In-vitro detection sample data-based decompensated liver cirrhosis risk assessment method and related device

Abstract

The invention discloses an in-vitro sample detection data-based decompensated liver cirrhosis risk assessment method and a related device, and relates to the technical field of liver cirrhosis condition assessment. The method comprises the following steps: acquiring in-vitro detection sample data corresponding to a to-be-assessed blood sample, the in-vitro detection sample data comprises sCD146, prothrombin time, serum albumin and platelet count; performing data preprocessing on the in-vitro detection sample data, namely generating a serum albumin grade variable ALBgrade based on a preset grading threshold value according to the numerical value of the serum albumin; inputting the processed feature data into a pre-trained binary Logistic regression model, and outputting a decompensated liver cirrhosis risk value corresponding to the to-be-evaluated blood sample; the decompensated liver cirrhosis risk value is a prediction probability P output by the binary Logistic regression model. The method can effectively improve the evaluation accuracy of decompensated liver cirrhosis.

Description

Methods and related devices for risk assessment of decompensated cirrhosis based on in vitro test sample data Technical Field

[0001] This application relates to the field of liver cirrhosis disease assessment technology, and in particular to a method and related device for assessing the risk of decompensated liver cirrhosis based on in vitro test sample data. Background Technology

[0002] Currently, non-invasive methods used clinically for assessing cirrhosis mainly include routine biochemical and coagulation indicators, platelet and albumin levels, imaging examinations (such as ultrasound, CT, MRI, and liver / spleen elastography), and serological scoring models (such as FIB-4 and APRI). According to clinical staging, the assessment of decompensated cirrhosis is usually based on "evidence of cirrhosis" and the presence of serious portal hypertension-related complications such as ascites, esophageal and gastric variceal bleeding (EVB), hepatorenal syndrome (HRS), or hepatic encephalopathy (HE). Evidence of cirrhosis can be derived from imaging and / or endoscopy, and histology can be referenced when necessary; however, histology is not a necessary condition for assessing decompensated cirrhosis.

[0003] However, existing non-invasive indicators or scoring models are easily affected by different etiologies, inflammatory activities, cholestasis and comorbidities, and are still insufficient for decompensation-related stratification (such as identification of pre-decompensation subclinical forms, risk assessment of first decompensation, or differentiation of stable / unstable decompensation). There is an urgent need for more stable and scalable non-invasive biomarkers and combined models to improve the identification and stratification of decompensated cirrhosis. Summary of the Invention

[0004] The purpose of this application is to provide a method and related device for risk assessment of decompensated cirrhosis based on in vitro test sample data, which can effectively improve the accuracy of assessment of decompensated cirrhosis.

[0005] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for assessing the risk of decompensated cirrhosis based on in vitro test sample data, comprising: S01, acquiring in vitro test sample data corresponding to the blood sample to be assessed; the in vitro test sample data is a record of in vitro test indicator data corresponding to a single blood sample; the in vitro test sample data includes sCD146, prothrombin time, serum albumin, and platelet count; S02, performing data preprocessing on the in vitro test sample data to obtain processed feature data; the data preprocessing includes: generating a serum albumin grade variable ALBgrade based on the serum albumin value and a preset grading threshold; S03, inputting the processed feature data into a pre-trained binary logistic regression model to output the risk value of decompensated cirrhosis corresponding to the blood sample to be assessed; the risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model; wherein the independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

[0006] Optionally, the formula expression for the binary logistic regression model is: Where P is the predicted probability of decompensated cirrhosis corresponding to the blood sample to be evaluated; β0 is a constant term; β1, β2, β3, and β4 are the regression coefficients of sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time, respectively.

[0007] Optionally, before performing S03, the method further includes: constructing the binary logistic regression model, specifically including: obtaining an in vitro test sample dataset; the in vitro test sample dataset includes in vitro test sample data of subjects with decompensated cirrhosis and in vitro test sample data of subjects with compensated cirrhosis who have not experienced decompensation events; using whether the subjects in the in vitro test sample dataset have decompensated cirrhosis as the dependent variable, and using the final set of independent variables determined by a stepwise regression variable screening strategy as the independent variables, the maximum likelihood estimation method is used to fit β0, β1, β2, β3, and β4 in the binary logistic regression model; the final set of independent variables includes sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

[0008] Optionally, the in vitro test sample data includes sCD146, prothrombin time, serum albumin, platelet count, gender, age, alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, alkaline phosphatase, and gamma-glutamyl transferase; the binary logistic regression model is constructed by selecting the final set of independent variables from candidate independent variables through a stepwise regression variable screening strategy; the candidate independent variables are determined based on the in vitro test sample data; the stepwise regression variable screening strategy is a backtracking method, in each iteration, the significance test P-value of each candidate independent variable is obtained based on the Wald test, and the candidate independent variable with the largest P-value that is greater than the preset significance threshold α is removed in each round, and the model is refitted after removal; the iteration stops when the P-value of all candidate independent variables in the current model is not greater than the preset significance threshold α, and the final set of independent variables is obtained.

