Accurate early warning auxiliary diagnosis model system based on high-risk HPV positive

Abstract

The invention discloses an accurate early warning auxiliary diagnosis model system based on high-risk HPV positive, and relates to the technical field of medical diagnosis, according to the accurate early warning auxiliary diagnosis model system based on high-risk HPV positive, by obtaining the relevance and action intensity of conventional blood and urine test indexes and high-risk human papilloma virus (HR-HPV) infection, the accurate early warning auxiliary diagnosis model system based on high-risk HPV positive can be used for early warning of high-risk human papilloma virus (HR-HPV) infection. A basis is provided for deep research in the field; an HR-HPV infection positive early warning model taking health screening big data as a basis and hematuria biological indexes as elements is formed, and the problems that HPV detection depends on subjective judgment of doctors or active selection of clients, an accurate measurement basis is lacked, and the screening coverage rate is not enough are solved. Therefore, a simple, convenient, rapid and economical means is provided for gynecological evaluation, reliable indications are provided for HPV detection, the screening efficiency is improved, and help is provided for prevention of HPV infection and cervical diseases.

Description

Technical Field

[0001] The present invention relates to the field of medical diagnosis technology, and in particular to a precise early warning auxiliary diagnosis model system based on high-risk HPV positivity. Background Art

[0002] The existing accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity still has the following drawbacks in actual use:

[0003] High-risk human papillomavirus (HR-HPV) infection is a major risk factor for cervical cancer and precancerous lesions, posing a serious threat to women's health. Currently, clinical detection methods for HR-HPV infection have many limitations in their application.

[0004] Existing HPV testing relies heavily on the subjective judgment of doctors or the active choice of customers, and lacks precise quantitative basis. This affects the objectivity and accuracy of the test results, making missed or misdiagnosed cases more likely to occur. At the same time, due to limitations in testing methods and insufficient public awareness of HPV screening, the screening coverage rate is less than ideal, and many potential infected people are not discovered in time, missing the best opportunity for intervention and treatment.

[0005] Routine blood and urine tests are commonly used basic examination items in clinical diagnosis, and their indicators can reflect the physiological and pathological status of the human body to a certain extent; however, current research on the correlation and intensity of the relationship between routine blood and urine test indicators and HR-HPV infection is not in-depth and systematic enough, and no effective basis for positive warning of HR-HPV infection has yet to be formed. Summary of the Invention

[0006] The purpose of the present invention is to provide an accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity to solve the above-mentioned problems.

[0007] To achieve the above objectives, the present invention provides the following technical solutions: a precise early warning auxiliary diagnosis model system based on high-risk HPV positivity, comprising:

[0008] An indicator correlation analysis module is used to study the quantitative relationship between routine blood and urine biological indicators and HR-HPV positivity;

[0009] The indicator association analysis module uses a correlation analysis algorithm when studying quantitative relationships;

[0010] The model building module is used to analyze and determine useful indicator factors, establish a "determined indicator" and a positive warning model for HR-HPV based on the corresponding multivariate analysis algorithm, and determine the warning threshold or standard.

[0011] Furthermore, the indicator correlation analysis module includes:

[0012] The conventional blood and urine biological indicators include:

[0013] There are 24 blood routine indicators, including white blood cell count, red blood cell count, hemoglobin, platelet count, neutrophil percentage, etc.

[0014] There are 14 types of urine routine indicators, including urine protein, urine sugar, urine occult blood, urine bilirubin, urobilinogen, etc.

[0015] Liver function and 7 liver enzymes, specifically alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, indirect bilirubin, etc.;

[0016] There are three types of renal function indicators, specifically creatinine, urea nitrogen, and uric acid;

[0017] There are five blood lipid and blood sugar indicators, namely total cholesterol, triglycerides, fasting blood sugar, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol;

[0018] More than 7 other basic indicators, including age, body mass index, blood pressure, potassium ions, sodium ions, etc.;

[0019] The quantitative relationship includes correlation or distribution difference and effect intensity, and the HR-HPV includes 17 genotypes, specifically HPV16, HPV18, HPV31, HPV33, etc.

[0020] Furthermore, the correlation analysis algorithm includes a Pearson correlation coefficient algorithm, the formula of which is: ;

[0021] in, is the Pearson correlation coefficient, 、 The two variables are observations, 、 are the means of the two variables, is the number of observations.

[0022] Furthermore, the “determined index” is an index with high correlation and high effect strength; the multivariate analysis algorithm includes a logistic regression algorithm, and its formula is:

[0023] ;

[0024] in, is the predicted probability of HR-HPV positivity, is the intercept term, is the regression coefficient, is the value of the "determined indicator"; the multivariate analysis algorithm also includes a support vector machine and a random forest, and multiple algorithms can be used in combination or in a comprehensive manner.

