Gout recurrence risk assessment method and system based on ratio of neutrophil to lymphocyte

Abstract

The invention provides a gout recurrence risk assessment method and system based on the ratio of neutrophil to lymphocyte. The method comprises the steps of data collection, data preprocessing and generation of gout recurrence risk assessment sample data through data standardization processing; establishing a combined model of a Cox proportional risk regression model and a competitive risk model, and evaluating the correlation between the ratio NLR of neutrophil to lymphocyte and gout recurrence; taking the NLR as a characteristic variable of a gout recurrence risk prognosis factor, adding the NLR into the baseline model, and constructing a plurality of gout recurrence risk candidate prediction models; screening to obtain a target gout recurrence risk assessment model; generating an evaluation result; and performing visual display. According to the method, the problems of incomplete and skewed distribution of clinical data can be effectively solved, the quality and stability of the data are improved, the accuracy of evaluation and prediction is improved by screening out the effective target gout recurrence risk evaluation model, and reliable auxiliary decision support is provided for clinicians.

Description

Technical Field

[0001] The present application relates to the field of model prediction technology, and in particular to a method and system for assessing gout recurrence risk based on the neutrophil-to-lymphocyte ratio. Background Art

[0002] Gout is a chronic inflammatory disease with a global prevalence of approximately 55.8 million people, of whom approximately 70% have uncontrolled disease. Its relapsing nature and associated complications, such as cardiovascular disease and renal impairment, pose significant challenges to clinical management. Traditional markers such as serum uric acid (UA) levels and tophi are commonly used to monitor gout. However, these indicators have limitations in predicting gout recurrence, particularly in patients with normal serum uric acid levels or those receiving uric acid-lowering therapy, thus providing insufficient guidance for the use of anti-inflammatory drugs. This highlights the need for more reliable, sensitive markers with higher predictive accuracy to better guide treatment strategies to prevent gout recurrence.

[0003] In recent years, research on the pathophysiology of gout has shown that inflammation and immune dysregulation are key drivers of disease recurrence. As an easily accessible marker, the neutrophil-to-lymphocyte ratio (NLR) has shown prognostic value in a variety of inflammatory diseases because it reflects the balance between inflammatory response and immune regulation in the body. To date, no studies have clearly demonstrated the superiority of NLR in predicting gout recurrence, highlighting the need for further research to clarify its potential role in gout management. This is particularly important when traditional markers (such as UA and tophi) and common inflammatory indicators (such as C-reactive protein [CRP]) have been shown to be insufficient to reliably predict gout recurrence.

[0004] Therefore, how to rigorously evaluate the predictive value of NLR for gout recurrence, compare it with the effectiveness of traditional markers (including UA, CRP and tophi), and develop a prediction model to evaluate whether the predictive utility of NLR is still valid in a specific clinical context, such as in patients with normal UA levels, no tophi and receiving uric acid-lowering treatment, is an important topic for technicians in this field. Summary of the Invention

[0005] In view of this, it is necessary to provide a gout recurrence risk assessment method and system based on the neutrophil-to-lymphocyte ratio, including: data acquisition, data preprocessing, and generating gout recurrence risk assessment sample data through data standardization; establishing a combined model of the Cox proportional hazards regression model and the competing risk model to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence; adding NLR as a characteristic variable of the prognostic factor of gout recurrence risk to the baseline model to construct multiple candidate gout recurrence risk prediction models; screening to obtain a target gout recurrence risk assessment model; generating assessment results; and visualizing them. This method can effectively handle the incompleteness and skewed distribution of clinical data, improve data quality and stability, and improve the accuracy of assessment prediction by screening out effective target gout recurrence risk assessment models, providing reliable auxiliary decision support for clinicians.

[0006] This study aims to rigorously evaluate the predictive value of NLR for gout recurrence and compare its effectiveness with that of traditional markers, including UA, CRP, and tophi. In addition, subgroup analyses will be performed and prediction models will be developed to assess whether the predictive utility of NLR remains valid in specific clinical settings, such as in patients with normal UA levels, the absence of tophi, and those receiving urate-lowering therapy.

[0007] In a first aspect, the present invention provides a method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio, the method comprising:

[0008] S1, collects electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system;

[0009] S2, preprocessing the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generating gout recurrence risk assessment sample data through data standardization;

[0010] S3, a combined model of Cox proportional hazards regression and competing risk model was established to evaluate the association between neutrophil-to-lymphocyte ratio (NLR) and gout recurrence;

[0011] S4. When the evaluation association judgment condition is met, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk;

[0012] S5, respectively using C statistics, net reclassification index NRI, and integrated discrimination improvement index IDI to evaluate the multiple candidate gout recurrence risk prediction models for assessing the predictive effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model;

[0013] S6, predicting the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generating a corresponding risk assessment result;

[0014] S7, construct and display a visualization interface to evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence and compare its efficacy with traditional markers.

[0015] Optionally, in another implementation of the first aspect of the present invention, in step S1, the same target population is a set of inpatients with gout ranging from those without gout symptoms to those with gout symptoms, and the electronic medical record data includes: neutrophil-to-lymphocyte ratio NLR value, gout recurrence time, recurrence status, and covariates, and the covariates include: demographic characteristics, lifestyle factors, laboratory tests, physical examinations, comorbidities, and drug use.

[0016] Optionally, in another implementation of the first aspect of the present invention, the S2 preprocesses the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generates gout recurrence risk assessment sample data through data standardization, including:

[0017] Check out outliers through box plots and remove them using the IQR rule;

[0018] View the distribution of missing data values ​​through visualization and perform filling;

[0019] The filled data is standardized, and the processing formula is:

[0020]

[0021] Where X is the original data, The mean value of the original data, σ is the standard deviation of the original data;

[0022] The standardized data were statistically analyzed. The normality of continuous variables was assessed using the Shapiro-Wilk test. Continuous variables were compared between different NLR groups using independent t-tests or nonparametric tests. Categorical variables were analyzed using the chi-square test or Fisher's exact test. Missing data were handled using the multiple imputation by chained equations (MICE) method combined with the classification and regression tree (CART) method. The results of each imputed data set were combined to obtain the final estimate.

[0023] Optionally, in another implementation of the first aspect of the present invention, S3, establishing a Cox proportional hazards regression model and a competing risk model to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence, comprises:

[0024] S3.1, use the Lasso method to select the variables that have the greatest impact on the model, including: The goal of the Lasso method is to minimize the following loss function:

[0025]

[0026] Among them, y i is the target value of the i-th observation, x i is the jth feature of the i-th observation, β0 is the intercept term, β j is the regression characteristic coefficient of the jth feature, λ is the regularization parameter that controls the strength of regularization, n is the number of samples, p is the number of features, and L1 regularization term Some coefficients are compressed to zero, thereby achieving feature selection and obtaining a data set (x1, x2, ..., x p ) T ;

[0027] S3.2, the filtered data are substituted into the Cox proportional hazard regression model for regression to obtain the regression characteristic coefficients β1, β2, ...β p , where the estimator β j satisfy:

[0028]

[0029] in, represents the sample composition of the Cox model, T i represents the survival time, C i represents the censoring time, δ i represents the event identification vector;

[0030] S3.3, using regression characteristic coefficients β1, β2, ... β p For the data set (x1,x2,...,x p ) T By formula Transformation, obtain a new data set (x1',x2',...,x p ') T ;

[0031] S3.4, divide the new dataset into training set and validation set according to the preset ratio;

[0032] S3.5, training the competing risk model using the training set and testing and evaluating it using the validation set to obtain the competing risk model;

[0033] S3.6. The optimal cutoff value of NLR was determined using the Grey test statistic maximization method in the competing risk model and the log-rank test statistic maximization method in the Cox proportional hazards survival model. Sensitivity analysis was performed using the competing risk model and Kaplan-Meier survival analysis to confirm the robustness of the research results under different model settings.