[0009] Optionally, the stepwise regression variable selection strategy specifically includes: S11, constructing and fitting a binary logistic regression model based on the candidate independent variables; S12, obtaining the significance test P-value of each candidate independent variable based on the Wald test in each iteration, and calculating the odds ratio and confidence interval of the candidate independent variable; S13, in each iteration, if there is a candidate independent variable with a significance test P-value greater than a preset significance threshold α, then the candidate independent variable with the largest significance test P-value greater than the preset significance threshold α is removed, resulting in updated candidate independent variables, and the binary logistic regression model is refitted based on the updated candidate independent variables; if there is no candidate independent variable with a significance test P-value greater than the preset significance threshold α, then the iteration is stopped, and the final set of independent variables is obtained.

[0010] Optionally, the formula for calculating the ratio is: In the formula, β i OR represents the regression coefficient corresponding to candidate independent variable i. i This represents the ratio of candidate independent variable i; the formula for solving the confidence interval is: In the formula, SE(βi) is the standard error of the candidate independent variable i.

[0011] Optionally, S02 specifically includes: removing abnormal sample records from the in vitro test sample data that exceed a preset value range, and for repeated test records of the same subject within a preset time window, retaining the most recent test record and removing the remaining repeated test records to obtain cleaned in vitro test sample data; based on the cleaned in vitro test sample data, determining whether there are missing items in the in vitro test sample data corresponding to the blood sample to be evaluated; when there are missing items, outputting a missing indicator message to indicate the missing test indicator, and outputting a preset invalid risk flag to terminate the calculation of the risk value of decompensated cirrhosis; when there are no missing items, obtaining a data with no missing items. The in vitro test sample data; Z-score standardization is performed on the continuous indicators in the in vitro test sample data without missing items; wherein, in the process of Z-score standardization of the continuous indicators, the same set of standardized parameters is used to process the continuous indicators when the binary logistic regression model is constructed and when the binary logistic regression model is applied, and the standardized parameters are calculated and fixed from the training data when the binary logistic regression model is constructed; the continuous indicators include sCD146, platelet count and prothrombin time; serum albumin is graded and assigned a grade variable ALBgrade.

[0012] Secondly, this application provides a device for assessing the risk of decompensated cirrhosis based on in vitro test sample data, used to implement the aforementioned method for assessing the risk of decompensated cirrhosis based on in vitro test sample data, comprising: a data acquisition module for acquiring in vitro test sample data corresponding to the blood sample to be assessed, wherein the in vitro test sample data includes sCD146, prothrombin time, serum albumin, and platelet count; a data preprocessing module for preprocessing the in vitro test sample data to obtain processed feature data; the data preprocessing includes: generating a serum albumin grade variable ALBgrade based on the serum albumin value and a preset grading threshold; and a prediction module for inputting the processed feature data into a pre-trained binary logistic regression model to output a risk value of decompensated cirrhosis corresponding to the blood sample to be assessed; wherein the risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model; wherein the independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

[0013] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for risk assessment of decompensated cirrhosis based on in vitro test sample data as described above.

[0014] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for assessing the risk of decompensated cirrhosis based on in vitro test sample data as described above.

