating the concentration of the marker component from the spectrum of detection light in a measured medium having known scattering characteristics;
A procedure for obtaining a relationship between a variation in scattering characteristics and a deviation occurring in an estimated value by the calibration function of the marker component concentration;
An optical method comprising: obtaining an estimated value of the marker component concentration from the detected light spectrum in a measured medium having an unknown scattering characteristic by the calibration function, and obtaining an unknown scattering characteristic from a deviation occurring in the estimated value. Scattering characteristic estimation method. A method in which light is incident on a medium to be measured including a scatterer and a marker component, and a scattering characteristic of the medium to be measured is estimated from a spectrum of detection light that passes through the medium to be measured and is emitted from the medium to be measured. There,
A procedure for obtaining a first data group indicating a relationship between a concentration of the marker component and a spectrum of the detection light in a measured medium having a known first scattering characteristic;
Obtaining a calibration function for calculating the concentration of the marker component from the spectrum of the detection light from the first data group;
Obtaining a second data group indicating a spectrum of the detection light with respect to a known first concentration of the marker component in a measured medium having a known second scattering characteristic;
A procedure for obtaining a first estimated value for the concentration of the marker component from the second data group using the calibration function;
Obtaining a deviation between the first concentration and the first estimated value, and obtaining a relationship between the amount of change in the scattering characteristic and the deviation of the estimated value;
A procedure for obtaining, by measurement, a third data group indicating a spectrum of the detection light with respect to a known second concentration of the marker component in a measured medium having an unknown scattering characteristic;
A procedure for obtaining a second estimated value for the concentration of the marker component from the third data group using the calibration function;
A step of obtaining a deviation between the second concentration and the second estimated value, and obtaining the unknown scattering characteristic based on the relationship between the amount of change in the scattering characteristic and the amount of change in the estimated value. Scattering characteristic estimation method. An optical scattering property estimation method according to claim 2,
The optical scattering characteristic estimation method, wherein the first data group and the second data group are obtained by a stochastic statistical simulation of light propagation. An optical scattering property estimation method according to claim 3,
The method for estimating an optical scattering characteristic, wherein the calibration function is obtained by applying multivariate analysis to the first data group. An optical scattering property estimation method according to any one of claims 1 to 4,
An optical scattering characteristic estimation method, wherein the detection light is transmitted light through which the incident light has passed through the medium to be measured, and transmission absorbance is obtained as a spectrum of the detection light. An optical scattering property estimation method according to any one of claims 1 to 4,
An optical scattering characteristic estimation method, wherein the incident light detects diffusely reflected light reflected and scattered by the medium to be measured as the detection light. The optical scattering property estimation method according to any one of claims 1 to 6,
The method for estimating an optical scattering characteristic, wherein the light incident on the measured medium is near infrared light having a wavelength of 700 to 2500 nm. Light is incident on a measured medium (7) including a scatterer and a marker component, and the measured object is measured by a spectrum of detected light that passes through the measured medium (7) and is emitted from the measured medium (7). An apparatus for estimating the scattering properties of a medium (7),
A light source section (5) for making light incident on the measured medium (7);
A detector (6) for detecting a spectrum of detection light that has passed through the measured medium (7);
A control calculation unit (2) for controlling the light source unit (5) and the detection unit (6) and performing calculation processing of the spectrum of the detection light,
The control calculation unit (2)
Means (223) for storing a calibration function for calculating the concentration of the marker component from the spectrum of the detection light in the measured medium (7) having a known first scattering characteristic;
When the scattering characteristic of the measured medium (7) changes from the first scattering characteristic, the relationship between the amount of change in the scattering characteristic and the deviation that occurs in the estimated value of the concentration of the marker component by the calibration function Means for storing (224),
The spectrum of the detected light in the measured medium (7) having unknown scattering characteristics is measured, the calibration function is applied to the spectrum to obtain an estimated value for the concentration of the marker component, and the deviation generated in the estimated value An optical scattering characteristic estimation device for obtaining the unknown scattering characteristic from The optical scattering characteristic estimation device according to claim 8,
The optical scattering characteristic estimation apparatus in which the light incident on the measurement medium (7) is near infrared light having a wavelength of 700 to 2500 nm.

Images