A blood drug concentration detection method, device and system for clinical pharmacy

Abstract

The present invention relates to the field of optical testing technology, and in particular to a blood drug concentration detection method, device and system for clinical pharmacy, including obtaining plasma sample spectral data and image graph, determining the waveform characteristic value of each peak according to the shape characteristics of the plasma sample spectral data peak, obtaining the changing trend and correlation characteristics of the waveform characteristic value with drug concentration, determining the correlation between each peak and blood drug concentration according to the waveform characteristic value of the peak, the changing trend and correlation characteristics of the waveform characteristic value with drug concentration, constructing a histogram of the image graph corresponding to each peak, determining the contribution of each peak according to the correlation and histogram distribution characteristics of each peak with blood drug concentration, performing data dimension reduction on the plasma sample spectral data according to the contribution of each peak, and establishing a detection model to detect blood drug concentration. The present invention reduces the complexity of data processing and improves the accuracy and reliability of blood drug concentration detection.

Description

Technical Field

[0001] The present invention relates to the field of optical testing technology, and in particular to a blood drug concentration detection method, device and system for clinical pharmacy. Background Art

[0002] By detecting blood drug concentrations, doctors and pharmacists can more accurately monitor drug concentrations in patients so as to adjust treatment plans, optimize drug dosages and avoid potential drug interactions. Existing technologies usually use hyperspectral imagers to detect blood drug concentrations. Hyperspectral imagers collect hyperspectral images of blood samples, obtain the absorption and reflection information of blood samples, and thus calculate the drug concentration. The results are transmitted to doctors or pharmacists in real time, improving the safety and effectiveness of drug treatment.

[0003] Hyperspectral data can reflect subtle differences in spectral features, but due to the strong correlation between spectral curve peaks, especially the strong correlation between adjacent peaks, the distribution of data in high-dimensional space is sparse and severely irregular, which will affect the analysis results of spectral data, resulting in reduced accuracy of blood drug concentration detection and affecting the accuracy of blood drug concentration detection. Summary of the invention

[0004] The present invention provides a blood drug concentration detection method, device and system for clinical pharmacy to solve the existing problems.

[0005] The present invention provides a blood drug concentration detection method, device and system for clinical pharmacy using the following technical solutions:

[0006] An embodiment of the present invention provides a method for detecting blood drug concentration in clinical pharmacy, the method comprising the following steps:

[0007] Obtain spectral data and images of plasma samples at different drug concentrations;

[0008] According to the shape characteristics of the peaks in the plasma sample spectral data, the waveform characteristic value of each peak is determined, and the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration are obtained. According to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, the correlation between each peak and the blood drug concentration is determined, and a histogram of the image corresponding to each peak is constructed. According to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak, the contribution of each peak is determined;

[0009] Obtain the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, determine the importance of each peak to the blood drug concentration detection based on the contribution of each peak and the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, reduce the dimension of the plasma sample spectral data based on the importance of the peak, establish a detection model, and detect the blood drug concentration.

[0010] Furthermore, the specific method of determining the waveform characteristic value of each peak includes:

[0011] ;

[0012] In the formula, Indicates The waveform characteristic value of each peak; Indicates The depth of a peak, wherein the depth is the difference between the reflectivity corresponding to the peak and the minimum reflectivity closest to the peak; Indicates The width of a peak, which is the distance between the nearest minimum points on the left and right sides of the peak; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the left side of the peak and the horizontal axis; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the right side of the peak and the horizontal axis; Indicates taking the absolute value; Represents an exponential function with a natural constant as base.

[0013] Furthermore, the determination of the correlation between each peak and the blood drug concentration includes the following specific methods:

[0014] A line graph was drawn with drug concentration as the horizontal axis and peak characteristic value as the vertical axis to obtain a changing line of drug concentration and peak characteristics. The line was fitted using the least squares method to obtain several fitted straight lines, and the Pearson correlation coefficients between different peaks and blood drug concentrations were obtained. The correlation between each peak and blood drug concentration was determined based on the number of plasma samples, the waveform characteristic value of the peak, the slope of the fitting straight line where the peak was located, and the Pearson correlation coefficient between the peak and blood drug concentration.

[0015] Furthermore, the specific calculation method of the correlation between each peak and blood drug concentration is:

[0016] ;

[0017] In the formula, Indicates The correlation between the peaks and blood drug concentration; represents the number of plasma samples; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates The peaks correspond to the slopes of the fitted lines; Indicates the maximum slope of the fitted straight line; Indicates the minimum slope of the fitted straight line; Indicates Pearson correlation coefficient between each peak and blood drug concentration; Indicates taking the absolute value.

