Method and system for developing primary sicca syndrome disease screening model

Abstract

The invention provides a primary sicca syndrome disease screening model development method and system. The method comprises the following steps: collecting tongue picture factors and tongue pictures of subjects including primary sicca syndrome patients and healthy individuals; the tongue picture factors are analyzed through single-factor analysis of two groups of data of the whole subject and the training set; developing and constructing an optimal tongue picture feature model; performing internal verification on the tongue picture feature model by adopting a bootstrap method, and performing external test on the optimal prediction model by using test set data; and displaying the optimal prediction model by using a webpage, and calculating the disease probability through the optimal prediction model. According to the method, the efficient and accurate screening model is constructed by adopting a machine learning algorithm, a simple, convenient and feasible screening means is provided, and the reliability of the model is improved; a prediction result is given in real time, the rapidity and accuracy of screening are improved, and the screening precision is improved; operation is easy and convenient, applicability is high, popularization and promotion are convenient, medical resources are saved, and patient experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of primary Sjögren's syndrome disease screening models; in particular, to a primary Sjögren's syndrome disease screening model development method and system. Background Art

[0002] Primary Sjögren's syndrome is a systemic autoimmune disease characterized by lymphocytic infiltration, mainly affecting the salivary and lacrimal glands. Primary Sjögren's syndrome also presents with a variety of extraglandular manifestations, leading to diversity in disease classification. Since the first screening classification criteria for Sjögren's syndrome (SS) were established in 1965, 13 different criteria have been proposed. Among them, the American-European Consensus Group Classification criteria (AECG) is widely used, which includes six evaluation items, two of which are subjective symptoms. The 2012 American College of Rheumatology (ACR) SS classification eliminated the influence of subjective symptoms and improved the specificity of screening. The 2016 American College of Rheumatology / European League Against Rheumatism (ACR-EULAR) classification criteria place more emphasis on extraglandular manifestations than previous classification criteria. Despite these updates, minor salivary gland biopsy (MSGB) and serum autoantibody testing remain essential for screening for primary Sjögren's syndrome.

[0003] However, at present, there is a lack of means to develop screening models for primary Sjögren's syndrome, and a lack of high-quality non-invasive screening methods for primary Sjögren's syndrome. Some patients suspected of primary Sjögren's syndrome cannot be screened due to fear of biopsy or lack of antibody testing and pathological screening conditions. Summary of the invention

[0004] In view of this, the purpose of the present invention is to design and develop a model for online predictive screening of primary Sjögren's syndrome based on tongue characteristics. The model adopts a machine learning algorithm to construct an efficient and accurate screening model by training a large number of tongue characteristics, thereby providing a simple and easy screening method for primary Sjögren's syndrome, and improving the reliability of the model based on a large amount of data training. By filling in the user's tongue characteristics, the web page can automatically analyze and give the screening results in real time, thereby improving the speed, accuracy and precision of the screening, and making the operation simple and highly applicable, so as to facilitate popularization and application.

[0005] The present invention provides a method for developing a primary Sjögren's syndrome disease screening model, comprising the following steps:

[0006] S1. Collect tongue image factors and tongue pictures of subjects including patients with primary Sjögren's syndrome and healthy individuals. The tongue image factors include age and gender. The tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating.

[0007] Exemplarily, patients with primary Sjögren's syndrome who were treated in the Department of Traditional Chinese Medicine Rheumatology of a hospital from December 2021 to December 2023 and healthy individuals who were examined in a physical examination center were included.

[0008] Inclusion criteria included: (1) patients with primary Sjögren's syndrome who met the 2016 ACR-EULAR classification criteria for primary Sjögren's syndrome; and (2) patients aged 18-75 years.

[0009] Exclusion criteria included: (1) concomitant with other connective tissue diseases, such as rheumatoid arthritis and systemic lupus erythematosus; (2) concomitant with tumors, infections, severe cardiovascular or cerebrovascular diseases, and metabolic diseases; (3) long-term use of antibiotics or hormones; and (4) the presence of oral diseases, such as gum disease.

