Diabetic nephropathy risk prediction method and system based on protein markers

Abstract

The invention relates to the field of medical informatics, and provides a diabetic nephropathy risk prediction method and system based on a protein marker. The method comprises the following steps: detecting the concentration of a protein marker in a plasma sample of a diabetic to obtain plasma protein marker concentration data; performing longitudinal correlation analysis on the protein markers according to the plasma protein marker concentration data to obtain the kidney disease related protein markers; performing feature screening on the kidney disease related protein marker combination to obtain a core protein marker; establishing a nephropathy risk prediction model through the clinical parameters of the diabetic patient and the core protein marker combination; and outputting a nephropathy risk prediction result of the patient to be detected through the nephropathy risk prediction model. The advanced intervention of the diabetic nephropathy is realized, and the prediction accuracy of the diabetic nephropathy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical informatics, and in particular to a method and system for predicting the risk of nephropathy in diabetic patients based on protein markers. Background Art

[0002] Diabetic nephropathy is one of the common complications of diabetes patients, which seriously affects the quality of life of patients and increases the risk of renal failure. Currently, the clinical assessment of the risk of diabetic nephropathy mainly relies on the detection of microalbumin in urine.

[0003] However, the existing methods for detecting urine microalbumin can only detect damage to the kidneys, and cannot provide early warning in the early stages of pathology or when there are no significant clinical symptoms. In addition, a single marker, such as urine protein, is easily interfered with by other factors, which can easily lead to false positive or false negative results and a low accuracy rate. Summary of the Invention

[0004] The present invention provides a method and system for predicting the risk of nephropathy in diabetic patients based on protein markers, so as to overcome the defects of the prior art.

[0005] The present invention provides a method for predicting the risk of nephropathy in diabetic patients based on protein markers, comprising: S1: Detect the concentration of protein markers in plasma samples of diabetic patients to obtain plasma protein marker concentration data; S2: performing longitudinal correlation analysis on the protein markers based on the plasma protein marker concentration data to obtain kidney disease-related protein markers; S3: performing feature screening on the combination of kidney disease-related protein markers to obtain core protein markers; S4: Establish a kidney disease risk prediction model based on the clinical parameters of diabetic patients and the core protein marker combination; S5: Outputting the renal disease risk prediction result of the patient to be tested through the renal disease risk prediction model.

[0006] According to a method for predicting the risk of nephropathy in diabetic patients based on protein markers provided by the present invention, step S1 further comprises: S11: Establish patient cohort; S12: Collect EDTA-anticoagulated fasting plasma samples from diabetic patients in the patient cohort and establish a standardized biospecimen library; S13: Detecting the concentration of the protein marker to be screened in the standardized biological sample library to obtain the plasma protein marker concentration data.

[0007] According to a method for predicting kidney disease risk in diabetic patients based on protein markers provided by the present invention, the patient cohort in step S11 includes multiple type 2 diabetic patients with healthy kidneys, where healthy kidneys include no diabetic nephropathy and no other kidney diseases.

[0008] According to a method for predicting the risk of nephropathy in diabetic patients based on protein markers provided by the present invention, step S2 further comprises: S21: matching the plasma protein marker concentration data with the renal disease status assessment results within a preset follow-up period to obtain a protein combination to be screened; S22: Performing statistical correlation analysis on the protein combination to be screened using a Cox proportional hazard model to obtain kidney disease-related protein markers.

[0009] According to a method for predicting the risk of kidney disease in diabetic patients based on protein markers provided by the present invention, in step S3, the characteristics of the kidney disease-related protein markers are screened using a Lasso regression algorithm.

[0010] According to a method for predicting the risk of nephropathy in diabetic patients based on protein markers provided by the present invention, step S4 further comprises: S41: integrating the core protein markers with clinical parameters to obtain a feature vector; S42: Based on the feature vector, the hybrid model is trained to obtain a kidney disease risk prediction model.