[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method and related device for risk assessment of decompensated cirrhosis based on in vitro test sample data. By combining sCD146 with other indicators (platelet count, serum albumin grade variable ALBgrade and prothrombin time), the discriminative performance of decompensated cirrhosis can be significantly improved. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 is an application environment diagram of a risk assessment method for decompensated cirrhosis based on in vitro test sample data in one embodiment of this application; Figure 2 is a flowchart of a risk assessment method for decompensated cirrhosis based on in vitro test sample data provided in one embodiment of this application; Figure 3 is a nomogram constructed based on sCD146 concentration, platelet count, serum albumin grade variable ALBgrade, and prothrombin time provided in one embodiment of this application; Figure 4 is a schematic diagram of the receiver operating characteristic (ROC) curve of the nomogram combined assessment model predicting decompensated cirrhosis provided in one embodiment of this application; Figure 5 is a schematic diagram comparing the ROC curves of different non-invasive serological indicators and the nomogram combined assessment model predicting decompensated cirrhosis provided in one embodiment of this application; Figure 6 is a schematic diagram of the functional modules of a risk assessment device for decompensated cirrhosis based on in vitro test sample data provided in one embodiment of this application; Figure 7 is a schematic diagram of the structure of a computer device provided in one embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Cirrhosis is a common outcome of various chronic liver diseases. Chronic (persistent or recurrent) liver inflammation and necrosis from any cause can lead to diffuse liver fibrosis; based on this, the presence of pseudolobule formation and intrahepatic vascular structure / blood flow disturbances necessitates a diagnosis of cirrhosis. Decompensated cirrhosis (DC) refers to cirrhosis evidence found in clinical, laboratory, endoscopic, imaging, or histological examinations, coupled with the patient's history of serious portal hypertension-related complications such as clinical ascites, esophageal and gastric variceal bleeding, hepatorenal syndrome, or hepatic encephalopathy. Patients with decompensated cirrhosis have a poor prognosis; timely identification and intervention are crucial for improving outcomes.

[0020] Currently, clinical diagnosis of cirrhosis and its staging requires a comprehensive assessment of etiology, medical history, clinical manifestations and complications, laboratory indicators, and imaging and / or endoscopic results. Histological evidence may be considered when necessary. Existing risk assessment methods (such as routine biochemical indicators, liver / spleen elastography, and serological scoring models) are easily affected by different etiologies, inflammatory activity, or comorbidities. They are still insufficient for the identification and stratification of decompensation-related factors (such as identification of pre-decompensation subclinical forms, initial decompensation risk assessment, or differentiation of different decompensation states). There is an urgent need to develop more stable, accurate, and scalable risk assessment methods to improve the efficiency of identifying decompensated cirrhosis and reduce unnecessary invasive examinations. In addition, how to effectively combine novel biomarkers (such as sCD146) with routine clinical indicators and achieve accurate prediction through statistical modeling is a technical challenge. Specific difficulties include: (1) screening key indicators with independent predictive value; (2) handling the nonlinear relationship between continuous and ordinal variables; (3) ensuring the stability and universality of the model in different populations; and (4) transforming complex statistical results into clinically operable visualization tools.

[0021] Currently, risk assessment research on decompensated cirrhosis mainly focuses on laboratory indicators and imaging / elastography. While serological scoring models such as FIB-4 and APRI are widely used for screening and stratifying liver fibrosis or cirrhosis, they primarily reflect the degree of fibrosis and are easily influenced by factors such as inflammatory activity, thus their ability to identify and stratify decompensated cirrhosis remains limited. In recent years, some studies have begun to focus on the role of endothelial injury-related biomarkers (such as sCD146) in liver disease, finding them to be valuable in portal hypertension and microvascular complications. However, these biomarkers have not yet been systematically incorporated into assessment models. Furthermore, although logistic regression models are widely used in medical prediction, existing studies mostly employ empirical weighting or simple linear summation, failing to fully utilize data-driven weight estimation methods.

[0022] Therefore, this application provides a method and related device for risk assessment of decompensated cirrhosis based on in vitro test sample data. By combining sCD146 with conventional biochemical indicators and constructing a binary logistic regression and nomogram, it fills the gap in this field and provides a new approach for risk assessment of decompensated cirrhosis.

[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] The risk assessment method for decompensated cirrhosis based on in vitro test sample data provided in this application embodiment can be applied in the application environment shown in Figure 1. The terminal 102 communicates with the server 104 via a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated into the server 104, or placed in the cloud or on other servers.

[0025] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 104 can be implemented using a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.