[0010] Age, gender and tongue images of patients with primary Sjögren's syndrome and healthy individuals were collected. Tongue images were taken using the DS01-B tongue information acquisition system. On the day of taking the tongue image, it was confirmed that the participants had not consumed any food or medication that might change the color of the tongue or the color of the tongue coating, and had not eaten or drunk any water within one hour before taking the image. The integrity of the tongue images taken was checked. The tongue features were determined by two senior physicians with extensive clinical experience in TCM rheumatology. If there was a disagreement between the two physicians, a decision was made after discussion by the team. The tongue images recorded two dimensions: the tongue body and the tongue coating. The tongue body manifestations included light red tongue, light white tongue, red / purple tongue, cyanotic tongue, fat tongue, thin and thin tongue, cracked tongue, and tooth-marked tongue; the tongue coating manifestations included white coating, yellow coating, thick coating, thin coating, greasy coating, dry coating, and exfoliated coating. In the tongue body manifestations, red / purple tongue includes red tongue and cyanotic tongue. The cyanotic tongue includes the entire body being blue or purple, and also includes cyanotic spots of varying sizes appearing on the surface of the tongue.

[0011] S2. Analyze tongue image factors through univariate analysis of the data of the whole subjects and the training set; in the univariate analysis, include tongue image factors with a P value less than 0.1 into the binary logistic regression model, and develop and construct the optimal tongue image characteristic model through variable screening and model comparison;

[0012] S3, using the bootstrap method to perform internal verification on the tongue image feature model, and using the test set data to perform external testing on the tongue image feature model;

[0013] S4. Use a web page to display the best prediction model. Fill in the user's basic tongue information, gender and age through the web page. The probability of disease is calculated by the best prediction model and displayed on the web page.

[0014] Preferably, in the statistical calculation of the above-mentioned processing process, SPSS 27.0 and R4.3.2 software are used, the count data are expressed as frequency and percentage, and the non-normal distribution data are expressed as median and quartile (Q1, Q3); in the comparison between groups, the chi-square test is used to compare the count data, and the Mann-Whitney U test is used to compare the measurements that do not conform to the normal distribution; R4.3.2 is used for random segmentation of data sets, model construction, model verification, model testing and model presentation; two-sided tests are used for hypothesis testing, and P < 0.05 is considered to be a statistically significant level of significance.

[0015] Furthermore, the method for developing and constructing the optimal tongue image feature model in step S2 comprises the following steps:

[0016] S21. Develop three models for tongue image characteristics: a comprehensive model, a tongue body model, and a tongue coating model, wherein the comprehensive model includes the tongue body and the tongue coating, the tongue body model only includes the tongue body, and the tongue coating model only includes the tongue coating. The comprehensive model, the tongue body model, and the tongue coating model all use gender and age as correction factors.

[0017] S22. Use Akaike information criterion AIC combined with stepwise regression to screen and optimize the comprehensive model, tongue body model and tongue coating model.

[0018] Furthermore, the method for screening and optimizing the comprehensive model, tongue body model and tongue coating model in step S22 includes:

[0019] Among the three tongue image characteristic models, namely the comprehensive model, tongue body model and tongue coating model, the model discrimination was evaluated by the receiver operating characteristic curve (ROC), the model calibration was evaluated by the calibration curve, the overall performance of the model was evaluated by the Brier score, and the clinical applicability of the model was evaluated by the decision curve analysis (DCA). The best prediction model for screening primary Sjögren's syndrome based on tongue image characteristics was selected.

[0020] Furthermore, the method for internally verifying the tongue image feature model in step S3 includes:

[0021] The bootstrap method was used to internally validate the tongue image feature model. The Jordan index of the optimal prediction model was calculated using the training set data, and the optimal cutoff value for screening primary Sjögren's syndrome was found; the sensitivity, specificity, positive predictive value and negative predictive value of the optimal prediction model for screening primary Sjögren's syndrome were calculated in the training set.

[0022] Furthermore, the method of performing external testing on the tongue image feature model in step S3 includes:

[0023] For the external validation data set, based on the best model discrimination evaluated by the operating characteristic curve (ROC) and the model calibration evaluated by the calibration curve, the clinical characteristic curve (DCA) was calculated to evaluate the clinical applicability of the model, and the best prediction model for screening primary Sjögren's syndrome based on tongue features was selected.