[0011] According to a method for predicting the risk of kidney disease in diabetic patients based on protein markers provided by the present invention, the core protein markers in step S41 include: KIM-1, NGAL, Clusterin, IL-6, FGF-23, TNF-α, and the clinical parameters in step S41 include: HbA1c data, blood pressure, and BMI data.

[0012] According to a method for predicting the risk of nephropathy in diabetic patients based on protein markers provided by the present invention, step S5 further comprises: S51: collecting an EDTA-anticoagulated fasting plasma sample from the patient to obtain a plasma sample to be tested; S52: Detecting the concentration of the core protein marker according to the plasma sample to be tested to obtain concentration data to be tested; collecting clinical parameter data of the patient to be tested to obtain clinical parameters to be tested; S53: Inputting the concentration data to be measured and the clinical parameters to be measured into the kidney disease risk prediction model to obtain a kidney disease risk prediction result.

[0013] According to a method for predicting kidney disease risk in diabetic patients based on protein markers provided by the present invention, the kidney disease risk prediction results include a kidney disease risk prediction score and a kidney disease risk prediction stratification result, specifically including: When the renal disease risk prediction score is RS < 15 points, the renal disease risk prediction stratification result is low risk; When the renal disease risk prediction score is 15 ≤ RS ≤ 28 points, the renal disease risk prediction stratification result is medium risk; When the renal disease risk prediction score is RS>28 points, the renal disease risk prediction stratification result is high risk.

[0014] The present invention also provides a system for predicting the risk of nephropathy in diabetic patients based on protein markers, comprising: Acquisition module: used to collect plasma protein marker concentration data of protein markers in plasma samples of diabetic patients; Analysis module: used to perform longitudinal correlation analysis on protein markers based on the plasma protein marker concentration data to obtain kidney disease-related protein markers; Screening module: used to perform feature screening on the combination of kidney disease-related protein markers to obtain core protein markers; Training module: used to establish a kidney disease risk prediction model based on the clinical parameters of diabetic patients and the core protein marker combination; The prediction module is configured as the kidney disease risk prediction model established by the training module, and is used to predict the kidney disease risk of the patient to be tested and obtain a kidney disease risk prediction result.

[0015] The present invention provides a method and system for predicting the risk of kidney disease in diabetic patients based on protein markers. By constructing a method for predicting the risk of kidney disease in diabetic patients based on protein markers, the transformation of diabetic nephropathy from passive treatment to active prevention is achieved. The present invention successfully constructs an intelligent risk prediction model integrating biomarkers and clinical parameters by establishing a standardized cohort and biological sample library including patients with healthy kidneys and combining protein detection technology and machine learning algorithms. The Cox proportional hazard model is used for longitudinal correlation analysis in the present invention, which fully explores the intrinsic correlation between protein markers and the timeline of disease development. The core predictive factors are accurately screened out by the Lasso regression algorithm, avoiding the dimensional disaster and overfitting problems in traditional methods. The screened protein markers are then organically integrated with clinical parameters to construct a hybrid prediction model that not only has the accuracy at the molecular level, but also maintains the practicality and operability of clinical application. The final prediction model can output quantitative risk scores and intuitive risk stratification results, providing clinicians with a decision support tool that is both scientifically rigorous and easy to use.

[0016] Overall, the protein marker-based method for predicting diabetic nephropathy risk in patients constructed by this invention, through a systematic technical solution design, achieves technological innovation throughout the entire process, from cohort construction, sample processing, marker screening, model training, to risk prediction, forming a complete technical system for early warning of diabetic nephropathy. By identifying high-risk patients early, this invention enables preemptive intervention for diabetic nephropathy, significantly reducing the rate of disease progression and the burden of medical expenses. This not only provides strong technical support for the precise prevention and personalized management of diabetic nephropathy, but also has important clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A schematic flow chart of a method for predicting the risk of kidney disease in diabetic patients based on protein markers provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a system for predicting the risk of kidney disease in diabetic patients based on protein markers provided in an embodiment of the present invention; Figure 3 Schematic diagram of the ROC curve of the diabetic nephropathy risk prediction model based on the method of the present invention provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments, and they should not be understood as limitations on the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0020] The following describes embodiments of the present invention with reference to the accompanying drawings.