[0026] In an exemplary embodiment, as shown in FIG2, a method for risk assessment of decompensated cirrhosis based on in vitro test sample data is provided. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is described using server 104 in FIG1 as an example, and includes the following steps S01 to S03. Specifically: S01, acquiring in vitro test sample data corresponding to the blood sample to be assessed, wherein the in vitro test sample data includes sCD146 (ng / mL), platelet count (10... 9 S02. Perform data preprocessing on the in vitro test sample data to obtain processed feature data; the data preprocessing includes: generating a serum albumin grade variable ALBgrade based on the serum albumin value and a preset grading threshold; S03. Input the processed feature data into a pre-trained binary logistic regression model to output the risk value of decompensated cirrhosis corresponding to the blood sample to be evaluated; the risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model; wherein, the independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

[0027] In this embodiment, a combined assessment method integrating serum sCD146 protein with traditional clinical examination indicators is proposed. sCD146 protein expression is significantly increased in the serum of patients with cirrhosis, especially in patients with decompensated cirrhosis, where the concentration of sCD146 is significantly higher than in patients with compensated cirrhosis. By combining sCD146 with routine biochemical indicators, this combined detection method was found to significantly improve the accuracy of assessment in decompensated cirrhosis.

[0028] Therefore, this embodiment builds upon the limitations of existing serological indicators in assessing decompensated cirrhosis, particularly the previous research findings on sCD146 protein as a potential assessment biomarker. Through analysis of extensive in vitro sample data, it was verified that combining sCD146 with other clinical indicators can effectively improve discriminative performance. By constructing nomograms and ROC curves, this application demonstrates the superiority of this method in assessing patients with decompensated cirrhosis.

[0029] Specifically, when performing steps S01-S03, the following can be done: Obtain in vitro test sample data of the subject to be evaluated through case collection or laboratory testing. Data types include: sex (male / female), soluble CD146 (sCD146) protein level, age (years), platelet count (PLT), alanine aminotransferase (ALT), unit (U / L), aspartate aminotransferase (AST), unit (U / L), total bilirubin (TBIL), unit (μmol / L), direct bilirubin (DBIL), unit (μmol / L), serum albumin (ALB), alkaline phosphatase (ALP), unit (U / L), gamma-glutamyl transferase (GGT), unit (U / L), and prothrombin time (PT). Serum albumin was classified into four concentration ranges: 1 (<30g / L), 2 (30–35g / L), 3 (35–40g / L), and 4 (≥40g / L). All of these data types are routine test items of great significance in the assessment of decompensated cirrhosis.

[0030] Then, the in vitro test sample data is preprocessed to obtain processed feature data, specifically including: 1) Data cleaning and deduplication: Abnormal sample records exceeding the preset value range are removed from the in vitro test sample data. For repeated test records of the same subject within a preset time window, the most recent test record is retained and the remaining repeated test records are removed to obtain the cleaned in vitro test sample data. The preset value range is used to remove abnormal records that are obviously caused by input errors, unit errors, interface transmission errors, or test failures. The preset value range is not equivalent to the clinical normal reference range. The preset value range can be determined based on the preset test range and / or preset logical constraints of the test item. When the test result exceeds the preset value range or does not meet the preset logical constraints, the corresponding sample record is judged as an abnormal record and removed.

[0031] 2) Missing data handling: Based on the cleaned in vitro test sample data, determine whether there are missing items in the in vitro test sample data corresponding to the blood sample to be evaluated; when there are missing items, output a missing information to indicate the missing test indicator, and output a preset invalid risk label to terminate the calculation of the risk value of decompensated cirrhosis; when there are no missing items, obtain in vitro test sample data without missing items.

[0032] 3) Perform Z-score standardization on the continuous indicators in the in vitro test sample data without missing items; wherein, in the process of Z-score standardization of the continuous indicators, the same set of standardization parameters is used to process the continuous indicators during model construction and model application, and the standardization parameters are calculated and fixed from the training data during model construction; the continuous indicators include sCD146, platelet count and prothrombin time; serum albumin is graded and assigned to obtain the rank variable ALBgrade.

[0033] 4) Serum albumin grading assignment: Serum albumin is graded and assigned a grade variable ALBgrade. For example, ALBgrade can be assigned values ​​in the following ranges: 1 represents <30g / L, 2 represents 30–35g / L, 3 represents 35–40g / L, and 4 represents ≥40g / L.

[0034] 5) Categorical variable coding: Categorical variables (such as gender) are coded numerically; ALBgrade is a grade assignment variable that is directly input into the model to meet the input requirements of the binary logistic regression model.