[0018] Furthermore, the determination of the contribution of each peak includes the following specific methods:

[0019] According to several image images corresponding to each peak, a grayscale histogram of the image image corresponding to each peak is constructed. According to the number of pixel values ​​in the grayscale histogram of the image image corresponding to each peak and the frequency of occurrence of each pixel value, the information entropy of each peak is determined. According to the correlation between each peak and the blood drug concentration and the information entropy of each peak, the contribution of each peak is determined.

[0020] Furthermore, the specific calculation method of the contribution of each peak is:

[0021] According to The correlation between the first peak and the blood drug concentration and the The information entropy of the peaks determines the The contribution of the first peak The correlation between the first peak and the blood drug concentration, The information entropy of the first peak is The contribution of each peak is directly proportional.

[0022] Furthermore, the specific method for determining the importance of each peak to the blood drug concentration detection includes:

[0023] The importance of each peak to blood drug concentration detection is determined based on the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak. The contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak are directly proportional to the importance of each peak to blood drug concentration detection.

[0024] Furthermore, according to the importance of the peaks, the spectral data of the plasma samples are reduced in dimension, and a detection model is established to detect the blood drug concentration, including the following specific methods:

[0025] The plasma sample spectral data is reduced in dimension using the principal component analysis method, the importance of each peak is used as a weight, a covariance matrix is ​​constructed using the weighted peak data, eigenvectors and eigenvalues ​​are calculated from the covariance matrix, a suitable principal component is selected according to the size of the eigenvalue and the variance ratio, and the plasma sample spectral data is projected onto the selected suitable principal component to achieve dimensionality reduction of the plasma sample spectral data;

[0026] The sample spectral data after dimension reduction were analyzed using the partial least squares regression algorithm, and a linear regression model was established to detect the blood drug concentration.

[0027] A blood drug concentration detection system for clinical pharmacy, using any one of the blood drug concentration detection methods for clinical pharmacy, the system comprises the following modules:

[0028] A data acquisition module, used to obtain spectral data and image graphs of plasma samples at different drug concentrations;

[0029] A peak contribution determination module is used to determine the waveform characteristic value of each peak according to the shape characteristics of the peak in the plasma sample spectral data, obtain the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determine the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, construct a histogram of the image corresponding to each peak, and determine the contribution of each peak according to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak;

[0030] The blood drug concentration detection module is used to obtain the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, determine the importance of each peak to the blood drug concentration detection according to the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, and reduce the dimension of the plasma sample spectral data according to the importance of the peak, and establish a detection model to detect the blood drug concentration.

[0031] A blood drug concentration detection device for clinical pharmacy includes a blood drug concentration detection device and a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any step of the blood drug concentration detection method for clinical pharmacy is implemented.

[0032] The beneficial effects of the technical solution of the present invention are as follows: the embodiment of the present invention determines the waveform characteristic value of each peak through the shape characteristics of the peak of the plasma sample spectral data, accurately reflects the shape characteristics of each peak in the spectral data, obtains the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determines the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, reflects the changing characteristics of the peak with the blood drug concentration, and constructs a histogram of the image graph corresponding to each peak, determines the contribution of each peak according to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image graph corresponding to each peak, and determines the importance of each peak to the blood drug concentration detection according to the contribution of each peak, which helps to screen out the bands related to the blood drug concentration, reduce the complexity of data processing, and improve the accuracy and reliability of blood drug concentration determination. According to the importance of the peak, data dimension reduction is performed, a detection model is established, and the blood drug concentration is detected, which can effectively remove redundant information, reduce the amount of calculation and storage space requirements, and improve the system operation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 A flow chart of the steps of a blood drug concentration detection method for clinical pharmacy of the present invention;

[0035] Figure 2 It is a structural block diagram of a blood drug concentration detection system for clinical pharmacy of the present invention;

[0036] Figure 3 The present invention is a structural block diagram of a blood drug concentration detection device for clinical pharmacy. DETAILED DESCRIPTION

[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a blood drug concentration detection method, device and system for clinical pharmacy proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0039] The following is a detailed description of a specific scheme of a blood drug concentration detection method, device and system for clinical pharmacy provided by the present invention in conjunction with the accompanying drawings.

