System and method for liver cirrhosis prediction and noninvasive diagnosis of autoimmune hepatitis

Abstract

The invention relates to a liver cirrhosis prediction and noninvasive diagnosis system and method for autoimmune hepatitis, and belongs to the technical field of medicine. The system comprises a data acquisition module, a data processing module and a model construction module. The data acquisition module is used for acquiring blood detection indexes of a patient, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyltranspeptidase (GGT), urea, creatinine, platelet count (PLT) and the like. The model building module is combined with blood detection data and clinical pathology data, a statistical analysis method such as binary Logistic regression is adopted, and a liver cirrhosis prediction model is trained and built. Based on four key blood detection indexes, the method can effectively predict the occurrence of liver cirrhosis of AIH patients, and especially provides a scientific reference for noninvasive diagnosis and curative effect monitoring for patients who cannot perform liver puncture or reject liver puncture.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of medical technology, and in particular relates to a system and method for predicting and non-invasively diagnosing autoimmune hepatitis cirrhosis. Background Art

[0002] Autoimmune liver disease is a disease whose etiology and pathogenesis are not yet clear. It may be caused by the combined effects of genetic, immune, environmental and other factors. According to the pathological type, autoimmune liver disease can be divided into autoimmune hepatitis (AIH), primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), among which AIH has a higher clinical incidence. Epidemiological data show that about 23% of AIH patients have cirrhosis. Cirrhosis can affect the liver, and if the condition is severe, it can affect all systems of the body, causing great harm to people's health.

[0003] Liver biopsy histopathological examination is the "gold standard" for diagnosing liver fibrosis, but because it is an invasive and invasive operation, patients are generally difficult to accept and its clinical application is limited. Therefore, it is urgent to find non-invasive and alternative methods for assessing the degree of liver cirrhosis. Establishing non-invasive indicators to monitor and evaluate the efficacy based on routine laboratory tests can avoid repeated liver punctures. Studies have reported non-invasive diagnostic models based on blood test indicators, such as the fibrosis index based on 4 factors (FIB4), aspartate aminotransferase alanine aminotransferase (AAR) and aspartate aminotransferase / platelet count (APRI), which have certain application value in the diagnosis of liver fibrosis. However, previous studies have mainly been applied to the diagnosis of liver fibrosis and cirrhosis in other liver diseases such as chronic hepatitis C, chronic hepatitis B and non-alcoholic steatohepatitis, and have also been shown to have significant sensitivity. However, there are few studies on models constructed based on blood test indicators in the diagnosis of autoimmune liver disease cirrhosis. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis.

[0005] The present invention is achieved by providing a system for predicting and non-invasively diagnosing cirrhosis of autoimmune hepatitis, the method comprising:

[0006] S101: Blood samples from patients with autoimmune hepatitis were selected for data collection, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyl transpeptidase (GGT), urea, creatinine, platelet count (PLT) and other blood indicators;

[0007] S102: According to the clinical pathological data of the patients, the patients were divided into cirrhosis group and non-cirrhosis group;

[0008] S103: Perform correlation analysis between blood test indicators and patients’ clinical pathological data to screen key predictive factors.

[0009] Further, the S101 specifically includes:

[0010] A total of 48 AIH patients were selected for modeling analysis, including 10 males and 38 females with an average age of 53.6±11.9 years old; 10 AIH patients were selected as test set samples for model testing, including 2 males and 8 females.

[0011] Further, the S102 specifically includes:

[0012] Liver tissue was obtained by puncture and immediately fixed in 10% neutral formaldehyde solution, routinely dehydrated, paraffin embedded, sliced, HE stained and mounted; the degree of liver tissue damage was expressed by fibrosis stage S; the Scheuer scoring standard was used for fibrosis staging, S0 was no fibrosis, S1 was fibrosis limited to the portal area, S2 was star-shaped fibrosis or PP fiber septa appeared around the portal area, and the liver lobule structure remained intact, S3 was fibrosis began to destroy the liver lobule structure, PC fiber septa appeared, and there was no obvious cirrhosis, S4 was suspected or definite cirrhosis; S0-S3 was defined as the non-cirrhosis group, and S4 was defined as the cirrhosis group.