[0034] Optionally, in another implementation of the first aspect of the present invention, in S4, when the evaluation association judgment condition is passed, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk, including:

[0035] The evaluation association judgment conditions are:

[0036] To determine whether the neutrophil-to-lymphocyte ratio (NLR) is associated with an increased risk of gout recurrence during hospitalization and whether it improves predictive accuracy compared with traditional markers such as serum uric acid (UA), tophi, and C-reactive protein (CRP).

[0037] Optionally, in another implementation of the first aspect of the present invention, S5, respectively using the C statistic, the net reclassification index NRI, and the integrated discrimination improvement index IDI to evaluate the multiple candidate gout recurrence risk prediction models for evaluating the prediction effect of gout recurrence to obtain a target gout recurrence risk assessment model, including:

[0038] Construct multiple candidate prediction models for gout recurrence risk, including at least: Cox regression model, support vector machine (SVM) model, k-nearest neighbor (KNN) model, random survival forest model, and XGBoost model;

[0039] The incremental predictive performance of NLR was evaluated by C-statistic, net reclassification index (NRI), and integrated discrimination improvement index (IDI), and compared to screen multiple candidate prediction models for gout recurrence risk.

[0040] The ROC curve was generated by each candidate prediction model for gout recurrence risk and the area under the curve (AUC) of each model was calculated to evaluate the prediction accuracy;

[0041] The differences between AUCs were compared using the DeLong test;

[0042] Obtaining the best performing candidate prediction model as the target gout recurrence risk assessment model;

[0043] The net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indices were calculated to assess the incremental predictive value of incorporating NLR into the baseline model, and confidence intervals were obtained by bootstrap resampling to optimize the target gout recurrence risk assessment model.

[0044] Optionally, in another implementation of the first aspect of the present invention, the step S6, predicting the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generating a corresponding risk assessment result, includes:

[0045] inputting the patient's clinical data into the target gout recurrence risk assessment model;

[0046] Output potential risk assessment results.

[0047] In a second aspect, an embodiment of the present application provides a gout recurrence risk assessment system based on the neutrophil-to-lymphocyte ratio, which is applied to a gout recurrence risk assessment method based on the neutrophil-to-lymphocyte ratio as described in the first aspect, characterized in that the system comprises:

[0048] The data collection module is used to collect electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system;

[0049] A data processing module is used to pre-process the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generate gout recurrence risk assessment sample data through data standardization;

[0050] The NLR evaluation module is used to establish a combined model of the Cox proportional hazards regression model and the competing risk model to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence;

[0051] A model building module is used to add NLR as a characteristic variable of the prognostic factor of gout recurrence risk to the baseline model when the evaluation judgment conditions are met, so as to construct multiple candidate prediction models for gout recurrence risk;

[0052] A model selection module is used to evaluate the multiple candidate gout recurrence risk prediction models using C statistics, net reclassification index NRI, and integrated discrimination improvement index IDI to evaluate the prediction effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model;

[0053] A risk assessment module, configured to predict the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generate corresponding risk assessment results;

[0054] The comparative display module is used to construct a visual interface and display, evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence, and compare its efficacy with traditional markers.

[0055] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0056] processor;

[0057] a memory for storing processor-executable instructions;

[0058] Wherein, the processor is configured to implement a gout recurrence risk assessment method based on the neutrophil-to-lymphocyte ratio as described in the first aspect when executing the instructions.

[0059] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a program, and the program instructs a device to execute a gout recurrence risk assessment method based on the neutrophil-to-lymphocyte ratio as described in any one of claims 1 to 7.

[0060] The embodiment of the present application provides a method and system for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio. The method and system collect electronic medical record data, clinical pathological data, hospitalization and clinical follow-up data of the same target population within a preset time period from a data center system; pre-process the electronic medical record data, clinical pathological data, hospitalization and clinical follow-up data, and generate gout recurrence risk assessment sample data through data standardization; establish a joint model of the Cox proportional hazard regression model and the competing risk model to assess the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence; when the assessment association judgment condition is met, the NLR is used as the risk factor for gout recurrence. Characteristic variables of prognostic factors of risk are added to the baseline model to construct multiple candidate prediction models for gout recurrence risk; the C statistic, net reclassification index (NRI), and integrated discrimination improvement index (IDI) are used to evaluate the multiple candidate prediction models for gout recurrence risk to assess the predictive effect of gout recurrence, thereby obtaining a target gout recurrence risk assessment model; the patient's gout recurrence risk is predicted based on the patient's clinical data and the target gout recurrence risk assessment model, and the corresponding risk assessment results are generated; a visualization interface is constructed and displayed to evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence, and its efficacy is compared with traditional markers. This method can effectively handle the incompleteness and skewed distribution of clinical data, improve the quality and stability of the data, and improve the accuracy of assessment and prediction by screening out effective target gout recurrence risk assessment models, providing reliable auxiliary decision support for clinicians. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of a method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio provided in another embodiment of the present application.

[0062] Figure 2The present invention provides an embodiment of the present invention to provide the distribution and predictive value of the neutrophil-to-lymphocyte ratio (NLR) in determining the optimal cut-off time and the risk of gout recurrence during hospitalization.

[0063] Figure 3 This is another embodiment of the present application providing a graph of the cumulative incidence of gout recurrence and Kaplan-Meier survival curves during hospitalization for the GoutRe cohort and the MIMIC-IV cohort.

[0064] Figure 4 Stratified analysis of the GoutRe and MIMIC-IV cohorts by age, sex, eGFR, UA level, and presence of comorbid diseases is provided for one embodiment of the present application.

[0065] Figure 5 This is an NLR ROC curve for predicting gout recurrence during hospitalization in the GoutRe cohort and the MIMIC-IV cohort, provided in one embodiment of the present application.

[0066] Figure 6 This is a schematic diagram of a gout recurrence risk assessment system module based on the neutrophil-to-lymphocyte ratio provided in one embodiment of the present application.

[0067] Figure 7 A schematic diagram of an electronic terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0069] It should be noted that, in the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art in the art to which this application relates. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0070] It should be noted that, in the embodiments of the present application, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying an order. Features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way.