[0024] The sensitivity, specificity, positive predictive value, and negative predictive value of the optimal prediction model for screening primary Sjögren's syndrome were calculated in the test set.

[0025] Furthermore, the method of using a web page to display the optimal prediction model in step S4 includes:

[0026] The nomogram of the best prediction model was drawn and the best prediction model was explained by the nomogram.

[0027] The present invention also provides a primary Sjögren's syndrome disease screening model development system, which executes the primary Sjögren's syndrome disease screening model development method as described above, comprising:

[0028] Data collection module: used to collect tongue image factors and tongue pictures of subjects including primary Sjögren's syndrome patients and healthy individuals, wherein the tongue image factors include age and gender, and the tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating;

[0029] Model building and evaluation module: used to analyze tongue image factors through univariate analysis of two groups of data: the whole subject and the training set; in the univariate analysis, the tongue image factors with a P value less than 0.1 are included in the binary logistic regression model, and the optimal tongue image characteristic model is developed and constructed through variable screening and model comparison;

[0030] Model verification and testing module: used to use the bootstrap method to perform internal verification on the tongue image feature model, and use the test set data to perform external testing on the tongue image feature model;

[0031] Model display module: Use a web page to display the best prediction model. Fill in the user's basic tongue information, gender and age through the web page. The probability of disease occurrence is calculated by the best prediction model and displayed on the web page.

[0032] Furthermore, the model building and evaluation module includes:

[0033] Model building unit: used to develop three models for tongue image characteristics: comprehensive model, tongue body model, tongue coating model, wherein the comprehensive model includes tongue body and tongue coating, the tongue body model only includes tongue body, the tongue coating model only includes tongue coating, and the comprehensive model, tongue body model and tongue coating model all use gender and age as correction factors;

[0034] Screening and optimization unit: used to screen and optimize the comprehensive model, tongue body model and tongue coating model using the Akaike information criterion combined with stepwise regression.

[0035] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of the method for developing a primary Sjögren's syndrome disease screening model as described above are implemented.

[0036] The present invention also provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for developing a primary Sjögren's syndrome disease screening model as described above are implemented.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The primary Sjögren's syndrome disease screening model development method and system provided by the present invention adopt a machine learning algorithm, and construct an efficient and accurate screening model by training a large number of tongue image features, thereby providing a simple and easy screening method for primary Sjögren's syndrome, and improving the reliability of the model based on a large amount of data training; by filling in the user's tongue image features, the web page automatically analyzes and gives the screening results, thereby improving the screening speed, giving the predicted results in real time, improving the rapidity and accuracy of the screening, and improving the screening accuracy; and the operation is simple, and ordinary doctors and patients can also use it independently, with high applicability, easy popularization and promotion, saving medical resources, and improving patient experience, and has broad prospects for promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0040] In the attached picture:

[0041] Figure 1 A flow chart for constructing a primary Sjögren's syndrome screening model according to an embodiment of the present invention;

[0042] Figure 2 : is a ROC curve and a calibration curve diagram of the Logistic regression model of an embodiment of the present invention;

[0043] Figure 3 Graphs showing the ROC curve and DCA curve of the best prediction model in the internal validation of the embodiment of the present invention;

[0044] Figure 4 The ROC curve, calibration curve and DCA curve diagram of the best prediction model in the test set of the embodiment of the present invention;

[0045] Figure 5 This is a web page diagram of the best prediction model for primary Sjögren's syndrome screening based on tongue performance according to an embodiment of the present invention;

[0046] Figure 6 A web page diagram of an embodiment of the present invention for calculating the probability of a disease by filling in the user's basic tongue information, gender and age through a web page;

[0047] Figure 7 A flow chart of a method for developing a primary Sjögren's syndrome disease screening model according to the present invention;

[0048] Figure 8 A flow chart of the method for developing and constructing an optimal tongue image characteristic model for the present invention;

[0049] Fig. 9 The figure is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0050] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present disclosure. Instead, they are merely examples of systems and products consistent with some aspects of the present disclosure as detailed in the appended claims.

