Method and system for evaluating state of isolated organ through pH value change of perfusate

Abstract

The invention discloses a method and system for evaluating the state of an isolated organ through the pH value change of perfusate, and belongs to the technical field of biological detection.The method comprises the steps that pH value data are collected through a pH electrode arranged in a sampling pool at a perfusate inlet and outlet, and the pH electrode is immersed into constant-temperature perfusate for continuous sampling after being subjected to three-point calibration; carrying out abnormal value identification and replacement on the collected pH value data, and calculating statistical parameters and first-order and second-order derivatives of the pH value difference of the inlet and the outlet after moving average smoothing processing; calculating characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient based on the processed pH value data, and obtaining evaluation indexes after standardization and principal component analysis; and dynamically monitoring the evaluation index through a sliding time window, calculating short-term and long-term change rates of the evaluation index, and triggering an early warning signal when the change rates are lower than a preset threshold value. Through the scheme of the invention, a unified evaluation framework can be established, and the evaluation reliability is improved.

Description

Technical Field

[0001] The present application relates to the field of biological detection, and in particular to a method and system for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid. Background Art

[0002] Organ transplantation is an important means of treating end-stage organ failure in modern medicine, and the quality of transplanted organs directly affects the success rate of the operation and the quality of life of the recipient patient after surgery. During the process of organ acquisition and transplantation, ex vivo organs will undergo cold ischemia and reperfusion, which may cause different degrees of organ damage. The currently commonly used ex vivo organ evaluation method in clinical practice mainly relies on the experience and judgment of organ acquisition doctors and some morphological indicators, such as organ surface color, elasticity, and vascular filling. However, this evaluation method is highly subjective and cannot accurately reflect the functional status of the organ. At the same time, traditional biochemical index detection methods often require a long detection cycle, and some index detection processes will cause irreversible damage to the organ. With the development of technologies such as organ preservation fluids and perfusion devices, the preservation time of ex vivo organs has been extended, but the corresponding organ function evaluation technology is relatively lagging behind. Traditional evaluation technology has many limitations: first, it is impossible to achieve real-time monitoring of the functional status of organs; second, the evaluation indicators are single and it is difficult to fully reflect the functional status of organs; third, some detection methods are traumatic and may affect the quality of organ transplantation; finally, the lack of standardized evaluation processes and judgment criteria leads to poor comparability and repeatability of evaluation results.

[0003] In response to the above problems, researchers have developed a variety of new organ function assessment technologies. For example, by collecting metabolites in organ preservation fluid for mass spectrometry analysis, an organ function assessment model is established; fluorescence imaging technology is used to monitor changes in ATP content in organs in real time; near-infrared spectroscopy technology is used to non-invasively detect the oxygenation status of organ tissues. These new assessment technologies have overcome the shortcomings of traditional methods to a certain extent and can reflect the functional status of organs more objectively and comprehensively. However, these improved technologies still face many challenges in practical applications: first, the assessment equipment is bulky and inconvenient to use at the organ procurement site; second, the detection process requires professional technicians to operate, which increases labor costs; third, the repeatability and stability of some detection methods need to be improved, and the test results are easily affected by external environmental factors. At the same time, there are differences in the functional characteristics and evaluation requirements of different types of organs, and existing technologies are difficult to meet personalized evaluation needs. In addition, the interpretation and judgment criteria of the evaluation results have not yet been unified, and the results of different evaluation technologies are difficult to compare horizontally. Therefore, the development of a portable, reliable, and standardized in vitro organ function assessment method is of great significance for improving the success rate of organ transplantation surgery.

[0004] Therefore, there is an urgent need for a technical solution to establish a unified evaluation framework and improve evaluation reliability. Summary of the invention

[0005] In order to solve the deficiencies of the prior art, the present application discloses a method and system for evaluating the state of an isolated organ by changing the pH value of a perfusion fluid. The present application solves the technical problems of the prior art, such as the complicated operation and interference to the organ during the sampling process.

[0006] The embodiment of the present application discloses a method for evaluating the state of an ex vivo organ by changing the pH value of a perfusion fluid, comprising: collecting pH value data by a pH electrode disposed in a sampling pool at the inlet and outlet of the perfusion fluid, the pH electrode being immersed in a constant temperature perfusion fluid for continuous sampling after three-point calibration; identifying and replacing outliers in the collected pH value data, and calculating statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives after moving average smoothing processing; calculating characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient based on the processed pH value data, and obtaining evaluation indicators after standardization and principal component analysis; dynamically monitoring the evaluation indicators through a sliding time window, calculating their short-term and long-term change rates, and triggering a warning signal when the change rate is lower than a preset threshold.