[0025] Furthermore, when the model building module determines the warning threshold or standard, it draws the receiver operating characteristic curve, calculates the area under the curve, and selects the value corresponding to the maximum Youden index as the warning threshold. The calculation formula of the Youden index is: Youden index = sensitivity + specificity - 1, where sensitivity is the true positive rate and specificity is the true negative rate.

[0026] Furthermore, it also includes a model verification, evaluation and optimization module, which is used to evaluate the accuracy, stability and extrapolation reliability of the model through internal verification and external verification, and continuously adjust and optimize; the internal verification is the verification of modeling data, and is carried out at the same time as modeling, using the cross-validation method to divide the modeling data into Equal parts, each time 1 copy of data modeling, 1 copy of data verification, repeat The external verification includes verification of past data and verification of newly collected physical examination data, and the verification of newly collected physical examination data is further divided into verification of known test result data and verification of unknown result data.

[0027] Furthermore, when the model verification, evaluation and optimization module evaluates the accuracy of the model, it uses indicators such as accuracy, precision, and recall. The calculation formula for accuracy is: accuracy = (true positive + true negative) / (true positive + false positive + true negative + false negative); the calculation formula for precision is: precision = true positive / (true positive + false positive); the calculation formula for recall is: recall = true positive / (true positive + false negative).

[0028] Furthermore, it also includes a model application and long-term observation module, which is used to apply the optimized model to people undergoing physical examinations, etc. Specifically, the routine blood and urine biological indicator data of the physical examination population are obtained through the system interface, and the HR-HPV positive warning results are obtained after input into the model, and the results are fed back to relevant medical personnel; at the same time, long-term observation is carried out on the application population to record the degree of consistency between the actual HR-HPV infection situation and the warning results.

[0029] Furthermore, the model application and long-term observation module is also used to regularly feed back data to the model verification, evaluation and optimization module based on the long-term observation results to further optimize the model. The optimization content includes adjusting the selection of "determination indicators" and updating the parameters of the multivariate analysis algorithm.

[0030] Compared with the existing technology, the accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity provided by the present invention has the following beneficial effects:

[0031] This precise early warning auxiliary diagnosis model system based on high-risk HPV positivity provides a basis for in-depth research in this field by obtaining the correlation and effect intensity between routine blood and urine test indicators and high-risk human papillomavirus (HR-HPV) infection; it forms an HR-HPV infection positive early warning model based on health screening big data and blood and urine biological indicators as elements, and aims to solve the problems of HPV testing relying on the subjective judgment of doctors or active selection of customers, lack of precise quantitative basis and insufficient screening coverage; thereby providing a simple, fast and economical means for gynecological evaluation, providing reliable indications for HPV testing and thus improving screening efficiency, and providing assistance for the prevention of HPV infection and cervical disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0033] Figure 1 Schematic diagram of the ROC curve of the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0035] See also Figure 1 , a precise early warning auxiliary diagnosis model system based on high-risk HPV positivity, including:

[0036] Index correlation analysis module: The index correlation analysis module is used to study the quantitative relationship between routine blood and urine biological indicators and HR-HPV positivity;

[0037] When the indicator association analysis module studies the quantitative relationship, it uses the correlation analysis algorithm;

[0038] Model building module, the model building module is used to analyze and determine useful indicator factors, establish a "determined indicator" and HR-HPV positive warning model based on the corresponding multivariate analysis algorithm, and determine the warning threshold or standard.

[0039] The indicator correlation analysis module includes blood routine, urine biology, liver function and liver enzymes, kidney function, blood lipids and blood sugar and other basic indicators:

[0040] The blood test, urine biomarkers, liver function and liver enzymes, kidney function, blood lipids and blood sugar and other basic indicators include:

[0041] There are 24 blood routine indicators, including white blood cell count, neutrophil percentage, lymphocyte percentage, monocyte percentage, eosinophil percentage, basophil percentage, absolute neutrophil count, absolute lymphocyte count, absolute monocyte count, absolute eosinophil count, absolute basophil count, red blood cell count, hemoglobin concentration, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red blood cell distribution width, platelet count, hematocrit, mean platelet volume and platelet volume distribution width, as well as neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio;

[0042] There are 14 urine routine indicators, including urine occult blood, urine bilirubin, urine urobilinogen, urine ketones, urine protein, urine nitrite, urine glucose, urine pH, urine specific gravity, urine routine vitamin C, leukocyte esterase, urine microscopic examination of white blood cells, urine microscopic examination of red blood cells, and urine casts;

[0043] Liver function and 7 liver enzymes, specifically alanine aminotransferase, total bilirubin, direct bilirubin, indirect bilirubin, aspartate aminotransferase, γ-glutamyl transferase, and alkaline phosphatase;

[0044] There are three types of renal function indicators, specifically creatinine, urea nitrogen, and uric acid;

[0045] There are five blood lipid and blood sugar indicators, namely total cholesterol, triglycerides, fasting blood sugar, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol;

[0046] Other basic indicators: age, height, weight, body mass index, systolic blood pressure, diastolic blood pressure, pulse;

[0047] Quantitative relationships include correlation or distribution differences and effect strength, and HR-HPV includes 17 genotypes, specifically HPV16, HPV18, HPV31, HPV33, HPV35, HPV39, HPV45, HPV51, HPV52, HPV53, HPV56, HPV58, HPV59, HPV66, HPV68, HPV73, and HPV82, as shown in the table below.