[0013] Further, the S103 specifically includes:

[0014] The blood biochemical indicators tested include total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase AST, alanine aminotransferase ALT, alkaline phosphatase ALP, glutamyl transpeptidase GGT, urea, creatinine, etc.; routine blood tests include white blood cell count, neutrophil ratio, lymphocyte ratio, platelet count PLT; blood coagulation indicators include prothrombin time PT, activity, international normalized ratio, etc.; specific protein meter is used to detect serum IgG, IgM, IgA, C3 and C4 levels.

[0015] Further, the S104 specifically includes:

[0016] SPSS 16.0 software was used for statistical analysis and comparison of measurement data. t test was used for the data that did not conform to the normal distribution, which was represented by M (P25, P75). Mann Whitney rank sum test was used for the comparison of count data. 2Test; the key predictive factors for the difference between the cirrhosis group and the non-cirrhosis group were screened out, namely international normalized ratio, total bilirubin, direct bilirubin and IgA; the binary logistic regression method was used to combine the screened key predictive factors to construct a predictive diagnostic model; the receiver operating characteristic curve (ROC) was drawn and the area under the ROC curve was calculated to evaluate the efficacy of each indicator in diagnosing cirrhosis, and P<0.05 was considered statistically significant.

[0017] Another object of the present invention is to provide a system for predicting and non-invasively diagnosing cirrhosis of autoimmune hepatitis, comprising:

[0018] A data acquisition module is used to collect blood test indicators of patients, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyl transpeptidase (GGT), urea, creatinine, platelet count (PLT), international normalized ratio (INR), immunoglobulin A (IgA), etc.;

[0019] A data processing module is used to pre-process the blood test indicators, including data cleaning, standardization, and missing value processing to ensure data quality;

[0020] A model building module is used to train and build a liver cirrhosis prediction model based on blood test data and clinical pathology data. The model is built based on binary logistic regression analysis, and the prediction diagnosis formula is: logit (p) = 0.033 × INR-0.147 × total bilirubin (μmol / L) + 0.196 × direct bilirubin (μmol / L) + 0.588 × IgA (g / L)-2.494. When logit (p) ≥ 0.214, the prediction diagnosis is liver cirrhosis in patients with autoimmune hepatitis (AIH);

[0021] The result display module is used to display the prediction results of non-invasive diagnosis to doctors to help them make clinical decisions. The results are presented in the form of charts and text.

[0022] Another object of the present invention is a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis.

[0023] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis.

[0024] Another object of the present invention is to provide an information data processing terminal, which is used to implement a value evaluation system for the blood test indicators in constructing a prediction and non-invasive diagnosis model for cirrhosis in patients with autoimmune hepatitis.

[0025] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0026] First, the model established by the present invention based on four blood test indicators has a good predictive diagnostic value for liver cirrhosis in AIH patients, and provides a reference for non-invasive diagnosis and efficacy monitoring of liver cirrhosis in AIH patients who are unable to undergo liver biopsy due to contraindications or are afraid of liver biopsy.

[0027] Biochemistry, blood routine and coagulation tests are common clinical examination items, with convenient detection, easy clinical data, and wide application of related indicators. The present invention selects 22 indicators from biochemistry, blood routine and coagulation tests to construct a predictive diagnosis model for liver cirrhosis in AIH patients. Through preliminary comparative analysis, we found that among the 22 blood test indicators, the international normalized ratio, total bilirubin, direct bilirubin and IgA were statistically significant in the non-cirrhosis group and cirrhosis group of AIH patients, suggesting that these four indicators may be associated with whether AIH patients develop liver cirrhosis. Next, we further used the ROC curve method to analyze the efficacy of the four blood test indicators in the predictive diagnosis of liver cirrhosis in AIH patients. The results showed that the international normalized ratio, total bilirubin, direct bilirubin and IgA had a certain effect on the predictive diagnosis of liver cirrhosis in AIH patients.

[0028] The present invention combines the above four blood test indicators to construct a prediction and diagnosis model for liver cirrhosis in AIH patients. The results show that its AUC is 0.799, the sensitivity is 81.8%, and the specificity is 75.7%. Compared with FIB4 ((FIB4 = age (years) × AST (U / L) / [PLT (10 9 / L)×ALT(U / L) 1 / 2 ]), AUC: 0.698, sensitivity: 72.7%, specificity: 67.6%) were all improved, and the prediction and diagnosis of liver cirrhosis in AIH patients had a good effect. The prediction and diagnosis accuracy of the model was 81.2% using the prediction formula. Finally, the test set samples were selected to verify the constructed model, and the diagnostic accuracy of liver cirrhosis was 70.0%.