[0071] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0072] The present application provides a gout recurrence risk assessment method and system based on the neutrophil-to-lymphocyte ratio, which can effectively deal with the incompleteness and skewed distribution problems of clinical data, improve the quality and stability of the data, and improve the accuracy of assessment prediction by screening out an effective target gout recurrence risk assessment model, providing clinicians with reliable auxiliary decision support.

[0073] Figure 1 This is a flow chart of a method for assessing the risk of gout recurrence based on the neutrophil to lymphocyte ratio provided in one embodiment of the present application. Figure 1 As shown, a method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio, the method comprising:

[0074] S1, collects electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system.

[0075] Specifically, in the embodiment of the present application, in step S1, the same target population is a set of inpatient members of gout patients ranging from those without gout symptoms to those with gout symptoms, and the electronic medical record data includes: neutrophil-to-lymphocyte ratio NLR value, gout recurrence time, recurrence status, and covariates, and the covariates include: demographic characteristics, lifestyle factors, laboratory tests, physical examinations, comorbidities and drug use.

[0076] S2. Preprocess the electronic medical record data, clinical pathological data, hospitalization and clinical follow-up data, and generate gout recurrence risk assessment sample data through data standardization.

[0077] Specifically, in the embodiment of the present application, the S2 pre-processes the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generates gout recurrence risk assessment sample data through data standardization, including:

[0078] Check out outliers through box plots and remove them using the IQR rule;

[0079] View the distribution of missing data values ​​through visualization and perform filling;

[0080] The filled data is standardized, and the processing formula is:

[0081]

[0082] Where X is the original data, The mean value of the original data, σ is the standard deviation of the original data;

[0083] The standardized data were statistically analyzed. The normality of continuous variables was assessed using the Shapiro-Wilk test. Continuous variables were compared between different NLR groups using independent t-tests or nonparametric tests. Categorical variables were analyzed using the chi-square test or Fisher's exact test. Missing data were handled using the multiple imputation by chained equations (MICE) method combined with the classification and regression tree (CART) method. The results of each imputed data set were combined to obtain the final estimate.

[0084] S3. A combined Cox proportional hazards regression model and a competing risk model were established to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence.

[0085] Specifically, in the examples of the present application, S3, establishing a Cox proportional hazards regression model and a competing risk model to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence, includes:

[0086] S3.1, use the Lasso method to select the variables that have the greatest impact on the model, including: The goal of the Lasso method is to minimize the following loss function:

[0087]

[0088] Among them, y i is the target value of the i-th observation, x i is the jth feature of the i-th observation, β0 is the intercept term, β j is the regression characteristic coefficient of the jth feature, λ is the regularization parameter that controls the strength of regularization, n is the number of samples, p is the number of features, and L1 regularization term Some coefficients are compressed to zero, thereby achieving feature selection and obtaining a data set (x1, x2, ..., x p ) T ;

[0089] S3.2, the filtered data are substituted into the Cox proportional hazard regression model for regression to obtain the regression characteristic coefficients β1, β2, ...β p , where the estimator β j satisfy:

[0090]

[0091] in, represents the sample composition of the Cox model, T i represents the survival time, C i represents the censoring time, δ i represents the event identification vector;

[0092] S3.3, using regression characteristic coefficients β1, β2, ... β p For the data set (x1,x2,...,x p ) T By formula Transformation, obtain a new data set (x1',x2',...,x p ') T ;

[0093] S3.4, divide the new dataset into training set and validation set according to the preset ratio;

[0094] S3.5, training the competing risk model using the training set and testing and evaluating it using the validation set to obtain the competing risk model;

[0095] S3.6. The optimal cutoff value of NLR was determined using the Grey test statistic maximization method in the competing risk model and the log-rank test statistic maximization method in the Cox proportional hazards survival model. Sensitivity analysis was performed using the competing risk model and Kaplan-Meier survival analysis to confirm the robustness of the research results under different model settings.

[0096] Specifically, the model training process mainly consists of four steps: 1. Pre-training stage; 2. Supervised fine-tuning, also called instruction fine-tuning stage; 3. Reward model training stage; 4. Reinforcement learning fine-tuning stage. The training steps specifically include: 1) Data acquisition and preprocessing: Collect and organize the data sets needed for training. 2) Model design and construction: Design and build a suitable model according to task requirements. 3) Model training: Use the training data set to train the model, and continuously adjust the model parameters so that the model can better fit the data. 4) Model evaluation and optimization: Use the test data set to evaluate the trained model, and optimize the model based on the evaluation results. 5) Model deployment and use:

[0097] Deploy the trained model to actual application scenarios and apply it.

[0098] S4. When the evaluation association judgment conditions are met, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk.

[0099] Specifically, in the embodiment of the present application, in S4, when the evaluation association judgment condition is passed, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk, including:

[0100] The evaluation association judgment conditions are:

[0101] To determine whether the neutrophil-to-lymphocyte ratio (NLR) is associated with an increased risk of gout recurrence during hospitalization and whether it improves predictive accuracy compared with traditional markers such as serum uric acid (UA), tophi, and C-reactive protein (CRP).

[0102] S5. C statistics, net reclassification index NRI and integrated discrimination improvement index IDI are used to evaluate the multiple candidate prediction models for gout recurrence risk to assess the predictive effect of gout recurrence, so as to obtain the target gout recurrence risk assessment model.

[0103] Specifically, in the embodiment of the present application, the S5 uses the C statistic, the net reclassification index NRI, and the integrated discrimination improvement index IDI to evaluate the multiple gout recurrence risk candidate prediction models for evaluating the prediction effect of gout recurrence to obtain the target gout recurrence risk assessment model, including:

[0104] Construct multiple candidate prediction models for gout recurrence risk, including at least: Cox regression model, support vector machine (SVM) model, k-nearest neighbor (KNN) model, random survival forest model, and XGBoost model;

[0105] The incremental predictive performance of NLR was evaluated by C-statistic, net reclassification index (NRI), and integrated discrimination improvement index (IDI), and compared to screen multiple candidate prediction models for gout recurrence risk.

[0106] The ROC curve was generated by each candidate prediction model for gout recurrence risk and the area under the curve (AUC) of each model was calculated to evaluate the prediction accuracy;

[0107] The differences between AUCs were compared using the DeLong test;

[0108] Obtaining the best performing candidate prediction model as the target gout recurrence risk assessment model;

[0109] The net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indices were calculated to assess the incremental predictive value of incorporating NLR into the baseline model, and confidence intervals were obtained by bootstrap resampling to optimize the target gout recurrence risk assessment model.

[0110] S6. Predict the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generate corresponding risk assessment results.

[0111] Specifically, in the embodiment of the present application, the step S6 predicts the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generates a corresponding risk assessment result, including:

[0112] inputting the patient's clinical data into the target gout recurrence risk assessment model;

[0113] Output potential risk assessment results.

[0114] S7, construct and display a visualization interface to evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence and compare its efficacy with traditional markers.

[0115] Experiment 1

[0116] This study aims to rigorously evaluate the predictive value of NLR for gout recurrence and compare its effectiveness with that of traditional markers, including UA, CRP, and tophi. In addition, subgroup analyses will be performed and prediction models will be developed to assess whether the predictive utility of NLR remains valid in specific clinical settings, such as in patients with normal UA levels, the absence of tophi, and those receiving urate-lowering therapy.