[0007] In a possible implementation, pH value data is collected by a pH electrode in a sampling pool disposed at the inlet and outlet of the perfusion fluid, and the pH electrode is immersed in a constant temperature perfusion fluid for continuous sampling after three-point calibration, including: a first pH electrode and a second pH electrode are respectively disposed in sampling pools at the inlet and outlet of the organ through which the perfusion fluid flows, and the temperature of the perfusion fluid in the sampling pool is kept constant; the pH electrode is calibrated by a three-point calibration method, and after the calibration is completed, the sensitive head of the pH electrode is completely immersed in the perfusion fluid; and pH value data at the inlet and outlet are continuously collected at preset time intervals to obtain a pH value data set.

[0008] In a possible implementation, outliers are identified and replaced for the collected pH data, and statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives are calculated after moving average smoothing, including: calculating the pH value change rate at adjacent time points, marking data points exceeding a preset threshold as outliers; replacing the outliers with the arithmetic mean of the two normal values ​​before and after, and performing moving average smoothing on the replaced data; calculating the mean and standard deviation of the inlet and outlet pH value difference; and using numerical differentiation to calculate the first-order derivative and second-order derivative of the pH data.

[0009] In a possible implementation, characteristic parameters including change rate, fluctuation intensity, deviation time and / or autocorrelation coefficient are calculated based on the processed pH value data, and evaluation indicators are obtained after standardization and principal component analysis, including: calculating the absolute mean of the pH change rate based on the first-order derivative; calculating the fluctuation intensity of the pH change based on the second-order derivative; counting the cumulative time of the pH value deviating from the reference interval; calculating the autocorrelation coefficient of the pH change; standardizing the characteristic parameters, and determining the weight coefficient of each characteristic parameter through principal component analysis; calculating the evaluation index based on the weight coefficient and the standardized characteristic parameters.

[0010] In a possible implementation, the evaluation index is dynamically monitored through a sliding time window, and its short-term and long-term change rates are calculated, and a warning signal is triggered when the change rate is lower than a preset threshold, including: setting a sliding time window to dynamically monitor the evaluation index; calculating the change rate of the evaluation index in a short-term time period; calculating the change rate of the evaluation index in a long-term time period; when the short-term change rate or the long-term change rate is lower than the preset threshold, a warning signal is triggered.

[0011] In a possible implementation, the characteristic parameters are standardized and the weight coefficient of each characteristic parameter is determined by principal component analysis, including: converting the characteristic parameters into a numerical value between zero and one according to a standardized formula, and organizing them into a matrix form of the number of samples multiplied by the number of characteristics; calculating the covariance matrix based on the standardized data matrix, and solving the covariance matrix characteristic equation by a decomposition method to obtain an eigenvalue sequence; using the proportion of the eigenvalue to the total as the weight coefficient, and judging by verifying the sample test, when the accuracy is insufficient, adding training samples to repeat the weight calculation.

[0012] In a possible implementation, the covariance matrix is ​​calculated based on the standardized data matrix, and the characteristic equation of the covariance matrix is ​​solved by the decomposition method to obtain the eigenvalue sequence, including: multiplying the standardized data matrix with its transposed matrix to obtain the covariance matrix; constructing a characteristic equation based on the covariance matrix; using the matrix decomposition method to solve the characteristic equation to obtain the eigenvalues; and arranging the eigenvalues ​​in descending order.

[0013] In a possible implementation, the proportion of the eigenvalue to the total is used as the weight coefficient, and the verification sample test is used to determine that when the accuracy is insufficient, additional training samples are added to repeat the weight calculation, including: calculating the proportion of each eigenvalue to the total eigenvalue; determining the proportion as the weight coefficient of the corresponding feature parameter; using the verification sample to verify the accuracy of the weight coefficient; when the accuracy does not reach a preset threshold, increasing the number of training samples and repeating the weight determination step.

[0014] The present application also discloses a system for evaluating the state of an ex vivo organ by changing the pH value of a perfusion fluid, which is applied to a method as described in any of the above embodiments, and includes: a sampling unit, a calculation unit, and a judgment unit; wherein the sampling unit is used to collect pH value data through a pH electrode disposed in a sampling pool at the inlet and outlet of the perfusion fluid, and the pH electrode is immersed in a constant temperature perfusion fluid for continuous sampling after a three-point calibration; the calculation unit is used to identify and replace abnormal values ​​of the collected pH value data, and calculate the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives after moving average smoothing; characteristic parameters including the rate of change, fluctuation intensity, deviation duration and / or autocorrelation coefficient are calculated based on the processed pH value data, and an evaluation index is obtained after standardization and principal component analysis; the judgment unit dynamically monitors the evaluation index through a sliding time window, calculates its short-term and long-term change rates, and triggers a warning signal when the change rate is lower than a preset threshold.