[0040] See also Figure 1 , which shows a flow chart of a method for detecting blood drug concentration for clinical pharmacy provided by an embodiment of the present invention, the method comprising the following steps:

[0041] Step S001, obtaining spectral data and image graphs of plasma samples at different drug concentrations.

[0042] It should be noted that the hyperspectral imager will collect the reflection spectra of the samples, which will reflect the chemical composition information in the plasma, especially the concentration information of the drugs in the plasma. In order to perform blood drug concentration detection in this embodiment, it is necessary to first obtain the spectral data of the plasma samples.

[0043] In order to implement the blood drug concentration detection method for clinical pharmacy proposed in this embodiment, it is first necessary to obtain the plasma sample spectral data. The specific process is as follows:

[0044] Prepare several plasma samples without adding drugs, add different concentrations of drugs to these samples to obtain plasma at several drug concentrations, use a hyperspectral imager to scan the plasma samples after adding drugs, and obtain spectral curve images of the samples. Each sample corresponds to a spectral curve, and each spectral curve corresponds to several images. Perform baseline correction, noise reduction, normalization and other preprocessing on the obtained spectral curves, and record the spectral curves after preprocessing as plasma sample spectral data.

[0045] At this point, the spectral data and image of the plasma sample are obtained through the above method.

[0046] Step S002, according to the shape characteristics of the peaks in the plasma sample spectral data, determine the waveform characteristic value of each peak, obtain the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determine the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, construct a histogram of the image corresponding to each peak, and determine the contribution of each peak according to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak.

[0047] Step 2.1, determining the waveform characteristic value of each peak according to the shape characteristics of the peaks in the plasma sample spectral data.

[0048] It should be noted that in spectral data, drug molecules will show reflection characteristics at different wavelengths, thereby showing the local maximum of the spectral curve, that is, the peak in the spectrum. These reflection characteristics are the key features of hyperspectral data for identifying drugs and detecting their concentrations. The width and depth of the peak in the spectrum can be used to obtain information on drug concentration. Generally, the wider the peak and the deeper the depth, the higher the drug concentration in the sample. At the same time, due to the purity and molecular structure of the drug, the peak will be symmetrical.

[0049] Specifically, the waveform characteristic value of each peak is determined according to the shape characteristics of each peak in the plasma sample spectral data.

[0050] As an embodiment, the specific calculation method of the waveform characteristic value of each peak is:

[0051] ;

[0052] In the formula, Indicates The waveform characteristic value of each peak; Indicates The depth of a peak, wherein the depth is the difference between the reflectivity corresponding to the peak and the minimum reflectivity closest to the peak; Indicates The width of a peak, which is the distance between the nearest minimum points on the left and right sides of the peak; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the left side of the peak and the horizontal axis; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the right side of the peak and the horizontal axis; Indicates taking the absolute value; Represents an exponential function with a natural constant as base.

[0053] It should be noted that each sample corresponds to a spectral curve, and each spectral curve has multiple peaks. The larger the value, the The greater the drug concentration in the sample corresponding to the peak, The larger the value, the The greater the drug concentration in the sample corresponding to the peak, Indicates The symmetry of the left and right sides of the wave peak, The smaller the value, the The more symmetrical the left and right sides of the wave peak are, the The larger the value.

[0054] Step 2.2, obtain the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, and determine the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration.

[0055] It should be noted that the wavelength range related to drug absorption and metabolism can be identified through the characteristics of the peaks, and the relationship between each peak and blood drug concentration can be quantified by analyzing the correlation between the peaks and blood drug concentration.

[0056] With drug concentration as the horizontal axis and peak characteristic value as the vertical axis, for the wavelength range of any peak, obtain several waveform characteristic values ​​of the spectral curve of the plasma sample at different drug concentrations within the wavelength range of the peak, and obtain several broken lines of drug concentration and peak characteristic changes based on the several waveform characteristic values. Use the least squares method to fit the several broken lines to obtain several fitted straight lines, and obtain the Pearson correlation coefficients between different peaks and blood drug concentrations. According to the number of plasma samples, the waveform characteristic values ​​of the peaks, the slope of the fitting straight line where the peaks are located, and the Pearson correlation coefficient between the peaks and blood drug concentrations, determine the correlation between each peak and blood drug concentration.

[0057] As an example, the specific calculation method of the correlation between the peak and the blood drug concentration is:

[0058] ;

[0059] In the formula, Indicates The correlation between the peaks and blood drug concentration; represents the number of plasma samples; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates The peaks correspond to the slopes of the fitted lines; Indicates the maximum slope of the fitted straight line; Indicates the minimum slope of the fitted straight line; Indicates Pearson correlation coefficient between each peak and blood drug concentration; Indicates taking the absolute value.