[0117] This retrospective study was based on two cohorts: the GoutRe cohort, which included patients from five hospitals, and the MIMIC-IV cohort, which contains deidentified health-related data from a series of hospitalized patients. The GoutRe cohort included patients hospitalized between January 1, 2010, and July 30, 2024, while the MIMIC-IV cohort used data from 2008 to 2019. The inclusion of a diverse population enhances the generalizability and robustness of the study findings.

[0118] This international, multicenter, retrospective cohort study included gout patients from five hospitals in China (the GoutRe cohort) and the MIMIC-IV database (the MIMIC-IV cohort) for external validation. Participants in both cohorts were diagnosed with gout according to the 1977 ACR and / or 2015 ACR / EULAR gout classification criteria. The GoutRe cohort included patients hospitalized between January 1, 2010, and July 30, 2024, while the MIMIC-IV cohort used data collected between 2008 and 2019.

[0119] A total of 97,086 patients in both cohorts (GoutRe cohort = 78,140; MIMIC-IV cohort = 18,946) met the 1977 ACR and / or 2015 ACR / EULAR gout classification criteria and had a gout-related ICD-10 code on their discharge diagnosis. Exclusion criteria included patients admitted for acute gout flare, chronic gouty arthritis, other autoimmune diseases, other musculoskeletal conditions requiring nonsteroidal anti-inflammatory drugs or glucocorticoids, patients reporting joint pain without a clear diagnosis, patients with difficulty identifying a gout flare, and a hospital length of stay (LOS) of less than 3 days (eFigure 1 in the Supplementary Material). After applying these exclusion criteria, 6,526 patients were initially enrolled in the GoutRe cohort and 6,475 patients were initially enrolled in the MIMIC-IV cohort.

[0120] We further excluded patients with missing NLR values ​​(301 patients in the GoutRe cohort and 3,286 patients in the MIMIC-IV cohort) or with more than 30% missing data (641 patients in the GoutRe cohort). Furthermore, we excluded patients who died during hospitalization (597 patients in the MIMIC-IV cohort) or were transferred to the ICU (583 patients in the MIMIC-IV cohort). The final study population included 5,584 patients in the GoutRe cohort and 2,019 patients in the MIMIC-IV cohort (eFigure 1 in the Supplementary Material).

[0121] Covariates included demographic characteristics, lifestyle factors, laboratory tests, physical examinations, comorbidities, and medication use. Demographic characteristics included age, sex, race, and weight change. Lifestyle factors included smoking and alcohol history. Laboratory tests included complete blood count parameters and renal function tests. Physical examination findings included the presence of tophi. Comorbidities assessed included hypertension, diabetes, cardiovascular disease, heart failure, stroke, dyslipidemia, fatty liver, kidney disease, thyroid disease, cancer, history of kidney stones, and metabolic syndrome. Medication use included antigout medications, hypoglycemic medications, cardiovascular medications, anticoagulants, lipid-lowering medications, and mannitol.

[0122] Measurement of NLR: In the GoutRe cohort, baseline NLR values ​​were determined on the first day of hospital admission, before the start of any treatment. Neutrophil and lymphocyte counts were obtained by complete blood count analysis of blood samples and expressed as ×10 3 The NLR is reported in cells / μL. The NLR is calculated by dividing the absolute neutrophil count by the absolute lymphocyte count. In the MIMIC-IV database, laboratory data for patients with recurrent gout were selected and analyzed from the earliest available record after admission, confirming that the data were collected before the gout flare. The calculation method is the same as described above.

[0123] Statistical analysis: First, the normality of continuous variables was assessed using the Shapiro-Wilk test. When necessary, independent t-tests or nonparametric tests were used to compare continuous variables between different NLR groups. Categorical variables were analyzed using the chi-square test or Fisher's exact test. Missing data were handled using the multiple imputation by chained equations (MICE) method combined with the classification and regression tree (CART) method, and the results of each imputed dataset were combined to obtain the final estimate. The optimal cutoff value of NLR was determined using the Grey test statistic maximization method in the competing risk model and the log-rank test statistic maximization method in the Cox proportional hazards survival model. Sensitivity analyses were performed using the competing risk model and Kaplan-Meier survival analysis to confirm the robustness of the study results under different model settings.

[0124] To assess the association between NLR and gout recurrence, Cox proportional hazards regression models were used to calculate hazard ratios (HRs), 95% confidence intervals (CIs), and P values. Model 1 was unadjusted; Model 2 was adjusted for NLR, serum uric acid (UA), estimated glomerular filtration rate (eGFR), and tophi; and Model 3 was further adjusted for age, sex, smoking history, alcohol consumption, and comorbidities (e.g., hypertension, diabetes, and cardiovascular disease). Subgroup analyses explored the consistency of the NLR effect across age, sex, and comorbidities. Interaction P values ​​were calculated using likelihood ratio tests comparing models with and without the interaction terms for each stratification factor.

[0125] To further evaluate predictive performance, machine learning models were employed in addition to the traditional Cox regression model. These included random survival forest, support vector machine (SVM), k-nearest neighbor (KNN), and XGBoost models. Predictive accuracy was assessed by generating receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC) for each model. The DeLong test was used to compare the statistical significance of differences between AUCs.

[0126] In addition, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indices were calculated to assess the incremental predictive value of incorporating the NLR into the baseline model, and confidence intervals were obtained by bootstrap resampling. Decision curve analysis (DCA) was performed to evaluate the clinical utility of the NLR-based model at a range of threshold probabilities, assessing the net benefit of various prediction models in clinical settings. All statistical tests were two-sided, and P values ​​less than 0.05 were considered statistically significant. All analyses were performed using R version 4.3.2 (R Project for Statistical Computing).

[0127] Main Outcomes and Measures: The primary outcome was gout recurrence rate, assessed using hospitalization and clinical follow-up data. Cox proportional hazards regression and competing risk models were used to assess the association between NLR and gout recurrence. The incremental predictive performance of NLR was assessed using the C statistic, net reclassification index (NRI), and integrated discrimination improvement index (IDI), with additional comparisons performed across multiple modeling approaches, including XGBoost and Cox regression.

[0128] Results: A total of 7,603 participants were enrolled (GoutRe cohort: 5,584; MIMIC-IV cohort: 2,019). In the GoutRe cohort, the recurrence rate was higher in the high-NLR group than in the low-NLR group (39.7% vs. 20.1%; P < .001), and similar results were observed in the MIMIC-IV cohort (18.2% vs. 10.1%; P < .001). Patients in the high-NLR group were older (mean age, 65.0 years vs. 60.5 years; P < .01) and had more comorbidities, including tophi (7.97% vs. 3.83%; P < .01), nephrolithiasis (44.0% vs. 40.5%; P < .01), and heart failure (8.52% vs. 3.65%; P < .01). Use of diuretics (24.49% vs. 9.13%; P < .01) and beta-blockers (27.96% vs. 18.57%; P < .01) was also more common.