[0051] The terms used in this disclosure are for the purpose of describing specific embodiments only and are not intended to limit the disclosure. The singular forms of "a", "said" and "the" used in this disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0052] It should be understood that although the terms first, second, third, etc. may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present disclosure, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0053] The embodiments of the present invention are described in further detail below.

[0054] The present invention provides a method for developing a primary Sjögren's syndrome disease screening model. Figure 7 As shown, the following steps are included:

[0055] S1. Collect tongue image factors and tongue pictures of subjects including patients with primary Sjögren's syndrome and healthy individuals. The tongue image factors include age and gender. The tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating.

[0056] In this embodiment, patients with primary Sjögren's syndrome who were treated in the Department of Traditional Chinese Medicine Rheumatology of a hospital from December 2021 to December 2023 and healthy individuals who were examined in a physical examination center were included.

[0057] Inclusion criteria included: (1) patients with primary Sjögren's syndrome who met the 2016 ACR-EULAR classification criteria for primary Sjögren's syndrome; and (2) patients aged 18-75 years.

[0058] Exclusion criteria included: (1) concomitant with other connective tissue diseases, such as rheumatoid arthritis and systemic lupus erythematosus; (2) concomitant with tumors, infections, severe cardiovascular or cerebrovascular diseases, and metabolic diseases; (3) long-term use of antibiotics or hormones; and (4) the presence of oral diseases, such as gum disease.

[0059] Age, sex, and tongue images (e.g. Figure 1). Tongue images were taken using the DS01-B tongue information acquisition system. On the day of taking the tongue image, it was confirmed that the participants had not consumed any food or medication that might change the color of the tongue or the color of the tongue coating, and had not eaten or drunk anything within one hour before taking the image. The integrity of the tongue images taken was checked. The tongue features were determined by two senior physicians with extensive clinical experience in TCM rheumatology. If there was a disagreement between the two physicians, a decision was made after discussion by the team. The tongue images recorded two dimensions: the tongue body and the tongue coating. The tongue body manifestations included pale red tongue, pale white tongue, red / crimson tongue, cyanotic tongue, fat tongue, thin and thin tongue, cracked tongue, and tooth-marked tongue; the tongue coating manifestations included white coating, yellow coating, thick coating, thin coating, greasy coating, dry coating, and exfoliated coating. In the tongue body manifestations, red / crimson tongue includes red tongue and crimson tongue. Cyanotic tongue includes blue or purple as a whole, and also includes cyanotic spots of varying sizes appearing on the surface of the tongue.

[0060] S2. Analyze tongue image factors through univariate analysis of the data of the whole subjects and the training set; in the univariate analysis, include tongue image factors with a P value less than 0.1 into the binary logistic regression model, and develop and construct the optimal tongue image characteristic model through variable screening and model comparison;

[0061] The method for developing an optimal tongue feature model includes the following steps (see Figure 8 shown):

[0062] S21. Develop three models for tongue image characteristics: a comprehensive model, a tongue body model, and a tongue coating model, wherein the comprehensive model includes the tongue body and the tongue coating, the tongue body model includes only the tongue body, and the tongue coating model includes only the tongue coating, and the comprehensive model, the tongue body model, and the tongue coating model all use gender and age as correction factors;

[0063] S22. Akaike information criterion combined with stepwise regression was used to screen and optimize the comprehensive model, tongue body model and tongue coating model.

[0064] In the three tongue image feature models, namely the comprehensive model, tongue body model and tongue coating model, the model discrimination was evaluated by the receiver operating characteristic curve (ROC), the model calibration was evaluated by the calibration curve, the overall performance of the model was evaluated by the Brier score, and the clinical applicability of the model was evaluated by the decision curve analysis (DCA).

[0065] S3, using the bootstrap method to perform internal verification on the tongue image feature model, and using the test set data to perform external testing on the tongue image feature model;

[0066] Among the three tongue image feature models, namely the comprehensive model, tongue body model and tongue coating model, the model discrimination evaluated by the operating characteristic curve ROC and the model calibration evaluated by the calibration curve were used to calculate the clinical applicability of the model evaluated by the clinical characteristic curve DCA, and the optimal prediction model for screening primary Sjögren's syndrome based on tongue image characteristics was selected.