[0015] In a method and system for evaluating the state of an in vitro organ by changes in the pH value of the perfusion fluid as disclosed above, the embodiment of the present application avoids the numerical instability problem that may occur in the direct calculation of the characteristic value in the traditional principal component analysis by extracting multiple characteristic parameters and solving the characteristic equation by the QR decomposition method to determine the weight coefficient. In the process of calculating the weight coefficient, the iterative convergence threshold is combined with the dynamic increase of training samples to ensure the calculation accuracy and convergence within a limited number of iterations. The embodiment of the present application not only overcomes the shortcomings of the traditional method that requires multiple samplings, and realizes non-invasive and continuous monitoring of the metabolic state of the organ, but also improves the reliability of the evaluation results through the comprehensive analysis of multiple characteristic parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A schematic flow chart of a method for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid disclosed in an embodiment of the present application;

[0018] Figure 2 This is a working characteristic curve diagram of evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid disclosed in an embodiment of the present application;

[0019] Figure 3 A schematic diagram of the correlation of a method for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid disclosed in an embodiment of the present application;

[0020] Figure 4 A scatter plot of the change trend of an evaluation indicator over time and its correlation disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0021] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure unless otherwise specifically stated.

[0022] Those skilled in the art can understand that the terms "first", "second" and the like in the embodiments of the present disclosure are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor represent the necessary logical order between them. It should also be understood that in the embodiments of the present disclosure, "multiple" can refer to two or more, and "at least one" can refer to one, two or more. It should also be understood that for any component, data or structure mentioned in the embodiments of the present disclosure, in the absence of explicit limitation or contrary revelation given in the context, it can generally be understood as one or more. In addition, the term "and / or" in the present disclosure is only a description of the association relationship of the associated objects, indicating that there can be three relationships, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present disclosure generally indicates that the associated objects before and after are an "or" relationship. It should also be understood that the description of each embodiment in the present disclosure emphasizes the differences between the embodiments, and the same or similar parts can refer to each other. For the sake of brevity, they will not be repeated one by one.

[0023] At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present disclosure and its application or use. The techniques, methods and devices known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and devices should be considered part of the specification. It should be noted that similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0024] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0025] Figure 1 The present invention discloses a method for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid, which is a flow chart of the present invention.

[0026] It should be understood that the existing methods based on pH monitoring mainly include pH electrode method, pH sensitive dye method and pH microarray sensor method. The pH electrode method uses a glass electrode to measure the pH value of the solution. The selective anion interference coefficient of the electrode is 10 -13 , the measurement accuracy is ±0.002 pH units within the physiological pH range, but the electrode response time is long, usually requiring 15-30 seconds to reach a steady-state reading. The pH-sensitive dye method uses fluorescent dyes such as BCECF and SNARF to calculate the pH value by measuring the ratio of fluorescence intensity at different excitation or emission wavelengths. This method has a short response time, but the dye itself will affect cell metabolism and there is photobleaching. The pH microarray sensor is composed of multiple ion-sensitive field-effect transistors, which can realize the spatial distribution measurement of pH value, but the manufacturing process is complicated, and the sensor surface is easily adsorbed with proteins, which affects the measurement results. These methods only focus on the absolute change of pH value and fail to make full use of the dynamic information contained in the pH change. For example, the pH change rate reflects the generation and transport rate of metabolites, the pH fluctuation intensity reflects the metabolic homeostasis regulation ability, and the autocorrelation of pH change reflects the periodic characteristics of metabolic oscillation. This information is of great value for evaluating the metabolic state of organs. In addition, the existing methods also have shortcomings in data processing. They mainly use simple threshold judgment or linear regression analysis, and fail to establish a reasonable mathematical model to comprehensively analyze multiple characteristic parameters.

[0027] In contrast, this scheme continuously monitors the difference in pH values ​​between inlet and outlet, uses numerical differentiation methods to calculate the pH change rate and fluctuation intensity, introduces autocorrelation analysis to study the periodic characteristics of pH changes, determines the weights of each characteristic parameter through principal component analysis, and establishes a data processing and evaluation model. This scheme avoids the interference of sampling operations on organs, obtains continuous metabolic state information, and the data processing method takes into account multiple characteristic parameters of pH changes, making the evaluation results more comprehensive and reliable.

[0028] like Figure 1As shown, in step S101, pH value data is collected by a pH electrode in a sampling pool at the inlet and outlet of the perfusion fluid, and the pH electrode is immersed in a constant temperature perfusion fluid for continuous sampling after three-point calibration. The method includes: respectively arranging a first pH electrode and a second pH electrode in the sampling pool at the inlet and outlet of the organ through which the perfusion fluid flows, and keeping the temperature of the perfusion fluid in the sampling pool constant; calibrating the pH electrode by a three-point calibration method, and completely immersing the sensitive head of the pH electrode in the perfusion fluid after the calibration is completed; and continuously collecting pH value data at the inlet and outlet at a preset time interval to obtain a pH value data set.