[0035] The above preprocessing steps improve model stability and predictive performance. The processed feature data are then input into a binary logistic regression model, using whether or not the patient has decompensated cirrhosis as the dependent variable and sCD146, PLT, ALBgrade, and PT as independent variables to obtain the predicted probability of decompensated cirrhosis. Specifically, the binary logistic regression model takes the following form: In constructing the binary logistic regression model, whether or not the patient has decompensated cirrhosis is used as the dependent variable Y, with decompensated cirrhosis denoted as 1 and compensated cirrhosis as 0. Clinical indicators such as sCD146, PLT, ALBgrade, and PT are used as independent variables. In the formula, P is the predicted probability of decompensated cirrhosis corresponding to the blood sample to be evaluated, β0 is a constant term, and β1, β2, β3, and β4 are the regression coefficients of sCD146, PLT, ALBgrade, and PT, respectively. Their magnitude and sign reflect the direction and intensity of the influence of the corresponding indicators on the risk of decompensated cirrhosis. ALBgrade is a rank variable obtained by assigning serum albumin grades.

[0036] The parameters of the binary logistic regression model are fitted to the training dataset using the maximum likelihood estimation method.

[0037] Specifically, a stepwise regression variable screening strategy can be adopted to eliminate variables that are not statistically significant or have limited contribution to prediction, and finally retain indicators such as sCD146, PLT, ALBgrade, and PT that have independent predictive value for decompensated cirrhosis, so as to construct a simple and stable optimal model.

[0038] Specifically, the stepwise regression variable selection strategy includes: S11, constructing and fitting a binary logistic regression model based on the candidate independent variables; S12, obtaining the significance test P-value of each candidate independent variable based on the Wald test in each iteration, and calculating the odds ratio and confidence interval of the candidate independent variable; S13, in each iteration, if there is a candidate independent variable with a significance test P-value greater than a preset significance threshold α, then the candidate independent variable with the largest P-value greater than the preset significance threshold α is removed, resulting in updated candidate independent variables, and the binary logistic regression model is refitted based on the updated candidate independent variables; if there is no candidate independent variable with a significance test P-value greater than the preset significance threshold α, then the iteration is stopped, and the final set of independent variables is obtained.

[0039] Compared with existing conventional non-invasive scoring models, the binary logistic regression model described in this embodiment has the following differences and improvements: (1) The endothelial injury-related biomarker sCD146 is introduced. Most existing models are mainly based on traditional biochemical or hematological indicators such as liver function, bilirubin, coagulation and platelets, without considering molecular markers related to endothelial activation / injury. This application incorporates sCD146 as a key independent variable into the model, which can reflect information related to portal hypertension and microvascular lesions, thereby improving the ability to identify decompensated events.

[0040] (2) Adopting a data-driven weight estimation method. Some existing scoring systems use empirical weighting or simple linear summation, making it difficult to accurately quantify the weights of different indicators. This application uses binary logistic regression to automatically learn the regression coefficients of each indicator from the data, achieving objective weight estimation based on population data, making the model more consistent with the distribution characteristics of actual disease incidence risk.

[0041] (3) Joint modeling of continuous variables and ordinal variables. In this application, continuous variables such as sCD146, PLT, and PT are included in the same regression model along with the graded ALB (ALBgrade). Nonlinear relationships are handled by transformations when necessary (such as taking logarithms, standardization, and piecewise assignment) to more precisely characterize the association between each indicator and the risk of decompensation, rather than simply using a single indicator threshold for judgment.

[0042] (4) Integration with nomogram construction. In this application, the Logistic model is not only used for statistical analysis, but also serves as the basis for subsequent nomogram construction. The regression coefficients can be directly mapped to the scoring scale in the nomogram, transforming the statistical results into a clinical visualization tool that is easy for clinicians to use quickly at the bedside.

[0043] In the specific analysis and prediction process, for any patient to be evaluated, their sCD146, PLT, ALBgrade, PT, and other indicators are first collected. These values ​​are then substituted into the aforementioned Logistic regression equation to calculate the linear predictive value η = β0 + β1(sCD146) + β2(PLT) + β3(ALBgrade) + β4(PT). This predictive value is then converted into the probability of decompensated cirrhosis using the Logistic function. .

[0044] The resulting P represents the predicted probability (risk value) that the patient has decompensated cirrhosis. Through the above method, the binary logistic regression model in this embodiment achieves the transformation from multi-indicator joint analysis to individualized risk prediction, providing quantitative risk assessment basis for clinical practice.