[0021] like Figure 1 As shown, the present invention provides a method for predicting the risk of nephropathy in diabetic patients based on protein markers, comprising: S1: Detect the concentration of protein markers in plasma samples of diabetic patients to obtain plasma protein marker concentration data.

[0022] Wherein, step S1 further includes: S11: Establish a patient cohort.

[0023] The patient queue in step S11 includes a plurality of type 2 diabetic patients with healthy kidneys, where healthy kidneys include patients without diabetic nephropathy and other kidney diseases.

[0024] The present invention screens patients with type 2 diabetes who have no diabetic nephropathy or other kidney diseases at baseline to establish the above-mentioned patient cohort. That is, the patients enrolled must simultaneously meet the dual conditions of having been diagnosed with type 2 diabetes but having normal renal function. When establishing the patient cohort, the present invention conducts detailed medical screening, including glycated hemoglobin testing to confirm the diagnosis of diabetes, renal function biochemical index testing to exclude nephropathy, and imaging examination to exclude structural kidney disease. Finally, a research cohort of 1,000 eligible patients is established.

[0025] S12: Collect EDTA-anticoagulated fasting plasma samples from diabetic patients in the patient cohort and establish a standardized biospecimen library.

[0026] Furthermore, in step S12, the present invention collects EDTA-anticoagulant fasting plasma samples, wherein EDTA is used as an anticoagulant to chelate calcium ions in the blood to prevent blood coagulation, while maintaining the original structure and activity state of proteins in the plasma. The collected and separated plasma is immediately packaged into pre-labeled cryovials, and the subsequent establishment of a standardized biological sample library adopts a programmed cooling method, and the plasma samples are finally transferred to a -80°C ultra-low temperature refrigerator for storage.

[0027] S13: Detecting the concentration of the protein marker to be screened in the standardized biological sample library to obtain the plasma protein marker concentration data.

[0028] The present invention uses Olink PEA technology to detect the concentration of protein markers to be screened. Specifically, the present invention first obtains multiple plasma samples from the aforementioned biological sample library, each plasma sample requiring a detection volume of 50 microliters. Then, the Olink Explore 384 platform can simultaneously detect different protein markers related to renal fibrosis, inflammatory response, and vascular endothelial damage in the plasma. Finally, through steps including sample dilution, antibody incubation, washing, PCR amplification, and fluorescence detection, the relative protein concentration data expressed as NPX value is output.

[0029] S2: Based on the plasma protein marker concentration data, longitudinal correlation analysis is performed on the protein markers to obtain kidney disease-related protein markers.

[0030] Wherein, step S2 further includes: S21: Matching the plasma protein marker concentration data with the renal disease status assessment results within a preset follow-up period to obtain a protein combination to be screened.

[0031] Furthermore, step S21 aims to match the plasma protein marker concentration data with the renal disease status assessment results within a preset follow-up period, wherein the preset follow-up period, in a specific embodiment, refers to a ten-year observation period calculated from the collection of the baseline plasma sample, during which a renal disease status assessment is performed once a year, including a comprehensive assessment of indicators such as urine protein-creatinine ratio detection, serum creatinine determination, glomerular filtration rate calculation, and urine microalbumin detection, and then the renal disease status assessment results are output. The renal disease status assessment results are coded in a binary classification manner, where 0 indicates that diabetic nephropathy has not occurred according to the assessment of the indicators, and 1 indicates that diabetic nephropathy has occurred according to the assessment of the indicators. The diagnostic criteria for diabetic nephropathy are two consecutive tests with a urine protein-creatinine ratio greater than 300 mg / g or a decrease in glomerular filtration rate of more than 30% of the baseline value and lasting for more than three months. Finally, the baseline protein marker concentration data of each patient are matched one-to-one with the time and status of renal disease occurrence during the follow-up period to form a structured data set including the patient identification code, baseline protein concentration vector, follow-up time, and renal disease status assessment results.