[0029] In one embodiment, glass pH electrodes are respectively arranged at the inlet and outlet of the perfusate flowing through the organ, and the electrodes are connected to a pH meter. The pH meter is calibrated by a three-point calibration method, and a pH standard buffer solution (pH 4.01, 7.00, 9.21) is used for calibration. After the calibration is completed, the pH electrode is placed in a sampling pool at the inlet and outlet, and the volume of the sampling pool is 2 mL, ensuring that the electrode sensitive head is completely immersed in the perfusate. The sampling frequency is set to once every 30 seconds, and the recording is continuous for 60 minutes. During the recording process, the perfusate temperature is maintained at 37.0±0.2°C, and the flow rate is maintained at 9.8kPa.

[0030] At step S102, the collected pH data is subjected to outlier identification and replacement, and the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives are calculated after moving average smoothing. The process includes: calculating the pH value change rate at adjacent time points, marking the data points exceeding the preset threshold as outliers; replacing the outliers with the arithmetic mean of the two normal values ​​before and after, and performing moving average smoothing on the replaced data; calculating the mean and standard deviation of the inlet and outlet pH value difference; and using numerical differentiation to calculate the first-order and second-order derivatives of the pH data.

[0031] Preferably, when preprocessing the pH value data, the pH value change rate v(t) at two adjacent time points is first calculated: Where Δt is the sampling time interval (0.5 minutes). According to the in vitro organ experiment data, the pH value does not change more than 0.8 units / minute under normal physiological conditions, so the data points with v(t)>0.8 are marked as outliers. The marked outliers are replaced by the arithmetic mean of the two normal values ​​before and after. Then the three-point moving average method is used to smooth the data: Calculate the difference in pH between inlet and outlet: ΔpH(t) = pH out (t)-pH in (t), calculate the mean μ for the 60-minute ΔpH(t) data ΔpH and standard deviation σ ΔpH : Where N is the total number of sampling points.

[0032] Then, for the preprocessed data, the first-order derivative of pH change is calculated using the five-point numerical differentiation method: Compute the second-order derivative:

[0033] At step S103, characteristic parameters including change rate, fluctuation intensity, deviation time and / or autocorrelation coefficient are calculated based on the processed pH value data, and evaluation indicators are obtained after standardization and principal component analysis. Among them, it includes: calculating the absolute mean of pH change rate based on the first-order derivative; calculating the fluctuation intensity of pH change based on the second-order derivative; counting the cumulative time of pH value deviation from the reference interval; calculating the autocorrelation coefficient of pH change; standardizing the characteristic parameters, and determining the weight coefficient of each characteristic parameter through principal component analysis; calculating the evaluation index according to the weight coefficient and the standardized characteristic parameters.

[0034] Specifically, the characteristic parameter, the absolute mean value of the pH change rate V, is calculated based on the derivative value. pH : Divide the monitoring time into 120 time periods (0.5 minutes each), where t i is the midpoint of the i-th time period. The fluctuation intensity of pH change F pH : The cumulative time T of pH deviation from the baseline range dev :Set counter count=0 for each time point t:If |ΔpH(t)-μ ΔpH |>2σ ΔpH , then count=count+1, T dev = count × Δt. Autocorrelation coefficient of pH change R(τ): for τ from 0 to 30 minutes, with a step length of 0.5 minutes:

[0035] Furthermore, the characteristic parameters are standardized and the weight coefficients of each characteristic parameter are determined through principal component analysis, including: converting the characteristic parameters into a value between zero and one according to a standardized formula, and organizing them into a matrix form of the number of samples multiplied by the number of characteristics; calculating the covariance matrix based on the standardized data matrix, and solving the covariance matrix characteristic equation by a decomposition method to obtain an eigenvalue sequence; using the proportion of the eigenvalue to the total as the weight coefficient, and judging through verification sample testing, when the accuracy is insufficient, adding training samples to repeat the weight calculation.

[0036] Among them, the covariance matrix is ​​calculated according to the standardized data matrix, and the characteristic equation of the covariance matrix is ​​solved by the decomposition method to obtain the eigenvalue sequence, including: multiplying the standardized data matrix with its transposed matrix to obtain the covariance matrix; constructing a characteristic equation based on the covariance matrix; using the matrix decomposition method to solve the characteristic equation to obtain the eigenvalue; and arranging the eigenvalues ​​in descending order.

[0037] Among them, the proportion of the eigenvalue to the total is used as the weight coefficient, and the verification sample test is used to determine that when the accuracy is insufficient, the training samples are added to repeat the weight calculation, including: calculating the proportion of each eigenvalue to the total eigenvalue; determining the proportion as the weight coefficient of the corresponding feature parameter; using the verification sample to test the accuracy of the weight coefficient; when the accuracy does not reach the preset threshold, the number of training samples is increased and the weight determination step is repeated.

[0038] In an implementation scenario, the construction process of the evaluation indicator M is as follows:

[0039] Collect training sample data, sample the pH value of historical organ sample data for 60 minutes, and record the V of each sample. pH 、F pH , R(5) value, and at the same time record the known metabolic state of the sample (based on biochemical indicators such as ATP content and lactate content).