[0067] The bootstrap method was used to internally validate the tongue image feature model, the Jordan index of the optimal prediction model was calculated using the training set data, and the optimal cutoff value for screening primary Sjögren's syndrome was found;

[0068] The sensitivity, specificity, positive predictive value, and negative predictive value of the optimal prediction model for screening primary Sjögren's syndrome were calculated in both the training set and the test set.

[0069] S4, using a web page to display the optimal prediction model, and filling in the user's basic tongue information, gender and age (such as Figure 6 As shown), the probability of disease occurrence is calculated by the optimal prediction model and displayed on the web page.

[0070] In addition, in this embodiment, a web page graph of the best prediction model is drawn, and the best prediction model is explained through the web page graph.

[0071] Specifically, in the statistical calculations of the above processing process, SPSS 27.0 and R4.3.2 software were used for data processing. SPSS 27.0 was used for data comparison, and R4.3.2 was used for random segmentation of data sets, model construction, model verification, model testing, and model presentation; two-sided tests were used for hypothesis testing, and P < 0.05 was considered to be a statistically significant level.

[0072] The embodiment of the present invention further provides a primary Sjögren's syndrome disease screening model development system, which executes the primary Sjögren's syndrome disease screening model development method as described above, including:

[0073] Data collection module: used to collect tongue image factors and tongue pictures of subjects including primary Sjögren's syndrome patients and healthy individuals, wherein the tongue image factors include age and gender, and the tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating;

[0074] Model building and evaluation module: used to analyze tongue image factors through univariate analysis of two groups of data: the whole subject and the training set; in the univariate analysis, the tongue image factors with a P value less than 0.1 are included in the binary logistic regression model, and the optimal tongue image characteristic model is developed and constructed through variable screening and model comparison;

[0075] The model building and evaluation module includes:

[0076] Model building unit: used to develop three models for tongue image characteristics: comprehensive model, tongue body model, tongue coating model, wherein the comprehensive model includes tongue body and tongue coating, the tongue body model only includes tongue body, the tongue coating model only includes tongue coating, and the comprehensive model, tongue body model and tongue coating model all use gender and age as correction factors;

[0077] Screening and optimization unit: used to screen and optimize the comprehensive model, tongue body model and tongue coating model using the Akaike information criterion combined with stepwise regression;

[0078] Model verification and testing module: used to use the bootstrap method to perform internal verification on the tongue image feature model, and use the test set data to perform external testing on the tongue image feature model;

[0079] Model display module: Use a web page to display the best prediction model. Fill in the user's basic tongue information, gender and age through the web page. The probability of disease occurrence is calculated by the best prediction model and displayed on the web page.

[0080] Application Examples

[0081] (1) Analysis of tongue characteristics:

[0082] The results of univariate analysis of the data of the whole subjects and the two groups in the training set are shown in Table 1. In the training set, the two groups of subjects had statistical differences in gender, age, light red tongue, cyanotic tongue, thin and thin tongue, fissured tongue, tooth-marked tongue, thin fur, thick fur, dry fur, greasy fur, exfoliated fur, and white fur (p < 0.1), and these factors (variables) were included in the subsequent development of the tongue image feature model.

[0083] Table 1 Basic clinical information of the subjects

[0084]

[0085] pSS: primary Sjögren's syndrome

[0086] (2) Model construction and evaluation:

[0087] After variable selection, this application example constructed three tongue image feature models based on the Logistic regression model: comprehensive model, tongue body model, and tongue coating model.