[0029] In summary, the model established based on the four blood test indicators in the present invention has a good predictive diagnostic value for liver cirrhosis in AIH patients, and provides a reference for non-invasive diagnosis and efficacy monitoring of liver cirrhosis in AIH patients who are unable to undergo liver biopsy due to contraindications or are afraid of liver biopsy.

[0030] Second, the technical solution of the present invention solves the accuracy and non-invasiveness problems of diagnosing cirrhosis of patients with autoimmune hepatitis (AIH) in the prior art in industrial applications. Traditional diagnostic methods usually rely on invasive detection methods such as liver biopsy, which have problems such as patient discomfort, risk of complications and complex operation. The present invention provides an AIH cirrhosis prediction and non-invasive diagnosis system based on blood test indicators. By collecting and processing the patient's blood test data, a binary logistic regression model with high prediction accuracy is constructed, realizing non-invasive diagnosis of cirrhosis.

[0031] Significant technological advances include:

[0032] 1. Non-invasive diagnosis: The present invention uses blood test indicators to build a prediction model, which reduces the reliance on invasive methods such as liver biopsy, and reduces the patient's pain and diagnostic risks.

[0033] 2. High accuracy and sensitivity: The diagnostic formula established through binary logistic regression analysis has high accuracy and sensitivity, which can effectively improve the early prediction and diagnosis of cirrhosis and avoid misdiagnosis and missed diagnosis.

[0034] 3. Convenience and practicality: The system can quickly process blood data and provide real-time diagnostic results, helping clinicians make scientific judgments and improving diagnostic efficiency and convenience.

[0035] Therefore, the present invention improves the accuracy and efficiency of AIH cirrhosis diagnosis while improving the patient experience, and has significant clinical application value and industrial promotion potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a structural block diagram of a liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis provided by an embodiment of the present invention;

[0037] Figure 2 It is a flow chart of a blood test data processing method for constructing a prediction and non-invasive diagnosis model for liver cirrhosis in patients with autoimmune hepatitis provided by an embodiment of the present invention;

[0038] Figure 3 The present invention provides a receiver operating characteristic curve of a model for diagnosing liver cirrhosis in AIH patients by constructing a model combining international normalized ratio, total bilirubin, direct bilirubin and IgA indicators. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0040] like Figure 1 As shown, the embodiment of the present invention provides a system for predicting and non-invasively diagnosing cirrhosis of autoimmune hepatitis, including:

[0041] A data collection module is used to collect blood test indicators of patients, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase AST, alanine aminotransferase ALT, alkaline phosphatase ALP, glutamyl transpeptidase GGT, urea, creatinine, platelet count PLT, etc.;

[0042] The model building module is used to train and build a cirrhosis prediction model based on blood test data and clinical pathology data.

[0043] The data acquisition module provided by the embodiment of the present invention further includes a routine blood test submodule, a coagulation test submodule and a specific protein test submodule, which are respectively used to detect white blood cell count, neutrophil ratio, prothrombin time PT, IgG, IgM, IgA, C3, C4 and other related indicators.

[0044] The data processing module of the present invention is used to perform statistical analysis using SPSS 16.0 software, including t-test, Mann-Whitney rank sum test and chi-square test. 2 The prediction and diagnosis model of liver cirrhosis was further constructed by combining the binary logistic regression method with key predictive factors.

[0045] The embodiment of the present invention provides a model building module including a machine learning algorithm module, which is used to use multivariate regression or other machine learning algorithms to model and analyze blood test indicators and pathological data to generate a non-invasive diagnostic model for predicting cirrhosis.

[0046] The system provided by the embodiment of the present invention further includes a receiver operating characteristic (ROC) curve generation module, which is used to draw the ROC curve and calculate the area under the curve to evaluate the diagnostic efficacy of blood test indicators in the diagnosis of liver cirrhosis.

[0047] The system provided by the embodiment of the present invention also includes a remote data transmission module, which is used to transmit the patient's blood test results from the detection equipment to a remote server, and use the diagnostic model on the server to analyze and generate a prediction report to provide real-time cirrhosis risk assessment.