[0040] Data standardization is performed, and each feature parameter x is transformed as follows: where x min and x max are the minimum and maximum values ​​of the parameter in all samples respectively.

[0041] Principal component analysis determines the weights, constructs the feature matrix X (number of samples × number of features), and calculates the covariance matrix C = X T X, solve the characteristic equation |C-λI|=0 to obtain the eigenvalue λI, and the relative size of the eigenvalue is the weight of the corresponding characteristic parameter.

[0042] Verification and adjustment: Use a preset proportion of verification samples to test the accuracy of the evaluation results. If the accuracy is lower than the preset threshold, such as 99%, return to step 1 to increase the sample size. The weight coefficient obtained is: (pH change rate weight), (Volatility Intensity Weight), (Deviation duration weight), (Autocorrelation coefficient weight). Finally, the evaluation index is calculated: Parameters with a prime sign indicate normalized values.

[0043] In one embodiment, for each set of data, V is calculated according to the first and second steps.pH 、F pH , R(5) Four characteristic parameter values. The ATP content and lactate content of each sample were collected simultaneously. The ATP content was determined by fluorescence method, and the lactate content was determined by lactate oxidase colorimetry. The metabolic state was determined according to the ratio of ATP content to lactate content. An ATP / lactate ratio greater than 2.37 was determined as a normal metabolic state, a ratio between 1.52 and 2.37 was determined as a metabolic disorder state, and a ratio less than 1.52 was determined as a metabolic failure state.

[0044] The collected characteristic parameter data are organized into a 4×100 data matrix, where each column corresponds to a characteristic parameter and each row corresponds to a sample. For each characteristic parameter column in the data matrix, find the maximum and minimum values ​​of the column. The formula is used to standardize each data. pH For example, assuming that the minimum value of this parameter in 100 samples is 0.0186 and the maximum value is 0.0892, for a certain sample V pH The value is 0.0248, and its standardized value is (0.0248-0.0186) / (0.0892-0.0186) = 0.0877. The other three characteristic parameters are standardized by the same method to obtain a standardized 4×100 data matrix.

[0045] Based on the standardized data matrix, the covariance matrix is ​​constructed. The standardized data matrix is ​​denoted as X, and its transposed matrix is ​​denoted as X T , calculate the covariance matrix C = X by matrix multiplication T X. For the obtained 4×4 covariance matrix, construct the characteristic equation |C-λI|=0. Use QR decomposition method to solve the characteristic equation and obtain four eigenvalues ​​λ 1 , 2 , 3 , 4 The QR decomposition process is as follows: First, the covariance matrix C is expressed as C = QR, where Q is an orthogonal matrix and R is an upper triangular matrix. Then, by iterative calculation: C k+1 =R k Q k , where C k is the matrix after the kth iteration, R k and Q k C k The upper triangular matrix and orthogonal matrix obtained by QR decomposition. When |C k+1 -C k |Less than the preset threshold of 1.37×10 -6 Stop the iteration when C k+1The elements on the diagonal are the eigenvalues. Sort the obtained eigenvalues ​​from large to small, calculate the proportion of each eigenvalue in the total, and obtain four weight coefficients.

[0046] The 50 validation samples were subjected to feature parameter extraction and standardization according to the same method as above. The standardized feature parameters were multiplied and summed with the corresponding weight coefficients to obtain the evaluation index M value. The ATP content and lactate content of these samples were measured at the same time, and their metabolic state was determined according to the ATP / lactate ratio. The M value was compared with the metabolic state. If the consistency of the judgment results of the two was less than 95%, 100 training samples were added, and the above steps of feature parameter extraction, standardization, weight determination, etc. were repeated until the consistency of the judgment results of the validation samples reached more than 95%.

[0047] At step S104, the evaluation index is dynamically monitored through a sliding time window, and its short-term and long-term change rates are calculated. When the change rate is lower than a preset threshold, an early warning signal is triggered. This includes: setting a sliding time window to dynamically monitor the evaluation index; calculating the change rate of the evaluation index in a short-term period; calculating the change rate of the evaluation index in a long-term period; and triggering an early warning signal when the short-term change rate or the long-term change rate is lower than a preset threshold.

[0048] Preferably, a sliding window of 60 data points (corresponding to 30 minutes) is set, and the window moves one data point each time (0.5 minutes). For each window position, the short-term rate of change ΔM is calculated s :Get the last 10 points of the window (5 minutes), Calculate the long-term rate of change ΔM l : Use the data of the entire window (30 minutes), Then, set the warning trigger conditions. If any of the following conditions occurs, the warning will be triggered: 4 consecutive points (2 minutes) ΔM s <-0.1, 20 consecutive points (10 minutes) ΔM l <-0.05, M<0.5 for 6 consecutive points (3 minutes).