[0060] It should be noted that The larger the value, the The more synchronized the changes in the peaks and blood drug concentrations are, the greater the correlation between the two. Indicates The degree of change in the waveform characteristics of each peak due to changes in blood drug concentration. The larger the value, the When the blood drug concentration changes, the waveform characteristics of the first peak also change accordingly. The greater the correlation between a peak and blood drug concentration, Indicates The degree of change of the waveform characteristics of each peak among all peaks, The larger the value, the higher the The greater the trend of the waveform characteristics of the first peak changing with the change of blood drug concentration, the greater the degree of influence of blood drug concentration. The greater the correlation between the peak and the blood drug concentration.

[0061] Step 2.3, construct a histogram of the image graph corresponding to each peak, and determine the contribution of each peak based on the correlation between each peak and the blood drug concentration and the histogram characteristics of the image graph corresponding to each peak.

[0062] It should be noted that in order to measure the importance of a peak to the determination of blood drug concentration, it is necessary to obtain the contribution of each peak, which is used to represent the amount of information contained in the peak or the contribution to the model prediction performance. The amount of information can be represented by the information entropy of each peak. Peaks with higher information entropy contain richer information and have higher contributions.

[0063] Specifically, each wavelength in the spectral data corresponds to an image graph. According to the width of each peak, several image graphs are obtained, and a grayscale histogram of the image graph corresponding to each peak is constructed. Each bin in the grayscale histogram represents a pixel value range, and the height of the bin indicates the frequency of occurrence of the pixel value in the corresponding range. According to the number of pixel values ​​in the grayscale histogram of the image graph corresponding to each peak and the frequency of occurrence of each pixel value, the information entropy of each peak is determined. According to the correlation between each peak and the blood drug concentration and the information entropy of each peak, the contribution of each peak is determined.

[0064] As an embodiment, the specific calculation method of the contribution of each peak is:

[0065] ;

[0066] In the formula, Indicates The contribution of each peak; Indicates The correlation between the peaks and blood drug concentration; Indicates The information entropy of a peak.

[0067] It should be noted that Indicates The correlation between the peaks and blood drug concentration, The larger the value, the The greater the contribution of each peak, The larger the value, the The greater the information entropy of a peak, the greater the contribution of the peak.

[0068] So far, the contribution of the peak is obtained through the above method.

[0069] Step S003, obtaining the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, determining the importance of each peak to the blood drug concentration detection according to the contribution of each peak and the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, and according to the importance of the peak, reducing the dimension of the plasma sample spectral data, establishing a detection model, and detecting the blood drug concentration.

[0070] It should be noted that the amount of blood drug concentration information provided by the peak can be analyzed based on the contribution of the peak, the importance of the peak can be determined based on the contribution of the peak, and more important peaks can be selected based on the importance to improve the accuracy of the detection model.

[0071] Step 3.1, obtaining the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, and determining the importance of each peak to the blood drug concentration detection according to the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak.

[0072] It should be noted that peaks with higher contribution contain rich information and can better reflect the characteristics of plasma sample spectral data. In blood drug concentration detection, if a peak in the plasma sample spectral data can provide more information about drug absorption or metabolism, then the peak has higher importance. Therefore, the variance of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak is used to determine the importance of each peak to blood drug concentration detection.

[0073] The difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak are obtained, and the importance of each peak to the blood drug concentration detection is determined according to the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak.

[0074] As an example, the specific calculation method of the importance of each peak to the blood drug concentration detection is:

[0075] ;

[0076] In the formula, Indicates The importance of each peak to blood drug concentration detection; Indicates The contribution of each peak; Indicates The variance of the waveform characteristic values ​​corresponding to the peaks is used to represent the difference characteristics of the waveform characteristic values.

[0077] It should be noted that The larger the value, the The first peak reflects more drug information, then The more important a peak is to blood drug concentration detection, The larger the value, the The more information each peak contains, the The more important the peak is to the blood drug concentration detection.

[0078] Step 3.2, according to the importance of the peaks, the spectral data of the plasma samples are reduced in dimension, and a detection model is established to detect the blood drug concentration.

[0079] It should be noted that dimensionality reduction of data can identify and retain peaks with higher importance, better remove irrelevant information, and help improve the accuracy and stability of blood drug detection models. According to the importance of each peak, targeted dimensionality reduction of peaks with higher importance can reduce the amount of calculation and storage space requirements and improve efficiency. At the same time, by retaining peaks with higher importance, the data can be made more representative.