[0048] like Figure 2As shown, an embodiment of the present invention provides a blood test data processing method for constructing a prediction and non-invasive diagnosis model for liver cirrhosis in patients with autoimmune hepatitis, comprising the following steps:

[0049] S101: Blood samples from patients with autoimmune hepatitis were selected for data collection, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyl transpeptidase (GGT), urea, creatinine, platelet count (PLT) and other blood indicators;

[0050] S102: According to the clinical pathological data of the patients, the patients were divided into cirrhosis group and non-cirrhosis group;

[0051] S103: Perform correlation analysis between blood test indicators and patients’ clinical pathological data to screen key predictive factors.

[0052] The embodiment of the present invention provides that step S103 further includes using multivariate regression analysis, combined with blood test indicators, to construct a liver cirrhosis prediction model to identify key predictive factors related to liver fibrosis and liver cirrhosis;

[0053] The step S103 further includes constructing a statistical model for non-invasive diagnosis of liver cirrhosis by combining blood test indicators and pathological data through a binary logistic regression model.

[0054] The method provided by the embodiment of the present invention also includes:

[0055] S104: Evaluate the diagnostic performance of the blood test data, specifically including drawing a receiver operating characteristic curve (ROC) and calculating the area under the curve (AUC) to evaluate the diagnostic performance in the prediction of liver cirrhosis.

[0056] Further including:

[0057] S105: The patient's blood test results and analysis data are transmitted to the cloud server using a remote data transmission system, and a liver cirrhosis prediction report is generated for the patient based on the diagnostic model in the server to provide real-time risk assessment.

[0058] The S101 specifically includes:

[0059] A total of 48 AIH patients were selected for modeling analysis, including 10 males and 38 females with an average age of 53.6±11.9 years old; 10 AIH patients were selected as test set samples for model testing, including 2 males and 8 females.

[0060] The S102 specifically includes:

[0061] Liver tissue was obtained by puncture and immediately fixed in 10% neutral formaldehyde solution, routinely dehydrated, paraffin embedded, sliced, HE stained and mounted; the degree of liver tissue damage was expressed by fibrosis stage S; the Scheuer scoring standard was used for fibrosis staging, S0 was no fibrosis, S1 was fibrosis limited to the portal area, S2 was star-shaped fibrosis or PP fiber septa appeared around the portal area, and the liver lobule structure remained intact, S3 was fibrosis began to destroy the liver lobule structure, PC fiber septa appeared, and there was no obvious cirrhosis, S4 was suspected or definite cirrhosis; S0-S3 was defined as the non-cirrhosis group, and S4 was defined as the cirrhosis group.

[0062] The S103 specifically includes:

[0063] Blood biochemical indicators were tested, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase AST, alanine aminotransferase ALT, alkaline phosphatase ALP, glutamyl transpeptidase GGT, urea, creatinine, etc.; routine blood tests were performed using Sysmex XN3000, including white blood cell count, neutrophil ratio, lymphocyte ratio, platelet count PLT; blood coagulation tests were performed, including prothrombin time PT, activity, international normalized ratio, etc.; serum IgG, IgM, IgA, C3 and C4 levels were detected using Siemens specific protein analyzer.

[0064] The S104 specifically includes:

[0065] SPSS 16.0 software was used for statistical analysis and comparison of measurement data. t test was used for the data that did not conform to the normal distribution, which was represented by M (P25, P75). Mann Whitney rank sum test was used for the comparison of count data. 2 Test; screen the key predictive indicators of blood tests, and use binary logistic regression method combined with key predictive factors to construct a predictive diagnostic model; draw the receiver operating characteristic curve (ROC) and calculate the area under the ROC curve to evaluate the efficacy of each indicator in diagnosing liver cirrhosis. P<0.05 was considered statistically significant.

[0066] An embodiment of the present invention provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis.

[0067] An embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis.

[0068] An embodiment of the present invention provides an information data processing terminal, which is used to implement a value evaluation system of the blood test index in constructing a prediction and non-invasive diagnosis model for liver cirrhosis in patients with autoimmune hepatitis.

[0069] 2. Relevant evidence of the technical effects obtained by the embodiments of the present invention.

[0070] 1 Results

[0071] 1.1 Comparison of basic information and blood test indicators between non-cirrhotic and cirrhotic AIH patients

[0072] Among the 48 AIH patients, 37 (77.1%) were non-cirrhotic and 11 (22.9%) were cirrhotic by liver histopathological examination. The detailed basic information and blood test index results are shown in Table 1.