[0045] In one exemplary embodiment, during the model training phase, after obtaining the model output results, it is necessary to evaluate the variable effects. Specifically, this can be done as follows: using a binary logistic regression model, calculate the odds ratio (OR) and confidence interval (95% CI) for each variable (such as sCD146, PLT, etc.), and obtain the corresponding P-value based on the Wald test or likelihood ratio test to assess the statistical association (significance) between each variable and the occurrence of decompensated cirrhosis.

[0046] The specific calculation methods for the odds ratio and confidence interval are as follows: First, taking whether or not the patient has decompensated cirrhosis as the dependent variable and clinical indicators such as sCD146, PLT, ALBgrade, and PT as independent variables, the maximum likelihood estimation method is used to fit the parameters of the binary logistic regression model to obtain the regression coefficients β corresponding to each independent variable. i and its standard error SE(β) i For each independent variable, the odds ratio (OR) is calculated by taking the exponent of the regression coefficient, i.e.: In the formula, β i OR represents the regression coefficient corresponding to independent variable i. i This represents the ratio of the independent variable i.

[0047] Meanwhile, based on the Wald method, with β i Based on the standard error, calculate the 95% confidence interval for the regression coefficient of this variable: Then, by exponentiating both the upper and lower limits, we obtain the 95% confidence interval for the OR of this variable: The p-values ​​for each variable can be calculated based on the Wald test or likelihood ratio test to determine whether the statistical association between the variable and the occurrence of decompensated cirrhosis is significant. Using these methods, the strength and uncertainty of the influence of each indicator on the risk of decompensated cirrhosis can be quantitatively assessed.

[0048] Then, based on the regression analysis results, a binary logistic regression model with a better fit was selected for fitting, and a nomogram was further constructed, as shown in Figure 3. The nomogram visualizes the weights and corresponding scores of each in vitro detection indicator in the model, facilitating decision-making by clinicians based on the model output.

[0049] Based on the results of univariate and multivariate regression analyses, indicators closely related to the occurrence of decompensated cirrhosis and possessing independent predictive value were selected. The final fitting model selected in this application is a multivariate binary logistic regression model with prothrombin time (PT), platelet count (PLT), serum sCD146, and serum albumin grade (ALBgrade) as independent variables. Specifically, the dependent variable Y is whether the subject has decompensated cirrhosis, where decompensated cirrhosis is denoted as 1 and compensated cirrhosis as 0. PT, PLT, sCD146, and ALBgrade are used as independent variables. Logistic regression is performed in statistical software to obtain regression coefficients, and a nomogram model is constructed based on these coefficients, resulting in the following regression equation: logit(P) = ln(P / (1-P)) = β0 + β1(sCD146) + β2(PLT) + β3(ALBgrade) + β4(PT); where P is the predicted probability of decompensated cirrhosis corresponding to the blood sample to be evaluated, β0 is a constant term, and β1, β2, β3, and β4 are the regression coefficients of the corresponding independent variables; PT is input into the model in seconds, and PLT is input in 10... 9 / L is used as the unit input to the model, and sCD146 is input into the model with its actual measured concentration value (e.g., ng / mL).

[0050] As shown in Figure 4, the discriminative ability of the model was evaluated using the receiver operating characteristic (ROC) curve and its area under the curve (AUC). Among the candidate models, the model containing PT, ALBgrade, PLT, and sCD146 showed the best performance in terms of discriminative ability and stability, and was therefore selected as the final fitted model used for constructing the nomogram in this embodiment.

[0051] When constructing the nomogram, the regression coefficients of the binary logistic regression model are used as a basis, and the magnitudes of the coefficients of the independent variables are linearly mapped to the integral scale on the nomogram: the larger the absolute value of the coefficient, the higher the corresponding integral weight. For any patient to be evaluated, the integrals of PT, ALBgrade, PLT, and sCD146 are read on the corresponding axes of the nomogram, and the total integral is obtained by summing them. Then, the total integral is converted into the predicted probability of decompensated cirrhosis through the risk axis below the nomogram. In this way, the fitting results of the regression model are visualized as an intuitive scoring tool, which facilitates clinicians to make quick decisions and risk stratification based on the total score and corresponding probability.

[0052] Next, the AUC was calculated and the ROC curve was plotted to verify the model's effectiveness: As shown in Figure 4, the AUC (area under the curve) metric was used to evaluate the model's predictive ability. The closer the AUC value is to 1, the better the model's discriminative power. Simultaneously, the ROC curve was plotted to verify the model's sensitivity at different false positive rates (1-specificity), further evaluating its risk assessment efficacy.