[0129] In both the GoutRe cohort (HR, 2.05; 95% CI, 1.83-2.30; P < .001) and the MIMIC-IV cohort (HR, 2.84; 95% CI, 2.08-3.87; P < .001), elevated NLR was associated with an increased risk of gout recurrence. NLR was weakly positively correlated with CRP (R 2 =0.091; P < .001), but had no significant correlation with UA (R 2 =8.9e-05; P = .61). The AUC of NLR (0.629) was higher than that of UA (0.591) or CRP (0.613).

[0130] Subgroup analyses confirmed that elevated NLR was associated with a higher risk of gout recurrence in patients with normal uric acid levels (HR, 1.87; 95% CI, 1.45-2.41; P < .001) and those receiving uric acid-lowering therapy (HR, 2.15; 95% CI, 1.68-2.76; P < .001). Adding NLR to the baseline model improved predictive accuracy (increased C-statistics: GoutRe from 0.65 to 0.68, P < .001; MIMIC-IV from 0.80 to 0.81, P = .003). The added value of NLR was further supported by the NRI (0.12; P < .01) and IDI (0.05; P < .01).

[0131] Conclusions and Relevance: A high NLR was significantly associated with an increased risk of gout recurrence during hospitalization. These findings suggest that NLR may serve as a simple, reliable, and easily accessible marker for identifying individuals at high risk of recurrence, supporting its potential use in guiding clinical management and tailoring therapeutic interventions.

[0132] A total of 7603 participants were enrolled in this study, including 5584 patients from the GoutRe cohort (100% Asian; mean [SD] age, 62.7 [14.7] years; 86.3% male) and 2019 patients from the MIMIC-IV cohort (primarily white [72.9%]; mean [ST] age, 68.0 [12.9] years; 75.1% male). The GoutRe cohort consisted of 1659 patients with relapsing gout and 3925 patients with non-relapsing gout, while the MIMIC-IV cohort included 320 patients in the relapsing group and 1699 patients in the non-relapsing group.

[0133] like Figure 2 As shown, the participants were divided into two groups ( Figure 2In the GoutRe cohort, patients in the high NLR group (n = 2736) were older (mean [SD] age, 65.0 [14.4] years vs 60.5 [14.6] years, P < 0.01) and had a higher incidence of gout recurrence (39.69% vs 20.12%, P < 0.01), tophi (7.97% vs 3.83%, P < 0.01) and stones (44.0% vs 40.5%, P < 0.01), and heart failure (8.52% vs 3.65%, P < 0.01) compared with the low NLR group (n = 2848). They also showed higher serum uric acid levels (e.g., >10 mg / dL: 19.70% vs 16.22%, P < 0.01), lower eGFR levels (e.g., ≥90 mL / min / 1.73 m 2 : 13.01% vs. 26.54%, P < 0.01), and more use of diuretics (24.49% vs. 9.13%, P < 0.01), β-blockers (27.96% vs. 18.57%, P < -01), and NaHCO3 (30.30% vs. 20.01%, P < 0.01).

[0134] Similarly, in the MIMIC-IV cohort, compared with the low NLR group (n=582), the high NLR group had a higher rate of gout recurrence (18.16% vs. 10.14%, P<0.01), a higher prevalence of heart failure (41.61% vs. 29.73%, P<0.01), and more frequent use of diuretics (34.59% vs. 15.29%, P<0.01) and beta-blockers (10.58% vs. 5.84%, P<0.01). Baseline characteristics, including race, laboratory values ​​(such as UA, eGFR), and comorbidities, highlighted statistically significant differences between the high and low NLR groups (P<0.05).

[0135] The distribution of NLR in the GoutRe cohort is as follows Figure 2 As shown in Part B of Figure 3, most of the values ​​are concentrated between 2 and 3. There is a nonlinear relationship between NLR and gout recurrence ( Figure 2 The optimal NLR cutoff value of 2.69 was determined by combining the log-rank test maximization within the survival analysis framework and the Gray test within the competing risk model framework ( Figure 2 The predictive importance of NLR was comparable to that of known predictors including UA and tophi ( Figure 2 This finding was confirmed in the MIMIC-IV cohort, further supporting the robustness of NLR as an important predictor of gout recurrence ( Figure 2 E in the .

[0136] The cumulative incidence of gout recurrence during hospitalization was Figure 3 In both the GoutRe and MIMIC-IV cohorts, patients with elevated NLR showed a significantly higher risk of recurrence (log-rank P < 0.001). This association was consistent across all subgroups examined, including those with normal serum uric acid levels, those without tophi, those with normal serum UA levels but without tophi, and those receiving uric acid-lowering therapy and anti-inflammatory therapy ( Figure 4-5 ). Kaplan-Meier survival curve ( Figure 3 , Figure 4 ) further supported these observations, demonstrating a robust association between elevated NLR and increased relapse rates across various clinical settings. In a multivariable Cox model, elevated NLR was independently associated with an increased risk of relapse (GoutRe: HR, 2.05; 95% CI, 1.83-2.30; P < 0.01; MIMIC-IV: HR, 2.84; 95% CI 2.08-3.87; P < 0.01).

[0137] Linear regression analysis showed that there was a positive correlation between NLR and CRP (R 2 =0.091, P < 0.001)( Figure 3 However, no significant correlation was found between NLR and UA levels (R 2 =8.9e-05, P=0.61)( Figure 3 C in the C section), and no significant correlation was found between CRP and UA (R 2 =0.0017, P=0.026)( Figure 3 ROC curves were generated to compare the predictive abilities of NLR, UA, and CRP for gout recurrence ( Figure 3 (D in the Figure 3A). The AUC for NLR was 0.629, indicating better predictive performance compared with UA (AUC = 0.591) and CRP (AUC = 0.613). The DeLong test confirmed that the predictive performance of NLR was significantly better than that of UA (P = 0.025), while there was no significant difference between CRP and NLR (P = 0.623).

[0138] Figure 4Stratified analyses of the GoutRe and MIMIC-IV cohorts by age, sex, eGFR, UA level, and presence of comorbid diseases are shown. A significant association was observed between elevated NLR and an increased risk of gout recurrence in all subgroups. For example, in the GoutRe cohort, patients with UA levels between 7 and 7.9 mg / dL had an HR of 2.38 (95% CI, 1.80-3.15; P < 0.001), while patients with UA levels between 9 and 9.9 mg / dL had an HR of 2.83 (95% CI 2.08-3.84; P < 0.001).

[0139] In both the GoutRe and MIMIC-IV cohorts, elevated NLR remained a strong predictor of gout recurrence across a variety of clinical settings and comorbidities, even after adjusting for all covariates. These trends were consistent across subgroups including cancer, cardiovascular disease, diabetes, dyslipidemia, hypertension, heart failure, kidney disease, stroke, thyroid disease, and kidney stones, as well as across age, sex, eGFR, and UA levels. Notably, in the MIMIC-IV cohort, the HR for gout recurrence in patients with thyroid disease was 5.16 (95% CI, 2.38-11.19; P < 0.001), further emphasizing the strong predictive value of NLR in different patient populations ( Figure 4 ).