[0073] Table 1. Comparison of basic information and blood test indicators between the two groups of AIH patients

[0074]

[0075]

[0076] Note: * indicates P<0.05, the difference is statistically significant.

[0077] 1.2 Screening of blood test indicators as key indicators for non-invasive diagnosis of liver cirrhosis in AIH patients

[0078] As shown in Table 1, among the 22 blood test indicators, the international normalized ratio in hemagglutination test, total bilirubin, direct bilirubin and immune index IgA in biochemical test were found to be statistically significant between the cirrhosis group and the non-cirrhosis group in AIH patients (P<0.05), suggesting that the international normalized ratio, total bilirubin, direct bilirubin and IgA may be used for the prediction and diagnosis of cirrhosis in AIH patients. Therefore, the above four test indicators were initially selected as the key test indicators for the prediction and diagnosis of cirrhosis in AIH patients.

[0079] 1.3 Construction of a non-invasive diagnostic model for liver cirrhosis in AIH patients based on key test indicators

[0080] ROC curve analysis was further used to calculate the predictive diagnosis of the above four key test indicators for distinguishing cirrhosis from non-cirrhosis in AIH patients. As shown in Table 2, the AUCs of international normalized ratio, total bilirubin, direct bilirubin and IgA were 0.865, 0.713, 0.763 and 0.717, respectively, the sensitivities were 90.9, 100.0, 100.0 and 72.7%, respectively, and the specificities were 73.0, 37.8, 48.6 and 64.9%, respectively.

[0081] Table 2. Efficacy of the four key screening test indicators in diagnosing liver cirrhosis in AIH patients

[0082]

[0083] The four blood test indicators obtained by screening were further used to construct a non-invasive diagnosis model for liver cirrhosis in AIH patients using binary logistic regression analysis. Statistical analysis data showed that ( Figure 1 The predictive diagnostic formula is: l logit(p)=0.033×INR-0.147×total bilirubin (μmol / L)+0.196×direct bilirubin (μmol / L)+0.588×IgA(g / L)-2.494, AUC is 0.799 (95%CI:0.6420.955), sensitivity is 81.8%, specificity is 75.7%, P value is 0.003, the difference is statistically significant. When logit(p)≥0.214, the accuracy of predicting liver cirrhosis in AIH patients is 81.2%.

[0084] 1.4 Screening of test set AIH patients to verify the diagnostic efficacy of the constructed model

[0085] In order to verify the actual effectiveness of the non-invasive diagnostic model constructed by the present invention in diagnosing cirrhosis in AIH patients, the present invention again screened 7 AIH non-cirrhotic patients and 3 AIH cirrhotic patients as test set clinical samples to verify the prediction diagnosis formula and cutoff value. The general information of AIH patients in the test set is shown in Table 3. According to the international normalized ratio, total bilirubin, direct bilirubin and IgA of each patient, the logit (p) value corresponding to the AIH patient is calculated in combination with the above prediction formula. If logit (p) ≥ 0.214, the AIH patient is diagnosed with cirrhosis; if logit (p) < 0.214, the AIH patient is diagnosed as not having cirrhosis. The test set results of the present invention show that among the 7 AIH non-cirrhotic patients, 6 were correctly judged and 1 was wrongly judged; among the 3 AIH cirrhotic patients, 1 was correctly judged and 2 were wrongly judged; the prediction accuracy of the joint prediction model in the test set clinical samples was 70%. Therefore, the diagnosis model of liver cirrhosis in AIH patients constructed by combining international normalized ratio, total bilirubin, direct bilirubin and IgA also showed a high diagnostic accuracy in the clinical samples of the test set.

[0086] Table 3. General information of 10 AIH patients in the test set

[0087]

[0088]

[0089] 1.5 Evaluation of the efficacy of existing models in diagnosing cirrhosis in AIH patients

[0090] It has been reported in the literature that: FIB4 (FIB4 = age (years) × AST (U / L) / [PLT (10 9 / L)×ALT(U / L) 1 / 2 ]), AAR(AAR=AST(U / L) / ALT(U / L) and APRI(APRI=[AST / (AST-ULN)×100] / PLT(10 9 / L), where ULN is the upper limit of normal) may have certain clinical value for the prediction and diagnosis of liver cirrhosis. As shown in Table 4, FIB4, AAR and APRI were calculated in 48 AIH patients in the non-cirrhotic group and the cirrhotic group, respectively. It was found that FIB4 had a statistically significant difference between the two groups (P<0.05), while AAR and APRI had no statistically significant difference between the two groups (P>0.05). Further analysis of FIB4, which had a statistically significant difference between the two groups, showed that its AUC for diagnosing liver cirrhosis in AIH patients was 0.698, and the sensitivity and specificity were 72.7 and 67.6%, respectively. Therefore, the model constructed by the present invention has a high diagnostic efficiency.