[0053] Finally, the AUC was calculated and the ROC curve was plotted to verify the model's effectiveness. As shown in Figure 5, compared with other non-invasive serological indicators, the model's efficacy in assessing the risk of decompensated cirrhosis is superior to other traditional non-invasive serological indicators.

[0054] Based on the same inventive concept, this application also provides a device for assessing the risk of decompensated cirrhosis based on in vitro test sample data, which implements the aforementioned method for assessing the risk of decompensated cirrhosis based on in vitro test sample data. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the device for assessing the risk of decompensated cirrhosis based on in vitro test sample data provided below can be found in the limitations of the method for assessing the risk of decompensated cirrhosis based on in vitro test sample data described above, and will not be repeated here.

[0055] In an exemplary embodiment, as shown in FIG6, a device for assessing the risk of decompensated cirrhosis based on in vitro test sample data is provided to implement the aforementioned method for assessing the risk of decompensated cirrhosis based on in vitro test sample data. The device includes: a data acquisition module 601 for acquiring in vitro test sample data corresponding to the blood sample to be assessed, wherein the in vitro test sample data includes sCD146, prothrombin time, serum albumin, and platelet count; and a data preprocessing module 602 for preprocessing the in vitro test sample data to obtain processed feature data. The data preprocessing includes: root... Based on the serum albumin value and a preset grading threshold, a serum albumin grade variable ALBgrade is generated. A prediction module 603 inputs the processed feature data into a pre-trained binary logistic regression model and outputs the risk value of decompensated cirrhosis corresponding to the blood sample to be evaluated. The risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model. The independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

[0056] In an exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram is shown in Figure 7. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores predicted probabilities. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for risk assessment of decompensated cirrhosis based on in vitro test sample data.

[0057] Those skilled in the art will understand that the structure shown in Figure 7 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0058] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0059] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0060] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0061] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0062] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0063] In summary, this application has the following technical advantages: The proposed risk assessment method for decompensated cirrhosis (DC) introduces the endothelial injury-related biomarker sCD146 and uses a data-driven weighting estimation method to jointly model continuous and ordinal variables. It combines sCD146 protein with conventional biochemical indicators such as PT, ALB, and PLT to construct an assessment model. After inputting the above indicators into the target subjects, the method ultimately outputs the predicted probability P of decompensated cirrhosis, achieving non-invasive quantitative assessment and risk stratification. The nomogram and ROC curve demonstrate that this combined assessment method not only significantly improves the accuracy of assessing decompensated cirrhosis and the ability to identify decompensated events, but also avoids the use of invasive examinations, providing a more convenient and accurate tool for early assessment and clinical management of cirrhosis.

[0064] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for assessing the risk of decompensated cirrhosis based on in vitro test sample data, characterized in that, include: S01. Obtain in vitro test sample data corresponding to the blood sample to be evaluated; The in vitro test sample data is a record of in vitro test indicator data corresponding to a single blood sample; the in vitro test sample data includes sCD146, prothrombin time, serum albumin, and platelet count; S02, the in vitro test sample data is preprocessed to obtain processed feature data; the data preprocessing includes: generating a serum albumin grade variable ALBgrade based on the serum albumin value and a preset grading threshold; S03, the processed feature data is input into a pre-trained binary logistic regression model to output the risk value of decompensated cirrhosis corresponding to the blood sample to be evaluated; the risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model; wherein, the independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

2. The method for risk assessment of decompensated cirrhosis based on in vitro test sample data according to claim 1, characterized in that, The formula for the binary logistic regression model is as follows: Where P is the predicted probability of decompensated cirrhosis corresponding to the blood sample to be evaluated; β0 is a constant term; β1, β2, β3, and β4 are the regression coefficients of sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time, respectively.

3. The method for assessing the risk of decompensated cirrhosis based on in vitro test sample data according to claim 2, characterized in that, Before executing S03, the process also includes: constructing the binary logistic regression model, specifically including: obtaining an in vitro test sample dataset; the in vitro test sample dataset includes in vitro test sample data of subjects with decompensated cirrhosis and in vitro test sample data of subjects with compensated cirrhosis who have not experienced decompensation events; using whether the subjects in the in vitro test sample dataset have decompensated cirrhosis as the dependent variable, and using the final set of independent variables determined by a stepwise regression variable screening strategy as the independent variables, the maximum likelihood estimation method is used to fit β0, β1, β2, β3, and β4 in the binary logistic regression model; the final set of independent variables includes sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