[0049] Next, we will take an isolated heart as an example. The perfusion solution used was a modified KH solution, whose composition (mmol / L) was: NaCl 118.0, KCl 4.7, KH 2 PO 4 1.2 MgSO 4 1.2 CaCl 2 2.5 NaHCO 3 25.0, glucose 11.0. The perfusion temperature was 37.0°C and the perfusion pressure was 9.8 kPa.

[0050] The execution steps are as follows: pH value acquisition, the inlet pH value is stable at 7.38±0.03, and the outlet pH value fluctuates between 7.20-7.35. After data preprocessing, we get: μΔpH=0.156, σΔpH=0.032. Feature parameter extraction, V pH =0.0248, F pH =0.0042, T dev =7.5 minutes, R(5) = 0.782. Evaluation index calculation, standardized characteristic parameter V' pH =0.825, F' pH =0.792, R'(5) = 0.856, weight coefficients w1 = 0.32, w2 = 0.18, w3 = 0.26, w4 = 0.24, calculated M = 0.847. Dynamic monitoring results: ΔM s The average value is 0.004 / min, ΔM l The mean value was 0.002 / min, and the M value was always higher than 0.8, indicating that the metabolic state of the heart was at a normal level.

[0051] Figure 2 A working characteristic curve diagram for evaluating the state of an in vitro organ by pH value changes disclosed in an embodiment of the present application. It includes a receiver operating characteristic (ROC) curve, a precision-recall (PR) curve, and a pH difference time series change curve. Among them, the abscissa of the ROC curve represents the false positive rate, ranging from 0 to 1.0, and the ordinate represents the true positive rate, ranging from 0 to 1.0. The ROC curve corresponding to the pH evaluation method rises convexly, with an area under the curve of 0.97, and the optimal working point is located at (0.09, 0.95); the ROC curve corresponding to the ATP / lactic acid ratio evaluation method also rises convexly, with an area under the curve of 0.81, and the optimal working point is located at (0.09, 0.78). The reference curve is a diagonal line, and the area below it is 0.5. From the geometric characteristics of the ROC curve, it can be seen that when the false positive rate is 0.2, the true positive rate of the pH evaluation method is higher than 0.95, and the true positive rate of the ATP / lactic acid ratio evaluation method is 0.75.

[0052] In the ROC curve comparison chart, the ROC curve of the pH method (black solid line) shows a steeper upward trend, reaching a TPR value of more than 0.95 at an FPR of 0.2, and the area under the curve (AUC) is 0.97; the ROC curve of the ATP / lactate ratio method (red solid line) rises more gently, with a TPR of about 0.75 at an FPR of 0.2, and an area under the curve of 0.81. The positions of the optimal working points (black dots) of the two methods show that the pH method can achieve a TPR of 0.95 at an FPR of 0.09, while the ATP / lactate ratio method can only achieve a TPR of 0.78 at the same FPR. The gray dotted line in the figure is the baseline for random classification (AUC = 0.5). According to the construction process of the M value, the high AUC value of the pH method comes from its fusion of V pH 、F pH , R(5) is the comprehensive evaluation capability of four characteristic parameters, which capture the rate of pH change, fluctuation intensity, deviation degree and autocorrelation characteristics respectively.

[0053] The PR curve comparison chart shows the change of recall rate at different precision levels. The PR curve (solid line) of the pH method can still maintain a high recall rate (>0.85) in the high precision range (>0.9), and the curve tends to be flat as a whole, indicating that the method has stable performance under different threshold settings. Preprocessing significantly reduces the impact of data noise on judgment. The morphological characteristics of the PR curve reflect that the pH method has a more accurate recognition ability for metabolic disorders with ATP / lactate ratios between 1.52 and 2.37.

[0054] The time series data comparison chart shows the dynamic change characteristics of the pH difference (ΔpH) within the 60-minute monitoring period. The black solid line represents the actual measured ΔpH value, which fluctuates within the range of 0.15±0.05, and the red dotted line marks the mean level of ΔpH (0.156). The fluctuation pattern shows that the ΔpH value under normal metabolic conditions exhibits quasi-periodic fluctuations around the mean.

[0055] Figure 3 This is a correlation diagram of a method for evaluating the state of an isolated organ by pH value changes disclosed in an embodiment of the present application. It shows a heat map of the correlation matrix between various indicators in the process of evaluating an isolated organ. Both the vertical and horizontal axes contain V pH 、F pH , μΔpH, σΔpH, R(5) and other pH evaluation characteristic parameters, as well as traditional biochemical indicators such as ATP content and lactate content. The colors in the heat map range from dark blue to dark red, reflecting the range of correlation coefficients from -0.95 to 0.95. Among them, V pHIt showed a strong positive correlation with the ATP / lactate ratio (r=0.82), indicating that the pH change rate can reflect the metabolic state of the organ; F pH It showed a moderate negative correlation with lactic acid content (r = -0.64), indicating that an increase in pH fluctuation intensity is often accompanied by lactic acid accumulation; T dev It showed a strong negative correlation with ATP content (r = -0.78), indicating that the ATP level tended to decrease when the time of pH deviation from the baseline interval increased; the correlation coefficient between R (5) and ATP / lactate ratio was 0.76, further supporting the association between pH change characteristics and metabolic state. The significance level is also marked in the heat map, where "*" indicates p < 0.05, "**" indicates p < 0.01, and "***" indicates p < 0.001. The dark red area on the matrix diagonal represents the complete correlation between each indicator and itself.