[0088] AIC is a widely used model selection criterion, and when comparing multiple models, the model with the smallest AIC value is selected. As shown in Table 2, among the three tongue image feature models, model 1 has the smallest AIC value. Figure 2 shows the ROC curves of all three tongue image feature models, among which model 1 has the largest area under the curve (AUC). The Brier score is an indicator for evaluating the calibration accuracy of a classification model, which is done by measuring the difference between the predicted probability and the actual observed result. In this case, model 1 showed the largest AUC and the lowest Brier score, indicating that model 1 has superior discrimination ability and calibration accuracy compared with model 2 and model 3 (A in Figure 2). The calibration curves were drawn on the training set using the bootstrap method, showing the good calibration performance of the three models (BD in Figure 2). In summary, model 1 was determined to be the best predictive model for screening primary Sjögren's syndrome based on tongue image. The ROC curve of the training set showed that the best cutoff value for screening primary Sjögren's syndrome using model 1 was 0.347. In the training set, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Youden index for screening primary Sjögren's syndrome using the best prediction model were 87.86%, 76.43%, 78.85%, 86.29%, and 0.64, respectively.

[0089] Table 2: Variable characteristics of the logistic regression model for tongue screening of primary Sjögren's syndrome

[0090]

[0091] Model 1: comprehensive model (including tongue body and tongue coating); Model 2: tongue coating model; Model 3: tongue body model. AIC: Akaike information criterion.

[0092] Figure 2 The ROC curve and calibration curve of the logistic regression model are shown. Figure 2 middle:

[0093] A is the ROC curve of the three models.

[0094] B is the calibration curve of model 1 (comprehensive model: tongue body and tongue coating).

[0095] C is the calibration curve of model 2 (tongue coating model).

[0096] D is the calibration curve of model 3 (tongue model).

[0097] The red line is: Model 1 (comprehensive model: tongue body and tongue coating), AUC=0.89.

[0098] The blue line is: Model 2 (tongue coating model), AUC=0.88.

[0099] The green line is: Model 3 (tongue model), AUC=0.82.

[0100] ROC is: receiver operating characteristic;

[0101] AUC is: Area Under the Curve.

[0102] (3) Internal validation of the model:

[0103] The bootstrap method was used to internally validate model 1. The results showed that model 1 had good clinical prediction accuracy with an AUC of 0.88 ( Figure 3 A in ).

[0104] DCA (decision curve analysis) evaluated the performance of the prediction model in clinical decision making. Internally validated DCA also showed good clinical applicability of model 1 ( Figure 3 B in the figure).

[0105] Figure 3 The ROC curve and DCA curve of the best prediction model in internal validation are shown.

[0106] Figure 3 A in the figure is the ROC curve of the best prediction model in internal validation. AUC=0.88.

[0107] Figure 3 B in the figure is the DCA curve of the best prediction model in internal validation.

[0108] ROC stands for: Receiver Operating Characteristic.

[0109] DCA stands for: Decision Curve Analysis.

[0110] (4) External testing of the model:

[0111] Use the test set to perform external testing on Model 1 and plot the ROC curve, calibration curve, and DCA curve of Model 1 in the external testing (e.g. Figure 4 The AUC results show that model 1 has a higher screening accuracy on the test set ( Figure 4 When the screening decision threshold of 0.347 obtained in the training set was applied to the test set, the sensitivity, specificity, positive predictive value, and negative predictive value of the model for primary Sjögren's syndrome were 93.33%, 71.67%, 76.71%, and 91.48%, respectively.

[0112] The calibration curve of model 1 shows that the model has excellent accuracy ( Figure 4 B in the figure).

[0113] DCA showed that model 1 had good clinical applicability ( Figure 4 C in.

[0114] Figure 4 The ROC curve, calibration curve, and DCA curve of the best prediction model in the test set are shown. Figure 4 A in the figure is the ROC curve of the best prediction model in the test set. AUC=0.93.

[0115] Figure 4 B in the figure is the calibration curve of the best prediction model in the test set.

[0116] Figure 4 C in the figure is the DCA curve of the best prediction model in the test set.

[0117] ROC stands for: Receiver Operating Characteristic.

[0118] DCA stands for: Decision Curve Analysis.

[0119] (5) Model display:

[0120] Use a web page to display the best prediction model (such as Figure 5 As shown). Through this web page diagram, after the user fills in the basic tongue information, gender and age, the web page will calculate the probability of disease through the above-mentioned optimal prediction model.