[0032] S22: Performing statistical correlation analysis on the protein combination to be screened using a Cox proportional hazard model to obtain kidney disease-related protein markers.

[0033] Furthermore, the Cox proportional hazards model is used as a survival analysis method in the present invention to establish a semiparametric relationship between covariates and hazard rates. In this invention, the covariate X corresponds to the concentration values ​​of various protein markers, the endpoint event is the occurrence of diabetic nephropathy, and the time variable t is the time interval from baseline to the occurrence of nephropathy or the end of follow-up. Specifically, the data processing process of the Cox model first requires converting the protein combination to be screened into a survival analysis format. Each row of data contains the patient ID, survival time, event occurrence status, and all protein marker concentration values. Survival time is calculated as the difference in days from the baseline sampling date to the date of kidney disease diagnosis. For patients who do not develop kidney disease, the difference in days to the last follow-up date is calculated and the event status is marked as missing.

[0034] The specific algorithm process for subsequent statistical correlation analysis includes the construction and maximization of the partial likelihood function. When judging, the Z statistic obeys the standard normal distribution. When the corresponding p-value is less than 0.05, it indicates that the protein marker is statistically significantly associated with the development of kidney disease. In addition, the screening criteria for kidney disease-related protein markers must consider biological plausibility and clinical relevance in addition to statistical significance. In this example, the hazard ratio is used to represent the multiple change in the risk of kidney disease for each unit increase in the concentration of the protein marker. A hazard ratio greater than 1 indicates that an increase in the concentration of the protein marker increases the risk of kidney disease, and a hazard ratio less than 1 indicates that the protein marker has a protective effect. Finally, combining the above criteria, protein markers significantly associated with the development of diabetic nephropathy were screened.

[0035] S3: Perform feature screening on the kidney disease-related protein marker combination to obtain a core protein marker.

[0036] Wherein, in step S3, the characteristics of the kidney disease-related protein markers are screened by using the Lasso regression algorithm.

[0037] Furthermore, the present invention uses the Lasso regression algorithm to perform feature screening for kidney disease-related protein markers. In the Lasso linear regression algorithm, when the regularization coefficient value gradually increases, the regression coefficients of some features will be compressed to zero, thereby achieving the purpose of feature selection. Specifically, when the regularization parameter value that minimizes the validation error is obtained through cross-validation, it is used as the optimal parameter. After determining the optimal value, multiple iterations are performed, and eventually the regression coefficients of protein markers with weak correlation with diabetic nephropathy are compressed to zero, screening out six core protein markers, including: KIM-1 (kidney injury molecule-1), NGAL (neutrophil gelatinase-associated lipocalin), Clusterin, IL-6 (interleukin-6), FGF-23 (fibroblast growth factor-23), and TNF-α (tumor necrosis factor-α).

[0038] S4: Establish a kidney disease risk prediction model based on the clinical parameters of diabetic patients and the core protein marker combination.

[0039] Wherein, step S4 further includes: S41: Integrate the core protein markers with clinical parameters to obtain a feature vector.

[0040] Among them, the core protein markers in step S41 include: KIM-1, NGAL, Clusterin, IL-6, FGF-23, TNF-α, and the clinical parameters in step S41 include: HbA1c data, blood pressure, and BMI data.

[0041] Step S41 integrates core protein markers with clinical parameters to obtain a feature vector. Core protein markers include six plasma proteins: KIM-1, NGAL, Clusterin, IL-6, FGF-23, and TNF-α. Clinical parameters include three clinical indicators: HbA1c data, blood pressure, and BMI data. Specifically, during the data processing process, the six protein marker concentration values ​​for each patient are first standardized. For clinical parameters, HbA1c data is recorded as a percentage, blood pressure data needs to be converted to mean arterial pressure (MAP), and BMI data is used directly in kg / m². Finally, the standardized six protein marker concentration values ​​and the three clinical parameter values ​​are arranged in a fixed order to form a 9-dimensional feature vector.