4. The method for risk assessment of decompensated cirrhosis based on in vitro test sample data according to claim 3, characterized in that, The in vitro test sample data includes sCD146, prothrombin time, serum albumin, platelet count, gender, age, alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, alkaline phosphatase, and gamma-glutamyl transferase. The binary logistic regression model is constructed by selecting the final set of independent variables from candidate independent variables using a stepwise regression variable selection strategy. The candidate independent variables are determined based on the in vitro test sample data. The stepwise regression variable selection strategy is a backtracking method, in which the significance test P-value of each candidate independent variable is obtained based on the Wald test in each iteration, and the candidate independent variable with the largest P-value that is greater than the preset significance threshold α is removed in each round, and the model is refitted after removal. The iteration stops when the P-value of all candidate independent variables in the current model is not greater than the preset significance threshold α, and the final set of independent variables is obtained.

5. The method for risk assessment of decompensated cirrhosis based on in vitro test sample data according to claim 4, characterized in that, The stepwise regression variable selection strategy specifically includes: S11, constructing and fitting a binary logistic regression model based on the candidate independent variables; S12, obtaining the significance test P-value of each candidate independent variable based on the Wald test in each iteration, and calculating the odds ratio and confidence interval of the candidate independent variable; S13, in each iteration, if there is a candidate independent variable with a significance test P-value greater than a preset significance threshold α, then the candidate independent variable with the largest significance test P-value greater than the preset significance threshold α is removed, resulting in updated candidate independent variables, and the binary logistic regression model is refitted based on the updated candidate independent variables; if there is no candidate independent variable with a significance test P-value greater than the preset significance threshold α, then the iteration is stopped, and the final set of independent variables is obtained.

6. The method for risk assessment of decompensated cirrhosis based on in vitro test sample data according to claim 5, characterized in that, The formula for calculating the ratio is: In the formula, β i OR represents the regression coefficient corresponding to candidate independent variable i. i This represents the ratio of candidate independent variable i; the formula for solving the confidence interval is: In the formula, SE(βi) is the standard error of the candidate independent variable i.

7. The method for risk assessment of decompensated cirrhosis based on in vitro test sample data according to claim 6, characterized in that, S02 specifically includes: removing abnormal sample records from the in vitro test sample data that exceed the preset value range; for repeated test records of the same subject within a preset time window, retaining the most recent test record and removing the remaining repeated test records to obtain cleaned in vitro test sample data; based on the cleaned in vitro test sample data, determining whether there are missing items in the in vitro test sample data corresponding to the blood sample to be evaluated; when there are missing items, outputting a missing indicator message to indicate the missing test indicator, and outputting a preset invalid risk flag to terminate the calculation of the risk value of decompensated cirrhosis; when there are no missing items, obtaining the body sample data without missing items. External test sample data; Z-score standardization is performed on the continuous indicators in the in vitro test sample data without missing items; wherein, in the process of Z-score standardization of the continuous indicators, the same set of standardized parameters is used to process the continuous indicators when the binary logistic regression model is constructed and when it is applied, and the standardized parameters are calculated and fixed from the training data when the binary logistic regression model is constructed; the continuous indicators include sCD146, platelet count and prothrombin time; serum albumin is graded and assigned to obtain the rank variable ALBgrade.

8. A device for assessing the risk of decompensated cirrhosis based on in vitro test sample data, used to implement the method for assessing the risk of decompensated cirrhosis based on in vitro test sample data as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire in vitro test sample data corresponding to the blood sample to be evaluated. The in vitro test sample data includes sCD146, prothrombin time, serum albumin and platelet count. The data preprocessing module is used to preprocess the in vitro test sample data to obtain processed feature data; The data preprocessing includes: generating a serum albumin grade variable ALBgrade based on the serum albumin value and a preset grading threshold; a prediction module is used to input the processed feature data into a pre-trained binary logistic regression model and output the risk value of decompensated cirrhosis corresponding to the blood sample to be evaluated; the risk value of decompensated cirrhosis is the predicted probability P output by the binary logistic regression model; wherein, the independent variables of the binary logistic regression model include sCD146, platelet count, serum albumin grade variable ALBgrade, and prothrombin time.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement a method for assessing the risk of decompensated cirrhosis based on in vitro test sample data, as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements a method for assessing the risk of decompensated cirrhosis based on in vitro test sample data, as described in any one of claims 1-7.

Images