[0121] Figure 5 Web page showing the best prediction model for primary Sjögren's syndrome screening based on tongue appearance. Based on the individual's tongue appearance, gender, and age, the web page will calculate the corresponding screening probability based on this study. Figure 6 Depicts the screening probability of primary Sjögren's syndrome in a 50-year-old woman who presented with an exfoliative, dark purple, cracked tongue with a thin coating but a fat tongue, with a prevalence of primary Sjögren's syndrome of 85.1%.

[0122] An embodiment of the present invention further provides a computer device, Fig. 9 is a schematic diagram of the structure of a computer device provided by an embodiment of the present invention; see the accompanying drawings Fig. 9 As shown, the computer device includes: an input system 23, an output system 24, a memory 22 and a processor 21; the memory 22 is used to store one or more programs; when the one or more programs are executed by the one or more processors 21, the one or more processors 21 implement the primary Sjögren's syndrome disease screening model development method provided in the above embodiment; wherein the input system 23, the output system 24, the memory 22 and the processor 21 can be connected via a bus or other means, Fig. 9The example of connecting through bus is taken in the following.

[0123] The memory 22 is a readable and writable storage medium of a computing device, and can be used to store software programs, computer executable programs, such as the program instructions corresponding to the primary Sjögren's syndrome disease screening model development method described in the embodiment of the present invention; the memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required for a function; the data storage area can store data created according to the use of the device, etc.; in addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device; in some instances, the memory 22 can further include a memory remotely arranged relative to the processor 21, and these remote memories can be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0124] The input system 23 may be used to receive input digital or character information, and to generate key signal input related to user settings and function control of the device; the output system 24 may include display devices such as display screens.

[0125] The processor 21 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory 22, that is, realizes the above-mentioned primary Sjögren's syndrome disease screening model development method.

[0126] The computer device provided above can be used to execute the primary Sjögren's syndrome disease screening model development method provided in the above embodiment, and has corresponding functions and beneficial effects.

[0127] The embodiment of the present invention also provides a storage medium containing computer executable instructions, which are used to execute the primary Sjögren's syndrome disease screening model development method provided by the above embodiment when executed by a computer processor, and the storage medium is any of various types of memory devices or storage devices, and the storage medium includes: installation media, such as CD-ROM, floppy disk or tape system; computer system memory or random access memory, such as DRAM, DDRRAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc.; the storage medium may also include other types of memory or a combination thereof; in addition, the storage medium may be located in a first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet); the second computer system may provide program instructions to the first computer for execution. The storage medium includes two or more storage media that can reside in different locations (for example, in different computer systems connected via a network). The storage medium can store program instructions (for example, specifically implemented as a computer program) that can be executed by one or more processors.

[0128] Of course, the storage medium containing computer executable instructions provided in an embodiment of the present invention, whose computer executable instructions are not limited to the primary Sjögren's syndrome disease screening model development method described in the above embodiment, can also execute related operations in the primary Sjögren's syndrome disease screening model development method provided in any embodiment of the present invention.

[0129] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0130] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for developing a primary Sjögren's syndrome disease screening model, characterized in that: The following steps are involved: S1. Collect tongue image factors and tongue pictures of subjects including patients with primary Sjögren's syndrome and healthy individuals. The tongue image factors include age and gender. The tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating. S2. Analyze tongue image factors through univariate analysis of the data of the whole subjects and the training set; in the univariate analysis, include tongue image factors with a P value less than 0.1 into the binary logistic regression model, and develop and construct the optimal tongue image characteristic model through variable screening and model comparison; S3, using the bootstrap method to perform internal verification on the tongue image feature model, and using the test set data to perform external testing on the tongue image feature model; S4. Use a web page to display the best prediction model. Fill in the user's basic tongue information, gender and age through the web page. The probability of disease is calculated by the best prediction model and displayed on the web page.

2. The method for developing a primary Sjögren's syndrome disease screening model according to claim 1, characterized in that: The method for developing and constructing the optimal tongue image characteristic model in step S2 comprises the following steps: S21. Develop three models for tongue image characteristics: a comprehensive model, a tongue body model, and a tongue coating model, wherein the comprehensive model includes the tongue body and the tongue coating, the tongue body model includes only the tongue body, and the tongue coating model includes only the tongue coating, and the comprehensive model, the tongue body model, and the tongue coating model all use gender and age as correction factors; S22. Akaike information criterion combined with stepwise regression was used to screen and optimize the comprehensive model, tongue body model and tongue coating model.