[0056] Figure 4 This is a scatter plot of the change trend of each evaluation index over time and its correlation disclosed in the embodiment of this application. The horizontal axis represents the monitoring time of 0-60 minutes, and the vertical axis includes the standardized V pH 、F pH , ATP / lactate ratio and other indicators. The color of the scattered points from light blue to dark red reflects the time process, and the size of the points is proportional to the M value. The horizontal dotted lines in the figure mark the key thresholds of ATP / lactate ratio 2.37 and 1.52, which correspond to the transition points from normal metabolism to disordered state and disorder to exhaustion state, respectively. The vertical shaded area marks the time period with a significance level of p<0.05. It can be seen from the scattered point distribution that V pH The value fluctuates in the range of 0.0186-0.0892, F pH The values ​​varied in the range of 0.0032-0.0058, and these changes had a significant temporal correlation with the trend of the ATP / lactate ratio. When the pH evaluation index approached the critical value, it was often accompanied by the transition of the ATP / lactate ratio to the corresponding threshold, reflecting the temporal consistency of the two evaluation methods in determining the organ status.

[0057] In summary, the present invention extracts V pH 、F pH , R(5) Multiple characteristic parameters. Among them, V pH Reflects the overall level of oxygen consumption and metabolite release in organ tissues; F pH The stability of metabolic activity was characterized; It indicates the degree to which the metabolic state deviates from the normal range; R(5) reflects the periodic variation characteristics of metabolic activity. The method of solving the characteristic equation to determine the weight coefficient by QR decomposition method avoids the numerical instability problem that may occur in the direct calculation of eigenvalues ​​in traditional principal component analysis. In the process of calculating the weight coefficient, the iterative convergence threshold of 1.37×10 -6 , which not only ensures the calculation accuracy, but also converges within a limited number of iterations. This scheme not only overcomes the shortcomings of the traditional method that requires multiple samplings, realizes non-invasive and continuous monitoring of the metabolic state of organs, but also improves the reliability of the evaluation results through the comprehensive analysis of multiple characteristic parameters. Compared with the existing technology, this scheme has made improvements in characteristic parameter extraction, data processing algorithms and evaluation model construction, making the evaluation results more objective and accurate.

[0058] Furthermore, an embodiment of the present application also discloses a system for evaluating the state of an ex vivo organ by changing the pH value of a perfusion fluid, which is applied to a method as described in any of the above embodiments, and includes: a sampling unit, a calculation unit, and a judgment unit; wherein the sampling unit is used to collect pH value data through a pH electrode arranged in a sampling pool at the inlet and outlet of the perfusion fluid, and the pH electrode is immersed in a constant temperature perfusion fluid for continuous sampling after a three-point calibration; the calculation unit is used to identify and replace abnormal values ​​of the collected pH value data, and calculate the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives after moving average smoothing processing; based on the processed pH value data, characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient are calculated, and evaluation indicators are obtained after standardization and principal component analysis; the judgment unit dynamically monitors the evaluation indicators through a sliding time window, calculates its short-term and long-term change rates, and triggers a warning signal when the change rate is lower than a preset threshold.

[0059] Furthermore, an embodiment of the present application also discloses a device for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid, comprising: a processor, a memory, and a system bus; the processor and the memory are connected via the system bus; the memory is used to store one or more programs, the one or more programs comprising instructions, which, when executed by the processor, cause the processor to execute any of the above methods.

[0060] Furthermore, an embodiment of the present application also discloses a computer program product, which, when executed on a terminal device, enables the terminal device to execute any of the above methods.

[0061] It can be known from the description of the above implementation mode that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0062] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0063] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0064] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating the state of an isolated organ by changes in the pH value of a perfusion fluid, characterized in that: include: The pH value data is collected by a pH electrode in a sampling pool at the inlet and outlet of the perfusion solution. After three-point calibration, the pH electrode is immersed in the constant temperature perfusion solution for continuous sampling. The collected pH data are subjected to outlier identification and replacement, and the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives are calculated after moving average smoothing. Based on the processed pH data, characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient are calculated, and evaluation indicators are obtained after standardization and principal component analysis; The evaluation indicators are dynamically monitored through a sliding time window, and their short-term and long-term change rates are calculated. When the change rate is lower than the preset threshold, a warning signal is triggered.