[0042] S42: Based on the feature vector, the hybrid model is trained to obtain a kidney disease risk prediction model.

[0043] Furthermore, step S42 trains the hybrid model based on the eigenvector to obtain a kidney disease risk prediction model, which is a model that combines the feature selection ability of Lasso regression and the survival analysis ability of Cox proportional hazard model. During the training process, 1000 patients were first randomly divided into a training set and a validation set in a ratio of 7:3. The training set contained 700 patients for model parameter estimation, and the validation set contained 300 patients for model performance evaluation. For each patient in the training set, its 9-dimensional eigenvector Xi, follow-up time and event occurrence status were recorded, where an event occurrence status of 1 indicates the occurrence of diabetic nephropathy, and an event occurrence status of 0 indicates that no diabetic nephropathy occurred at the end of follow-up. Then, for the risk function representation of the Cox regression model, the regression coefficients were solved by the maximum partial likelihood estimation method. After iterative optimization, the regression coefficients of the nine features were finally obtained: KIM-1 corresponding to 0.78, NGAL corresponding to 0.65, Clusterin corresponding to 0.53, IL-6 corresponding to 0.50, FGF-23 corresponding to 0.41, TNF-α corresponding to 0.89, HbA1c corresponding to 1.12, MAP corresponding to 0.076, and BMI corresponding to 0.34. The final risk score calculation formula is:

[0044] Finally, the model obtained by training after including the solved regression coefficients will be used as a kidney disease risk prediction model.

[0045] S5: Outputting the renal disease risk prediction result of the patient to be tested through the renal disease risk prediction model.

[0046] Wherein, step S5 further includes: S51: Collect an EDTA-anticoagulated fasting plasma sample from the patient to be tested to obtain a plasma sample to be tested; S52: Detect the concentration of the core protein marker based on the plasma sample to be tested to obtain concentration data to be tested; collect clinical parameter data of the patient to be tested to obtain clinical parameters to be tested.

[0047] In a specific example, a male patient with type 2 diabetes has an 8-year disease course, an HbA1c of 7.8%, a BMI of 26.5 kg / m², a blood pressure of 138 / 86 mmHg, and no history of proteinuria. The doctor recommends assessing his risk of diabetic nephropathy. During the assessment, fasting EDTA anticoagulated plasma is first drawn to test the concentrations of six protein markers.

[0048] S53: Inputting the concentration data to be measured and the clinical parameters to be measured into the kidney disease risk prediction model to obtain a kidney disease risk prediction result.

[0049] Among them, the kidney disease risk prediction results include the kidney disease risk prediction score and the kidney disease risk prediction stratification results, specifically including: when the kidney disease risk prediction score is RS <15 points, the kidney disease risk prediction stratification result is low risk; when the kidney disease risk prediction score is 15 ≤ RS ≤ 28 points, the kidney disease risk prediction stratification result is medium risk; when the kidney disease risk prediction score is RS>28 points, the kidney disease risk prediction stratification result is high risk.

[0050] Subsequently, parameters were calculated based on the data in steps S51 and S52: mean arterial pressure (MAP) = 86 + 1 / 3×(138-86) ≈ 103 mmHg. The test results were then input: KIM-1: 120 pg / mL, NGAL: 85 pg / mL, Clusterin: 200 pg / mL, IL-6: 3.5 pg / mL, FGF-23: 180 pg / mL, TNF-α: 8 pg / mL. Finally, the risk score was output based on the renal disease risk prediction model: RS = 0.78×120 + 0.65×85 + 0.53×200 + 0.50×3.5 + 0.41×180 + 0.89×8 + 1.12×7.8 + 0.76×(103 / 10) + 0.34×26.5 = 363.05 Based on the above risk stratification, the risk stratification result is RS = 36.3 > 28, indicating high risk, meaning a 10-year incidence rate > 35%. Once the predicted results are obtained, clinical intervention can be initiated: immediately start an SGLT2 inhibitor (such as dapagliflozin) to protect the kidneys, monitor the urine protein-creatinine ratio (UACR) every three months, and strengthen blood pressure control (target < 130 / 80 mmHg) and blood glucose management (HbA1c < 7%).