[0091] Table 4. Comparison of FIB4, AAR, and APRI in non-cirrhotic and cirrhotic AIH patients

[0092]

[0093] Note: * indicates P<0.05, the difference is statistically significant.

[0094] The liver cirrhosis prediction and non-invasive diagnosis system for autoimmune hepatitis of the present invention first selects samples for modeling analysis and model testing through step S101. In this step, a group of patients with autoimmune hepatitis with a clear diagnosis are selected as research subjects, and are divided into training sets and test sets by random sampling. The training set is used to establish a prediction model, while the test set is used to evaluate the accuracy and effectiveness of the model. The selection of samples must be sufficiently diverse, covering different stages of disease progression, to ensure the wide applicability of the model.

[0095] Step S102 involves a pathological histological examination of a liver biopsy, which is currently the "gold standard" for diagnosing cirrhosis. In this step, samples are obtained through a biopsy of liver tissue and a pathological evaluation is performed to determine the extent of liver fibrosis and cirrhosis. These pathological results will serve as an output reference for the model and will be compared with other non-invasive detection methods (such as blood indicators) to verify the accuracy of the model's predictions. Liver biopsy data will provide the model with actual disease status information and help determine key variables and indicators in the model.

[0096] Step S103 is a blood test, which aims to evaluate the correlation and diagnostic value of multiple blood indicators in the prediction of liver cirrhosis. Commonly used blood test indicators include liver function enzymes (such as alanine aminotransferase ALT, aspartate aminotransferase AST), bilirubin, albumin, platelet count, etc. These indicators can reflect the damage and functional state of the liver and are associated with the progression of liver fibrosis. By collecting blood samples from patients, the association between blood indicators and liver biopsy results is established and used as an important input variable of the model.

[0097] The statistical analysis of step S104 is the core of this method and is used to evaluate the effectiveness of blood test indicators in predicting and diagnosing cirrhosis. First, multivariate regression analysis or machine learning algorithms are used to determine which blood indicators are highly correlated with the progression of cirrhosis. Subsequently, the performance of the model is evaluated using statistical methods such as ROC curves, sensitivity, specificity, positive predictive value, and negative predictive value. This step aims to optimize the diagnostic ability of the model and ensure that it can achieve high-precision non-invasive diagnosis in clinical applications.

[0098] Based on the above steps, the model can not only predict whether patients with autoimmune hepatitis will develop cirrhosis, but also perform non-invasive diagnosis based on the results of blood tests. The advantage of this method is that it replaces liver biopsy with blood tests, greatly reducing the pain and risk of patients, and can achieve earlier prediction and intervention, thereby improving the prognosis of patients.

[0099] Ultimately, the evaluation and optimization of the model will ensure its stability and accuracy in different patient groups. Through validation and testing of multiple clinical samples, the robustness of the model will be further improved so that it can be widely used in clinical diagnosis and provide personalized disease management and treatment decision support for patients with autoimmune hepatitis.

[0100] The following are two specific application examples of the liver cirrhosis prediction and non-invasive diagnosis model built based on the blood test index in the industry:

[0101] Example 1: Application in medical diagnostic equipment

[0102] In the field of in vitro diagnostic (IVD) equipment, the model can be integrated into automated blood analyzers for routine examinations in hospitals or clinics. By inputting the patient's blood sample into the device, the device can quickly analyze various indicators in the blood based on the built-in algorithm model and automatically predict whether the patient is at risk of cirrhosis. This application has greatly improved the hospital's ability to diagnose cirrhosis early, allowing doctors to make accurate judgments based on non-invasive test results, avoiding the invasive operation of traditional liver biopsy, and improving diagnostic efficiency.