[0140] Adding NLR to the baseline prediction model significantly improved the accuracy of the model in predicting gout recurrence during hospitalization, with the C statistic increasing from 0.65 to 0.68 in the GoutRe cohort (P < 0.001) and from 0.80 to 0.81 in the MIMIC-IV cohort (P = 0.003). These improvements were further supported by the significant improvement in the IDI index. ROC curve ( Figure 5 ) visually demonstrated the improvements in sensitivity and specificity after incorporating NLR. Furthermore, decision curve analysis confirmed the clinical utility of the enhanced model, showing consistent net benefits across the threshold probability range in both cohorts.

[0141] Four different modeling methods were used to evaluate the prediction of gout recurrence: Cox regression, support vector machine (SVM), random survival forest, and XGBoost. The inclusion of NLR significantly improved the AUC of all models in the GoutRe and MIMIC-IV datasets. Specifically, the XGBoost model showed the highest AUC, indicating its superior performance (GoutRe: AUC = 0.73; MIMIC-IV: AUC = 0.85), followed by the random survival forest model (GoutRe: AUC = 0.72; MIMIC-IC-IV: ACC = 0.85. This highlights the robustness of the XGBoost and random survival forest models in predicting gout recurrence when combined with NLR ( Figure 5 ).

[0142] Given the lack of reliable predictors of gout recurrence, our study aimed to bridge this gap by thoroughly investigating the potential association between NLR and gout recurrence. To the best of our knowledge, this is the first study conducted to date in hospitalized patients with gout to evaluate the association between NLR and gout recurrence. Using data from the GoutRe cohort, including patients from five hospitals in China and the MIMIC-IV database, we found that higher NLR levels were independently associated with an increased risk of gout recurrence in the general population. Otherwise, we not only clearly validated the importance of NLR as a new predictor of gout recurrence, but also demonstrated its wide applicability in different patient populations, thereby significantly improving the predictive ability and accuracy of existing prediction models.

[0143] Based on the calculated critical value of 2.69, the patient population was accurately divided into two groups: a high NLR group and a low NLR group. Comparative analysis revealed a significantly increased risk of gout recurrence in the high NLR group compared with the low NLR group. These findings are consistent with previous studies showing that patients with acute gout have an increased percentage of neutrophils and a decreased percentage of lymphocytes in peripheral blood. Furthermore, patients with acute gout had a higher NLR compared with those in remission.

[0144] The predictive ability of NLR for gout recurrence stems from its reflection of the systemic inflammatory state. NLR is an intuitive and effective indicator for assessing the inflammatory response in the body. Many studies have shown that NLR is of great value in disease assessment, prediction, disease activity and treatment effect evaluation in patients with systemic lupus erythematosus and rheumatoid arthritis combined with lupus nephritis. Gout is an inflammatory disease, and its recurrence is closely related to the level of systemic inflammation. During a gout attack, neutrophils migrate to the affected joints after urate crystals are deposited and work to eliminate these crystals through phagocytosis, thus playing a key role.

[0145] This process triggers local inflammation, leading to joint swelling and pain. Lymphocytes play a crucial role in immune system regulation, and their reduction may indicate impaired immune regulation and ineffective suppression of inflammation, thereby increasing the risk of gout recurrence. Given the sensitivity of a single cell type as an inflammatory marker to fluctuations caused by various internal and external factors, it may not be sufficient to accurately describe the complex characteristics and severity of the immune-inflammatory state.

[0146] In contrast, the NLR, obtained by integrating neutrophil and lymphocyte counts, can simultaneously reflect changes in the status of these two key inflammatory cells and reveal the body's inflammatory and immune status at a broader level. Therefore, the combined ratio of multiple cell counts as an inflammatory marker may have higher reliability and accuracy than individual cell indicators. Therefore, an elevated NLR value indicates elevated levels of inflammation in the patient, which increases the likelihood of future relapse.

[0147] Compared with previous gout recurrence predictors, the gout recurrence model constructed in this study has significantly improved performance and a higher AUC. The research has made significant progress and in-depth exploration in multiple dimensions.

[0148] A comprehensive validation study was conducted to confirm the broad applicability of NLR. Although uric acid levels are the primary determinant of gout flares, some patients with normal uric acid levels may still experience recurrent flares, suggesting that relying solely on uric acid levels as a predictor may lack sufficient accuracy. Tophus is a manifestation of chronic gout caused by recurrent and prolonged flares, indicating progression of the disease to the chronic stage.

[0149] However, it should be noted that not all patients with gout will develop tophi. While uric acid-lowering medications can effectively lower uric acid levels, timely treatment and patient compliance may affect their effectiveness. Furthermore, gout flares can occur even with drug treatment. Given these factors, accurately predicting gout recurrence in patients with normal uric acid levels, no tophi formation, and receiving uric acid-lowering therapy is more challenging.

[0150] Subgroup analysis showed that NLR had consistent and robust predictive ability across different patient populations, particularly in individuals with normal serum uric acid levels, no tophi formation, and receiving uric acid-lowering therapy. This finding suggests that NLR may partially compensate for the limitations of traditional predictors in predicting gout recurrence and provide a new perspective and tool for the prediction of gout. It is proposed that due to the inflammatory nature of gout, mild local or systemic inflammation may persist even in patients with normal serum uric acid levels and no tophi formation. NLR can sensitively capture this potential inflammatory activity, leading to elevated NLR.

[0151] The predictive efficacy of NLR in patients receiving urate-lowering therapy may be due in part to its ability to identify non-responders or those with other persistent inflammatory factors. As a comprehensive inflammatory marker, NLR may more comprehensively reflect the impact of heterogeneity in the gout population on disease progression and relapse risk. Therefore, NLR shows relatively stable predictive efficacy across different patient populations. As an independent composite prognostic factor, the predictive efficacy of NLR is less affected by common clinical variables, indicating that it provides complementary information beyond traditional indicators and helps to comprehensively assess the risk of relapse in patients.

[0152] Furthermore, the incremental impact of incorporating NLR into each of the constructed machine learning models was assessed. The analysis revealed a striking finding that incorporating NLR significantly improved the accuracy of predicting gout recurrence, thus highlighting the independent and critical predictive value of NLR.

[0153] Finally, a comparative analysis of the sensitivity and specificity of NLR, CRP, and UA in predicting gout recurrence found that the AUC of NLR exceeded that of CRP and UA. A study investigating the efficacy of multiple inflammatory markers in the diagnosis of acute gout showed that CRP was superior to other inflammatory markers such as SIRI, ESR, NLR, MLR, SII, and PLR.

[0154] However, this study suggests that NLR is superior to CRP in predicting gout recurrence. This may be because NLR, as an indicator of the overall state of the body's immune system, is more sensitive to monitoring the dynamic fluctuations of cytokine release and inflammatory responses. Furthermore, NLR may be more closely related to the pathophysiology of gout, especially considering the key role of neutrophils in gout attacks. CRP is an acute-phase inflammatory biomarker synthesized by the liver in response to immune cytokine stimulation.