2. The method according to claim 1, characterized in that in, The pH value data is collected by the pH electrode in the sampling pool at the inlet and outlet of the perfusion fluid. After three-point calibration, the pH electrode is immersed in the constant temperature perfusion fluid for continuous sampling, including: The first pH electrode and the second pH electrode are respectively arranged in the sampling pool at the inlet and outlet of the perfusion fluid flowing through the organ, and the temperature of the perfusion fluid in the sampling pool is kept constant; The pH electrode was calibrated using the three-point calibration method. After the calibration was completed, the sensitive head of the pH electrode was completely immersed in the perfusion solution. The pH value data at the inlet and outlet are continuously collected at preset time intervals to obtain a pH value data set.

3. The method according to claim 1, characterized in that in, The collected pH data are identified and replaced with outliers, and the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives are calculated after moving average smoothing, including: Calculate the pH change rate at adjacent time points and mark the data points exceeding the preset threshold as outliers; The abnormal value is replaced by the arithmetic mean of the two normal values ​​before and after, and the replaced data is smoothed by moving average; Calculate the mean and standard deviation of the difference between the inlet and outlet pH values; Numerical differentiation was used to calculate the first and second derivatives of the pH data.

4. The method according to claim 1, characterized in that in, Characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient are calculated based on the processed pH data. Evaluation indicators are obtained after standardization and principal component analysis, including: The absolute mean of the pH change rate was calculated based on the first-order derivative; The fluctuation intensity of pH change is calculated based on the second-order derivative; Count the cumulative time that the pH value deviates from the baseline range; Calculate the autocorrelation coefficient of pH change; The characteristic parameters are standardized and the weight coefficient of each characteristic parameter is determined by principal component analysis; The evaluation index is calculated based on the weight coefficients and the normalized feature parameters.

5. The method according to claim 1, characterized in that in, Dynamically monitor the evaluation indicators through a sliding time window, calculate their short-term and long-term change rates, and trigger warning signals when the change rate is lower than the preset threshold, including: Set a sliding time window to dynamically monitor the evaluation indicators; Calculate the rate of change of the evaluation indicator over a short period of time; Calculate the rate of change of the evaluation indicator over a long period of time; When the short-term change rate or the long-term change rate is lower than the preset threshold, an early warning signal is triggered.

6. The method according to claim 4, characterized in that in, The characteristic parameters are standardized and the weight coefficients of each characteristic parameter are determined by principal component analysis, including: The characteristic parameters are converted into values ​​between zero and one according to the standardized formula, and are organized into a matrix form of the number of samples multiplied by the number of characteristics; The covariance matrix is ​​calculated based on the standardized data matrix, and the eigenvalue sequence is obtained by solving the characteristic equation of the covariance matrix through the decomposition method; The ratio of the eigenvalue to the total is used as the weight coefficient, and it is determined through verification sample testing. When the accuracy is insufficient, additional training samples are added to repeat the weight calculation.

7. The method according to claim 6, characterized in that in, The covariance matrix is ​​calculated based on the standardized data matrix, and the eigenvalue sequence is obtained by solving the covariance matrix characteristic equation through the decomposition method, including: Multiply the normalized data matrix with its transposed matrix to obtain the covariance matrix; Construct characteristic equation based on covariance matrix; The matrix decomposition method is used to solve the characteristic equation and obtain the eigenvalue; Sort the eigenvalues ​​in descending order.

8. The method according to claim 6, characterized in that in, The ratio of the eigenvalue to the total is used as the weight coefficient, and the test is determined by the verification sample. When the accuracy is insufficient, the training sample is added to repeat the weight calculation, including: Calculate the ratio of each eigenvalue to the total eigenvalues; Determine the proportion as a weight coefficient of the corresponding feature parameter; Use validation samples to test the accuracy of weight coefficients; When the accuracy does not reach the preset threshold, the number of training samples is increased and the weight determination step is repeated.

9. A system for evaluating the state of an isolated organ by changes in the pH value of a perfusate, applied to the method according to any one of claims 1 to 8, characterized in that: include: Sampling unit, calculation unit and judgment unit; wherein, The sampling unit is used to collect pH value data through a pH electrode arranged in a sampling pool at the inlet and outlet of the perfusion liquid. After three-point calibration, the pH electrode is immersed in the constant temperature perfusion liquid for continuous sampling; The calculation unit is used to identify and replace abnormal values ​​of the collected pH value data, calculate the statistical parameters of the inlet and outlet pH value difference and its first-order and second-order derivatives after moving average smoothing; calculate characteristic parameters including change rate, fluctuation intensity, deviation duration and / or autocorrelation coefficient based on the processed pH value data, and obtain evaluation indicators after standardization and principal component analysis; The judgment unit dynamically monitors the evaluation index through a sliding time window, calculates its short-term and long-term change rates, and triggers an early warning signal when the change rate is lower than a preset threshold.

Images