[0103] Example 2: Application in remote health management system

[0104] In telemedicine and health management systems, the model can be used to monitor disease progression in patients with autoimmune hepatitis. By connecting to home blood testing equipment, patients can regularly collect blood samples at home and transmit the data to a remote health management platform via the Internet. The platform uses the diagnostic model of the present invention for analysis, automatically generates health reports and provides prediction results for cirrhosis. This approach enables patients to obtain continuous health monitoring and timely risk warnings, reduces the need to go to the hospital for frequent testing, and improves the convenience and accuracy of disease management.

[0105] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A system for predicting and non-invasively diagnosing cirrhosis of autoimmune hepatitis, characterized in that: The system comprises: A data acquisition module is used to collect blood test indicators of patients, wherein the blood test indicators include total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyl transpeptidase (GGT), urea, creatinine, platelet count (PLT), international normalized ratio (INR), and immunoglobulin A (IgA); A data processing module is used to pre-process the blood test indicators, including data cleaning, standardization, and missing value processing to ensure data quality; A model building module is used to train and build a liver cirrhosis prediction model based on blood test data and clinical pathology data. The model is built based on binary logistic regression analysis, and the prediction diagnosis formula is: logit (p) = 0.033 × INR-0.147 × total bilirubin (μmol / L) + 0.196 × direct bilirubin (μmol / L) + 0.588 × IgA (g / L)-2.

494. When logit (p) ≥ 0.214, the prediction diagnosis is liver cirrhosis in patients with autoimmune hepatitis (AIH); The result display module is used to display the prediction results of non-invasive diagnosis to doctors to help them make clinical decisions. The results are presented in the form of charts and text.

2. The system according to claim 1, characterized in that The data acquisition module further includes a routine blood test submodule, a coagulation test submodule and a specific protein test submodule, which are respectively used to detect white blood cell count, neutrophil ratio, prothrombin time PT, IgG, IgM, IgA, C3, C4 and other related indicators.

3. The system according to claim 1, characterized in that The data processing module was statistically analyzed using SPSS 16.0 software, including t-test, Mann Whitney rank sum test and χ 2 The prediction and diagnosis model of liver cirrhosis was further constructed by combining the binary logistic regression method with key predictive factors.

4. The system according to claim 1, characterized in that The model building module includes a machine learning algorithm module, which is used to use multivariate regression or other machine learning algorithms to model and analyze blood test indicators and pathological data to generate a non-invasive diagnostic model for predicting cirrhosis.

5. The system according to claim 1, wherein: The system further includes a receiver operating characteristic (ROC) curve generation module for drawing a ROC curve and calculating the area under the curve to evaluate the diagnostic efficacy of blood test indicators in the diagnosis of liver cirrhosis.

6. The system according to claim 1, characterized in that The system also includes a remote data transmission module for transmitting the patient's blood test results from the testing device to a remote server, using the diagnostic model on the server to perform analysis and generate a prediction report, and providing real-time liver cirrhosis risk assessment.

7. A blood test data processing method for constructing a prediction and non-invasive diagnosis model for liver cirrhosis in patients with autoimmune hepatitis, characterized in that: The method comprises the following steps: S101: Blood samples from patients with autoimmune hepatitis were selected for data collection, including total protein, albumin, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), glutamyl transpeptidase (GGT), urea, creatinine, platelet count (PLT) and other blood indicators; S102: According to the collected clinical pathological data of the patients, the patients were divided into cirrhosis group and non-cirrhosis group; S103: Perform correlation analysis between blood test indicators and patients’ clinical pathological data to screen key predictive factors.

8. The blood test data processing method according to claim 7, characterized in that: The step S103 further includes using multivariate regression analysis, combined with blood test indicators, to construct a liver cirrhosis prediction model to identify key predictive factors related to liver fibrosis and liver cirrhosis; The step S103 further includes constructing a statistical model for non-invasive diagnosis of liver cirrhosis by combining blood test indicators and pathological data through a binary logistic regression model.

9. The blood test data processing method according to claim 7, characterized in that: The method further comprises: S104: Evaluate the diagnostic performance of the blood test data, specifically including drawing a receiver operating characteristic curve (ROC) and calculating the area under the curve (AUC) to evaluate the diagnostic performance in the prediction of liver cirrhosis.

10. The blood test data processing method according to claim 7, characterized in that: Further including: S105: The patient's blood test results and analysis data are transmitted to the cloud server using a remote data transmission system, and a liver cirrhosis prediction report is generated for the patient based on the diagnostic model in the server to provide real-time risk assessment.

Images