[0051] like Figure 2 As shown, the present invention also provides a system for predicting the risk of nephropathy in diabetic patients based on protein markers, comprising: Collection module 100: used to collect plasma protein marker concentration data of protein markers in plasma samples of diabetic patients; Analysis module 200: for performing longitudinal correlation analysis on protein markers based on the plasma protein marker concentration data to obtain kidney disease-related protein markers; Screening module 300: for performing feature screening on the kidney disease-related protein marker combination to obtain core protein markers; Training module 400: for establishing a kidney disease risk prediction model based on clinical parameters of diabetic patients and the core protein marker combination; The prediction module 500 is configured as the kidney disease risk prediction model established by the training module 400, and is used to predict the kidney disease risk of the patient to be tested and obtain a kidney disease risk prediction result.

[0052] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0053] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0054] The following describes a method and system for predicting kidney disease risk in diabetic patients based on protein markers according to another embodiment in conjunction with a specific experimental scenario.

[0055] First, the study cohort was established and samples were processed. For subject screening, this example set the inclusion criteria as follows: confirmed type 2 diabetes with a duration of ≥5 years; baseline eGFR ≥60 mL / min / 1.73 m²; and urine albumin / creatinine ratio (UACR) <30 mg / g. Exclusion criteria included: non-diabetic kidney disease (such as glomerulonephritis, autoimmune disease, etc.); use of nephrotoxic drugs (such as NSAIDs) within 3 months; and concomitant malignancy or end-stage liver disease.

[0056] Secondly, clinical data were collected, including baseline parameters including but not limited to age, gender, HbA1c, blood glucose, blood lipids, eGFR, urine albumin / creatinine ratio, etc.

[0057] During follow-up, the follow-up plan was set to follow up once a year, and the KDIGO criteria were used to define diabetic nephropathy.

[0058] Subsequent sample collection involves venous blood collection after fasting for 8 hours using EDTA anticoagulant tubes and centrifugation at 1500×g for 15 minutes at 4°C. After plasma separation, the samples are divided into 0.5 mL cryovials and stored in -80°C ultra-low temperature freezers to avoid repeated freezing and thawing (≤2 times).

[0059] After obtaining the samples, this example uses Olink PEA technology (Proximity Extension Assay) to detect plasma proteins, and then performs plasma marker screening and verification. That is, univariate analysis is used to screen significant proteins (p < 0.1). The obtained Lasso regression screens out six core protein markers: KIM-1, NGAL, Clusterin, IL-6, FGF-23, and TNF-α, as well as three clinical parameters: HbA1c, blood pressure, and BMI. LASSO regression is used to reduce the dimension and select key markers, and a multivariate Cox regression is used to establish a prediction model.

[0060] When validating the model, the model was divided into 70% training set and 30% validation set, and indicators such as C-statistic were calculated.