[0080] Firstly, principal component analysis is used to reduce the dimensionality of the plasma sample spectral data. The importance of each peak is used as a weight, and the weighted peak data is used to construct a covariance matrix. The eigenvectors and eigenvalues ​​are calculated from the covariance matrix. The appropriate principal components are selected according to the size of the eigenvalues ​​and the variance ratio. The plasma sample spectral data is projected onto the selected appropriate principal components to achieve dimensionality reduction of the plasma sample spectral data.

[0081] Then, the sample spectral data after dimension reduction were analyzed using the partial least squares regression algorithm, and a linear regression model was established to detect the blood drug concentration.

[0082] It should be noted that the partial least squares regression algorithm is an existing technology and will not be described in detail here.

[0083] Through the above steps, the detection of blood drug concentration is completed.

[0084] Second, see Figure 2 , which shows a blood drug concentration detection system for clinical pharmacy, the system includes the following modules:

[0085] A data acquisition module, used to obtain spectral data and image graphs of plasma samples at different drug concentrations;

[0086] A peak contribution determination module is used to determine the waveform characteristic value of each peak according to the shape characteristics of the peak in the plasma sample spectral data, obtain the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determine the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, construct a histogram of the image corresponding to each peak, and determine the contribution of each peak according to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak;

[0087] The blood drug concentration detection module is used to obtain the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, determine the importance of each peak to the blood drug concentration detection according to the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, and reduce the dimension of the plasma sample spectral data according to the importance of the peak, and establish a detection model to detect the blood drug concentration.

[0088] Third, see Figure 3 The present invention also proposes a blood drug concentration detection device for clinical pharmacy, which includes a blood drug concentration detection device and a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method described in steps S001 to S003 are implemented.

[0089] This embodiment determines the waveform characteristic value of each peak through the shape characteristics of the peak of the plasma sample spectral data, accurately reflects the shape characteristics of each peak in the spectral data, obtains the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determines the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, reflects the changing characteristics of the peak with the blood drug concentration, and constructs a histogram of the image graph corresponding to each peak. According to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image graph corresponding to each peak, the contribution of each peak is determined. According to the contribution of each peak, the importance of each peak to the blood drug concentration detection is determined, which helps to screen out the bands related to the blood drug concentration, reduce the complexity of data processing, and improve the accuracy and reliability of blood drug concentration determination. According to the importance of the peak, data dimension reduction is performed, a detection model is established, and the blood drug concentration is detected, which can effectively remove redundant information, reduce the amount of calculation and storage space requirements, and improve the system operation efficiency.

[0090] It should be noted that the The model is only used to represent negative correlation and constrain the output of the model to be in In the specific implementation, it can be replaced by other models with the same purpose. This embodiment is only based on The model is described as an example without any specific limitation. is the input to the model.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting blood drug concentration for clinical pharmacy, characterized in that: The method comprises the following steps: Obtain spectral data and images of plasma samples at different drug concentrations; According to the shape characteristics of the peaks in the plasma sample spectral data, the waveform characteristic value of each peak is determined, and the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration are obtained. According to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, the correlation between each peak and the blood drug concentration is determined, and a histogram of the image corresponding to each peak is constructed. According to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak, the contribution of each peak is determined; Obtain the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, determine the importance of each peak to the blood drug concentration detection based on the contribution of each peak and the difference characteristics of the waveform eigenvalues ​​of all peaks contained in the wavelength range of each peak, reduce the dimension of the plasma sample spectral data based on the importance of the peak, establish a detection model, and detect the blood drug concentration.

2. A method for detecting blood drug concentration for clinical pharmacy according to claim 1, characterized in that: The specific method of determining the waveform characteristic value of each peak includes: ； In the formula, Indicates The waveform characteristic value of each peak; Indicates The depth of a peak, wherein the depth is the difference between the reflectivity corresponding to the peak and the minimum reflectivity closest to the peak; Indicates The width of a peak, which is the distance between the nearest minimum points on the left and right sides of the peak; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the left side of the peak and the horizontal axis; In the spectral data, The peak and The area of ​​the closed region formed by the minimum point on the right side of the peak and the horizontal axis; Indicates taking the absolute value; Represents an exponential function with a natural constant as base.