[0155] Although it exhibits rapid elevation in inflammatory infections and other conditions, its ability to accurately reflect the extent of inflammation often exhibits a certain time delay, making it susceptible to various non-infectious factors. Consistent with the investigation by, previous studies have shown that GlycA can serve as a reliable long-term biomarker for assessing the state of neutrophil hyperactivity and has a stronger predictive ability for gout recurrence than UA levels. Although several studies have shown a close association between UA levels and the risk of gout recurrence, it remains challenging to capture the body's comprehensive inflammatory response through direct indicators of uric acid metabolism alone. Although hyperuricemia is a prerequisite for gout attacks, not all patients with hyperuricemia will develop gout, and some gout patients may show normal serum uric acid levels during a flare.

[0156] Therefore, the predictive value of UA levels in determining gout recurrence has certain limitations. As a new inflammatory marker, NLR has higher sensitivity and specificity than CRP and UA levels in predicting gout recurrence risk. By integrating information from neutrophils and lymphocytes, it can more comprehensively reflect the body's inflammatory response, providing a more reliable basis for assessing recurrence risk in gout patients.

[0157] The contribution of this study is the validation of NLR as a cost-effective, readily available, and simple biomarker for predicting gout recurrence. This finding provides a more precise basis for clinical decision-making and helps healthcare providers develop personalized prevention and treatment strategies, including adjusting medication regimens and strengthening lifestyle interventions. The ultimate goal is to effectively reduce the risk of recurrence in gout patients.

[0158] This study has some limitations. First, the analysis focused on the comparison of NLR, CRP, and UA, without a comprehensive assessment of other inflammatory markers. Future studies should include a wider range of inflammatory markers to provide a more in-depth comparative analysis. Second, due to the dynamic nature of NLR as a biomarker affected by multiple factors, a single measurement may not fully reflect an individual's true risk level. Therefore, future studies should expand the sample size to improve the generalizability of the findings and incorporate additional inflammatory markers and other relevant indicators to further improve the predictive performance of the model. In addition, the study was limited to hospitalized patients, which may affect the broad applicability of the results. Future studies should include more diverse patient groups to verify the generalizability of the conclusions.

[0159] The results of this study strongly support the use of NLR as a practical and sensitive biomarker for predicting gout recurrence. Incorporating NLR into clinical practice could help clinicians accurately identify patients at high risk of gout recurrence, improve risk stratification and management of gout patients, help develop more personalized prevention strategies and optimize treatment interventions, ultimately reducing the risk of recurrence in patients.

[0160] In this international multicenter retrospective cohort study of 7603 hospitalized gout patients, a higher NLR (cutoff value >2.69) was significantly associated with an increased risk of gout recurrence across various clinical settings and comorbidities. Incorporating NLR into the prediction model significantly improved its performance. NLR had better predictive ability than UA and CRP. In this study, NLR is a simple, reliable, and accessible biomarker that can identify patients at high risk of gout recurrence, potentially improving clinical management and guiding tailored therapeutic interventions.

[0161] Figure 6A schematic diagram of a gout recurrence risk assessment system module based on the neutrophil-to-lymphocyte ratio provided in one embodiment of the present application. In this embodiment of the present application, the system includes:

[0162] The data collection module 11 is used to collect electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system.

[0163] The data processing module 12 is used to pre-process the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generate gout recurrence risk assessment sample data through data standardization.

[0164] NLR evaluation module 13 is used to establish a combined model of Cox proportional hazards regression model and competing risk model to evaluate the association between neutrophil-to-lymphocyte ratio (NLR) and gout recurrence.

[0165] The model building module 14 is used to add NLR as a characteristic variable of the prognostic factor of gout recurrence risk to the baseline model when the evaluation judgment condition is passed, so as to construct multiple candidate prediction models for gout recurrence risk.

[0166] The model selection module 15 is used to evaluate the multiple candidate gout recurrence risk prediction models using C statistics, net reclassification index NRI and integrated discrimination improvement index IDI to evaluate the prediction effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model.

[0167] The risk assessment module 16 is used to predict the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generate corresponding risk assessment results.

[0168] The comparison display module 17 is used to construct a visualization interface and display, evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio NLR for gout recurrence, and compare its efficacy with traditional markers.

[0169] It can be understood that the gout recurrence risk assessment system based on the neutrophil-to-lymphocyte ratio provided in the embodiment of the present application can effectively deal with the incompleteness and skewed distribution problems of clinical data, improve the quality and stability of the data, and improve the accuracy of assessment predictions by screening out effective target gout recurrence risk assessment models, thereby providing reliable auxiliary decision support for clinicians.

[0170] See also Figure 7 , Figure 7 This is an electronic terminal device provided by an embodiment of the present application. Figure 7The electronic terminal device shown includes at least the following parts: one or more processors, one or more input devices, one or more output devices, and one or more memories. The processors, input devices, output devices, and memories communicate with each other via a communication bus. The memory is used to store computer programs, which include program instructions. The processor is used to execute the program instructions stored in the memory. The processor is configured to call the program instructions to perform the following operations to perform the functions of the modules / units in the above-mentioned device embodiments, such as Figure 6 Functionality of the modules shown.

[0171] In an embodiment of the present application, a computer-readable storage medium includes instructions, and the instructions instruct a device to execute the system of the first aspect. For example, the instructions instruct the device to execute Figure 1 The steps shown in the figure are a method and system for assessing the risk of gout recurrence based on the ratio of neutrophils to lymphocytes.

[0172] It should be understood that in the embodiment of the present invention, the so-called processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. It should be noted that a part of the electronic device of the above embodiment may also be implemented by a computer. In this case, the program for implementing the control function may be recorded in a computer-readable recording medium, and the program recorded in the recording medium may be read into a computer and executed.

[0173] Input devices may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc. Output devices may include a display (LCD, etc.), a speaker, etc.

[0174] It should be noted that the "computer" mentioned here refers to a computer built into an electronic device, employing hardware including an operating system (OS) and peripheral devices. Furthermore, "computer-readable recording medium" refers to removable media such as floppy disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computers.

[0175] Furthermore, "computer-readable recording media" may include: media that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines; and media that store programs for a fixed period of time, such as volatile memory within computers acting as servers or clients in this context. Furthermore, the aforementioned program may be a program for implementing a portion of the aforementioned functions, or a program that can achieve the aforementioned functions by combining with a program already stored in a computer.

[0176] Furthermore, the electronic device in the above-described embodiments can also be implemented as a collection (device group) consisting of multiple devices. Each device constituting the device group may have a portion or all of the functions or functional blocks of the electronic device in the above-described embodiments. A device group only needs to have all the functions or functional blocks of the electronic device.

[0177] Those skilled in the art should recognize that the above embodiments are merely intended to illustrate the present application and are not intended to limit the present application. As long as they are within the spirit of the present application, appropriate changes and modifications to the above embodiments are within the scope of protection claimed in the present application.