[0061] Finally, if Figure 3 As shown in FIG, the ROC curve of the diabetic nephropathy risk prediction model of the present invention, wherein the horizontal axis is 1-specificity, i.e., false positive rate, and the vertical axis is sensitivity, i.e., true positive rate. Figure 3It can be seen that the risk prediction model established by the present invention has a sensitivity of 88% in the 10-year risk prediction of diabetic nephropathy, and the early prediction ability is greatly improved. It can effectively identify most potential high-risk patients, thereby providing the clinic with opportunities for early intervention and reducing the risk of disease progression. The specificity of the model is 85%, and the AUC of the model in the verification cohort reaches 0.87. The diagnostic efficacy is improved, which helps to reduce false positive results and ensure that only truly high-risk patients are marked as subjects requiring further examination or intervention, thereby optimizing the allocation of medical resources.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for predicting the risk of nephropathy in diabetic patients based on protein markers, characterized in that: include: S1: Detect the concentration of protein markers in plasma samples of diabetic patients to obtain plasma protein marker concentration data; S2: performing longitudinal correlation analysis on the protein markers based on the plasma protein marker concentration data to obtain kidney disease-related protein markers; S3: performing feature screening on the combination of kidney disease-related protein markers to obtain core protein markers; S4: Establish a kidney disease risk prediction model based on the clinical parameters of diabetic patients and the core protein marker combination; S5: Outputting the renal disease risk prediction result of the patient to be tested through the renal disease risk prediction model.

2. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: Step S1 further comprises: S11: Establish patient cohort; S12: Collect EDTA-anticoagulated fasting plasma samples from diabetic patients in the patient cohort and establish a standardized biospecimen library; S13: Detecting the concentration of the protein marker to be screened in the standardized biological sample library to obtain the plasma protein marker concentration data.

3. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 2, characterized in that: The patient queue in step S11 includes a plurality of type 2 diabetic patients with healthy kidneys, where healthy kidneys include patients without diabetic nephropathy and other kidney diseases.

4. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: Step S2 further comprises: S21: matching the plasma protein marker concentration data with the renal disease status assessment results within a preset follow-up period to obtain a protein combination to be screened; S22: Performing statistical correlation analysis on the protein combination to be screened using a Cox proportional hazard model to obtain kidney disease-related protein markers.

5. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: In step S3, the kidney disease-related protein markers are subjected to feature screening using a Lasso regression algorithm.

6. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: Step S4 further comprises: S41: integrating the core protein markers with clinical parameters to obtain a feature vector; S42: Based on the feature vector, the hybrid model is trained to obtain a kidney disease risk prediction model.

7. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 6, characterized in that: The core protein markers in step S41 include: KIM-1, NGAL, Clusterin, IL-6, FGF-23, TNF-α, and the clinical parameters in step S41 include: HbA1c data, blood pressure, and BMI data.

8. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: Step S5 further comprises: S51: measuring the core protein marker concentration in the plasma sample of the patient to be tested to obtain concentration data to be tested; collecting clinical parameter data of the patient to be tested to obtain clinical parameters to be tested; S52: Inputting the concentration data to be measured and the clinical parameters to be measured into the kidney disease risk prediction model to obtain a kidney disease risk prediction result.

9. The method for predicting the risk of nephropathy in diabetic patients based on protein markers according to claim 1, characterized in that: The kidney disease risk prediction results include the kidney disease risk prediction score and the kidney disease risk prediction stratification results, specifically including: When the renal disease risk prediction score is RS < 15 points, the renal disease risk prediction stratification result is low risk; When the renal disease risk prediction score is 15 ≤ RS ≤ 28 points, the renal disease risk prediction stratification result is medium risk; When the renal disease risk prediction score was RS > 28 points, the renal disease risk prediction stratification result was high risk.

10. A system for predicting the risk of kidney disease in diabetic patients based on protein markers, characterized in that: include: Acquisition module: used to collect plasma protein marker concentration data of protein markers in plasma samples of diabetic patients; Analysis module: used to perform longitudinal correlation analysis on protein markers based on the plasma protein marker concentration data to obtain kidney disease-related protein markers; Screening module: used to perform feature screening on the combination of kidney disease-related protein markers to obtain core protein markers; Training module: used to establish a kidney disease risk prediction model based on the clinical parameters of diabetic patients and the core protein marker combination; The prediction module is configured as the kidney disease risk prediction model established by the training module, and is used to predict the kidney disease risk of the patient to be tested and obtain a kidney disease risk prediction result.

Images