3. The method for developing a primary Sjögren's syndrome disease screening model according to claim 2, characterized in that: The method for screening and optimizing the comprehensive model, tongue body model and tongue coating model in step S22 includes: Among the three tongue image feature models, namely the comprehensive model, tongue body model and tongue coating model, the model discrimination was evaluated by the receiver operating characteristic curve (ROC), the model calibration was evaluated by the calibration curve, the overall performance of the model was evaluated by the Brier score, and the clinical applicability of the model was evaluated by the decision curve analysis (DCA), and the optimal prediction model for screening primary Sjögren's syndrome based on tongue image characteristics was selected.

4. The method for developing a primary Sjögren's syndrome disease screening model according to claim 3, characterized in that: The method for internally verifying the tongue image characteristic model in step S3 includes: The bootstrap method was used to internally validate the tongue image feature model, and the Jordan index of the optimal prediction model was calculated using the training set data, and the optimal cutoff value for screening primary Sjögren's syndrome was found; the sensitivity, specificity, positive predictive value and negative predictive value of the optimal prediction model for screening primary Sjögren's syndrome were calculated in the training set.

5. The method for developing a primary Sjögren's syndrome disease screening model according to claim 4, characterized in that: The method of performing external testing on the tongue image characteristic model in step S3 comprises: For the external validation data set, based on the best model discrimination evaluated by the operating characteristic curve (ROC) and the model calibration evaluated by the calibration curve, the clinical characteristic curve (DCA) was calculated to evaluate the clinical applicability of the model, and the best prediction model for screening primary Sjögren's syndrome based on tongue features was selected; The sensitivity, specificity, positive predictive value, and negative predictive value of the optimal prediction model for screening primary Sjögren's syndrome were calculated in the test set.

6. The method for developing a primary Sjögren's syndrome disease screening model according to claim 1, characterized in that: The method of using a web page to display the optimal prediction model in step S4 includes: Draw a visualization of the best prediction model and explain it through a visualization.

7. A primary Sjögren's syndrome disease screening model development system, which executes the primary Sjögren's syndrome disease screening model development method according to any one of claims 1 to 6, characterized in that: include: Data collection module: used to collect tongue image factors and tongue pictures of subjects including primary Sjögren's syndrome patients and healthy individuals, wherein the tongue image factors include age and gender, and the tongue pictures record the tongue image characteristics in two dimensions: tongue body and tongue coating; Model building and evaluation module: used to analyze tongue image factors through univariate analysis of two groups of data: the whole subject and the training set; in the univariate analysis, the tongue image factors with a P value less than 0.1 are included in the binary logistic regression model, and the optimal tongue image characteristic model is developed and constructed through variable screening and model comparison; Model verification and testing module: used to use the bootstrap method to perform internal verification on the tongue image feature model, and use the test set data to perform external testing on the tongue image feature model; Model display module: Use a web page to display the best prediction model. Fill in the user's basic tongue information, gender and age through the web page. The probability of disease occurrence is calculated by the best prediction model and displayed on the web page.

8. The primary Sjögren's syndrome disease screening model development system according to claim 7, characterized in that: The model building and evaluation module includes: Model building unit: used to develop three models for tongue image characteristics: comprehensive model, tongue body model, tongue coating model, wherein the comprehensive model includes tongue body and tongue coating, the tongue body model only includes tongue body, the tongue coating model only includes tongue coating, and the comprehensive model, tongue body model and tongue coating model all use gender and age as correction factors; Screening and optimization unit: used to use the Akaike information criterion combined with stepwise regression to screen and optimize the comprehensive model, tongue body model and tongue coating model, and finally screen out the optimal model.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for developing a primary Sjögren's syndrome disease screening model described in any one of claims 1 to 6 are implemented.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method for developing a primary Sjögren's syndrome disease screening model as described in any one of claims 1 to 6 are implemented.

Images