Claims

1. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio, characterized in that: The method comprises: S1, collects electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system; S2, preprocessing the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generating gout recurrence risk assessment sample data through data standardization; S3, a combined model of Cox proportional hazards regression and competing risk model was established to evaluate the association between neutrophil-to-lymphocyte ratio (NLR) and gout recurrence; S4. When the evaluation association judgment condition is met, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk; S5, respectively using C statistics, net reclassification index NRI, and integrated discrimination improvement index IDI to evaluate the multiple candidate gout recurrence risk prediction models for assessing the predictive effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model; S6, predicting the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generating a corresponding risk assessment result; S7, construct and display a visualization interface to evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence and compare its efficacy with traditional markers.

2. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 1, characterized in that: In step S1, the same target population is a set of inpatient members of gout patients ranging from those without gout symptoms to those with gout symptoms, and the electronic medical record data includes: neutrophil-to-lymphocyte ratio (NLR) value, gout recurrence time, recurrence status, and covariates, and the covariates include: demographic characteristics, lifestyle factors, laboratory tests, physical examinations, comorbidities, and drug use.

3. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 1, characterized in that: S2, preprocessing the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generating gout recurrence risk assessment sample data through data standardization, including: Check out outliers through box plots and remove them using the IQR rule; View the distribution of missing data values ​​through visualization and perform filling; The filled data is standardized, and the processing formula is: Where X is the original data, The mean value of the original data, σ is the standard deviation of the original data; The standardized data were statistically analyzed. The normality of continuous variables was assessed using the Shapiro-Wilk test. Continuous variables were compared between different NLR groups using independent t-tests or nonparametric tests. Categorical variables were analyzed using the chi-square test or Fisher's exact test. Missing data were handled using the multiple imputation by chained equations (MICE) method combined with the classification and regression tree (CART) method. The results of each imputed data set were combined to obtain the final estimate.

4. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 3, characterized in that: S3, Cox proportional hazards regression model and competing risk model were established to evaluate the association between neutrophil-to-lymphocyte ratio (NLR) and gout recurrence, including: S3.1, use the Lasso method to select the variables that have the greatest impact on the model, including: The goal of the Lasso method is to minimize the following loss function: Among them, y i is the target value of the i-th observation, x i is the jth feature of the i-th observation, β0 is the intercept term, β j is the regression characteristic coefficient of the jth feature, λ is the regularization parameter that controls the strength of regularization, n is the number of samples, p is the number of features, and L1 regularization term Some coefficients are compressed to zero, thereby achieving feature selection and obtaining a data set (x1, x2, ..., x p ) T ; S3.2, the filtered data are substituted into the Cox proportional hazard regression model for regression to obtain the regression characteristic coefficients β1, β2, ...β p , where the estimator β j satisfy: in, i=1,...,n, represents the sample composition of the Cox model, T i represents the survival time, C i represents the censoring time, δ i represents the event identification vector; S3.3, using regression characteristic coefficients β1, β2, ... β p For the data set (x1,x2,...,x p ) T By formula Transformation, obtain a new data set (x1',x2',...,x p ') T ; S3.4, divide the new dataset into training set and validation set according to the preset ratio; S3.5, training the competing risk model using the training set and testing and evaluating it using the validation set to obtain the competing risk model; S3.

6. The optimal cutoff value of NLR was determined using the Grey test statistic maximization method in the competing risk model and the log-rank test statistic maximization method in the Cox proportional hazards survival model. Sensitivity analysis was performed using the competing risk model and Kaplan-Meier survival analysis to confirm the robustness of the research results under different model settings.

5. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 4, characterized in that: In S4, when the evaluation association judgment condition is passed, NLR is added to the baseline model as a characteristic variable of the prognostic factor of gout recurrence risk to construct multiple candidate prediction models for gout recurrence risk, including: The evaluation association judgment conditions are: To determine whether the neutrophil-to-lymphocyte ratio (NLR) is associated with an increased risk of gout recurrence during hospitalization and whether it improves predictive accuracy compared with traditional markers such as serum uric acid (UA), tophi, and C-reactive protein (CRP).

6. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 5, characterized in that: S5, respectively using C statistics, net reclassification index NRI and integrated discrimination improvement index IDI to evaluate the multiple candidate gout recurrence risk prediction models for evaluating the prediction effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model, including: Construct multiple candidate prediction models for gout recurrence risk, including at least: Cox regression model, support vector machine (SVM) model, k-nearest neighbor (KNN) model, random survival forest model, and XGBoost model; The incremental predictive performance of NLR was evaluated by C-statistic, net reclassification index (NRI), and integrated discrimination improvement index (IDI), and compared to screen multiple candidate prediction models for gout recurrence risk. The ROC curve was generated by each candidate prediction model for gout recurrence risk and the area under the curve (AUC) of each model was calculated to evaluate the prediction accuracy; The differences between AUCs were compared using the DeLong test; Obtaining the best performing candidate prediction model as the target gout recurrence risk assessment model; The net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indices were calculated to assess the incremental predictive value of incorporating NLR into the baseline model, and confidence intervals were obtained by bootstrap resampling to optimize the target gout recurrence risk assessment model.

7. A method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to claim 6, characterized in that: S6, predicting the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generating a corresponding risk assessment result, including: inputting the patient's clinical data into the target gout recurrence risk assessment model; Output potential risk assessment results.

8. A gout recurrence risk assessment system based on the neutrophil-to-lymphocyte ratio, applied to a gout recurrence risk assessment method based on the neutrophil-to-lymphocyte ratio as claimed in any one of claims 1 to 7, characterized in that: The system comprises: The data collection module is used to collect electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data of the same target population within a preset time period from the data center system; A data processing module is used to pre-process the electronic medical record data, clinical pathology data, hospitalization and clinical follow-up data, and generate gout recurrence risk assessment sample data through data standardization; The NLR evaluation module is used to establish a combined model of the Cox proportional hazards regression model and the competing risk model to evaluate the association between the neutrophil-to-lymphocyte ratio (NLR) and gout recurrence; A model building module is used to add NLR as a characteristic variable of the prognostic factor of gout recurrence risk to the baseline model when the evaluation judgment conditions are met, so as to construct multiple candidate prediction models for gout recurrence risk; A model selection module is used to evaluate the multiple candidate gout recurrence risk prediction models using C statistics, net reclassification index NRI, and integrated discrimination improvement index IDI to evaluate the prediction effect of gout recurrence, so as to obtain a target gout recurrence risk assessment model; A risk assessment module, configured to predict the patient's gout recurrence risk based on the patient's clinical data and the target gout recurrence risk assessment model, and generate corresponding risk assessment results; The comparative display module is used to construct a visual interface and display, evaluate the optimal threshold and predictive value of the neutrophil-to-lymphocyte ratio (NLR) for gout recurrence, and compare its efficacy with traditional markers.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement a gout recurrence risk assessment method based on the neutrophil-to-lymphocyte ratio as described in any one of claims 1 to 7 when executing the instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and the program instructs a device to execute a method for assessing the risk of gout recurrence based on the neutrophil-to-lymphocyte ratio according to any one of claims 1 to 7.

Images