3. A method for detecting blood drug concentration for clinical pharmacy according to claim 1, characterized in that: The specific method for determining the correlation between each peak and the blood drug concentration includes: A line graph was drawn with drug concentration as the horizontal axis and peak characteristic value as the vertical axis to obtain a changing line of drug concentration and peak characteristics. The line was fitted using the least squares method to obtain several fitted straight lines, and the Pearson correlation coefficients between different peaks and blood drug concentrations were obtained. The correlation between each peak and blood drug concentration was determined based on the number of plasma samples, the waveform characteristic value of the peak, the slope of the fitting straight line where the peak was located, and the Pearson correlation coefficient between the peak and blood drug concentration.

4. A method for detecting blood drug concentration for clinical pharmacy according to claim 3, characterized in that: The specific calculation method of the correlation between each peak and blood drug concentration is: ； In the formula, Indicates The correlation between the peaks and blood drug concentration; represents the number of plasma samples; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates Blood sample No. The waveform characteristic value of each peak; Indicates The peaks correspond to the slopes of the fitted lines; Indicates the maximum slope of the fitted straight line; Indicates the minimum slope of the fitted straight line; Indicates Pearson correlation coefficient between each peak and blood drug concentration; Indicates taking the absolute value.

5. A method for detecting blood drug concentration for clinical pharmacy according to claim 1, characterized in that: The specific method of determining the contribution of each peak is as follows: According to several image images corresponding to each peak, a grayscale histogram of the image image corresponding to each peak is constructed. According to the number of pixel values ​​in the grayscale histogram of the image image corresponding to each peak and the frequency of occurrence of each pixel value, the information entropy of each peak is determined. According to the correlation between each peak and the blood drug concentration and the information entropy of each peak, the contribution of each peak is determined.

6. A method for detecting blood drug concentration for clinical pharmacy according to claim 5, characterized in that: The specific calculation method of the contribution of each peak is: According to The correlation between the first peak and the blood drug concentration and the The information entropy of the peaks determines the The contribution of the first peak The correlation between the first peak and the blood drug concentration, The information entropy of the first peak is The contribution of each peak is directly proportional.

7. A method for detecting blood drug concentration for clinical pharmacy according to claim 1, characterized in that: The specific method for determining the importance of each peak to the blood drug concentration detection includes: The importance of each peak to blood drug concentration detection is determined based on the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak. The contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak are directly proportional to the importance of each peak to blood drug concentration detection.

8. A method for detecting blood drug concentration for clinical pharmacy according to claim 1, characterized in that: According to the importance of the peaks, the spectral data of the plasma samples are reduced in dimension, and a detection model is established to detect the blood drug concentration, including the following specific methods: The plasma sample spectral data is reduced in dimension using the principal component analysis method, the importance of each peak is used as a weight, a covariance matrix is ​​constructed using the weighted peak data, eigenvectors and eigenvalues ​​are calculated from the covariance matrix, a suitable principal component is selected according to the size of the eigenvalue and the variance ratio, and the plasma sample spectral data is projected onto the selected suitable principal component to achieve dimensionality reduction of the plasma sample spectral data; The sample spectral data after dimension reduction were analyzed using the partial least squares regression algorithm, and a linear regression model was established to detect the blood drug concentration.

9. A blood drug concentration detection system for clinical pharmacy, using a blood drug concentration detection method for clinical pharmacy as claimed in any one of claims 1 to 8, characterized in that: The system includes the following modules: A data acquisition module, used to obtain spectral data and image graphs of plasma samples at different drug concentrations; A peak contribution determination module is used to determine the waveform characteristic value of each peak according to the shape characteristics of the peak in the plasma sample spectral data, obtain the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, determine the correlation between each peak and the blood drug concentration according to the waveform characteristic value of the peak, the number of samples, the changing trend and correlation characteristics of the waveform characteristic value with the drug concentration, construct a histogram of the image corresponding to each peak, and determine the contribution of each peak according to the correlation between each peak and the blood drug concentration and the histogram distribution characteristics of the image corresponding to each peak; The blood drug concentration detection module is used to obtain the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, determine the importance of each peak to the blood drug concentration detection according to the contribution of each peak and the difference characteristics of the waveform characteristic values ​​of all peaks contained in the wavelength range of each peak, and reduce the dimension of the plasma sample spectral data according to the importance of the peak, and establish a detection model to detect the blood drug concentration.

10. A blood drug concentration detection device for clinical pharmacy, comprising a blood drug concentration detection device, a memory contained in the blood drug concentration detection device, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images