[0048]

[0049] The area under the ROC curve for genotyping

[0050] Most genotypes have significantly higher predictions than the overall model, but HPV52 accounts for the largest proportion of these genotypes, and its poor results drag down the overall model's results. Because there are relatively few genotypes with good data, the models are relatively unstable, and the results may be biased upwards. However, these models have good predictive value for infection with individual genotypes.

[0051] Correlation analysis algorithms, including the Pearson correlation coefficient algorithm, are formulated as follows: ;

[0052] in, is the Pearson correlation coefficient, 、 The two variables are observations, 、 are the means of the two variables, is the number of observations, where represents the i cases in a variable. Among the two variables, y is only the HPV test result, and x is the corresponding various possible related variables, such as age, white blood cell count, etc.

[0053] "Determined indicators" are indicators with high correlation and high effect strength; multivariate analysis algorithms include logistic regression algorithms, whose formula is:

[0054] ;

[0055] in, is the predicted probability of HR-HPV positivity, is the intercept term, is the regression coefficient, is the value of "determine the index", where Represents the k variables entering the regression model; multivariate analysis algorithms also include support vector machines and random forests, and multiple algorithms can be used in combination or in a comprehensive manner, where x1, x2, and x3 represent the variables entering the logistic regression model, which are determined by the software through model analysis, and each group of values ​​corresponding to these variables is their respective group values.

[0056] When the model building module determines the warning threshold or standard, it draws the receiver operating characteristic curve (ROC curve), calculates the area under the curve (AUC), and selects the value corresponding to the maximum Youden index as the warning threshold. The calculation formula of the Youden index is: Youden index = sensitivity + specificity - 1, where sensitivity is the true positive rate and specificity is the true negative rate.

[0057] Case Handling Summary

[0058]

[0059] HPV detection status description table

[0060] a. The actual state of positive is 1, positive means HPV positive, and negative means negative. N is the sample size, that is, the number of positive and negative cases. The effective N is the actual sample size involved in the statistics. The larger the value of the test result variable, the more it can prove that the actual state is positive.

[0061] Area under the curve

[0062] Test outcome variables: predicted probability (as shown in the table below)

[0063]

[0064] Test outcome variable: The predicted probability has at least one tie between the positive and negative actual status groups; the statistic may be biased;

[0065] a. Under non-parametric assumptions;

[0066] b. Null hypothesis: real area = 0.5.

[0067] It also includes model validation, evaluation and optimization modules, which are used to evaluate the accuracy, stability and extrapolation reliability of the model through internal and external validation, and to continuously adjust and optimize; internal validation is the validation of modeling data, and is carried out simultaneously with modeling, using the cross-validation method to divide the modeling data into Equal parts, each time 1 copy of data modeling, 1 copy of data verification, repeat The average value is taken after times; external verification includes past data verification and newly collected physical examination data verification, and the newly collected physical examination data verification is further divided into known test result data verification and unknown result data verification.

[0068] When the model validation, evaluation and optimization module evaluates the accuracy of the model, it uses indicators such as accuracy, precision, and recall. The accuracy is calculated as follows: Accuracy = (True Positives + True Negatives) / (True Positives + False Positives + True Negatives + False Negatives); the precision is calculated as follows: Precision = True Positives / (True Positives + False Positives); and the recall is calculated as follows: Recall = True Positives / (True Positives + False Negatives).

[0069] It also includes a model application and long-term observation module. The model application and long-term observation module is used to apply the optimized model to people undergoing physical examinations, etc. Specifically, the routine blood and urine biological indicator data of the physical examination population are obtained through the system interface, and the HR-HPV positive warning results are obtained after input into the model, and the results are fed back to relevant medical personnel; at the same time, long-term observations are conducted on the application population to record the degree of consistency between the actual HR-HPV infection situation and the warning results.

[0070] The model application and long-term observation module is also used to regularly feed back data to the model verification, evaluation and optimization module based on long-term observation results to further optimize the model. The optimization content includes adjusting the selection of "determination indicators" and updating the parameters of the multivariate analysis algorithm.

[0071] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. A precise early warning auxiliary diagnosis model system based on high-risk HPV positivity, characterized by: include: An indicator correlation analysis module is used to study the quantitative relationship between routine blood and urine biological indicators and HR-HPV positivity; The indicator association analysis module uses a correlation analysis algorithm when studying quantitative relationships; The model building module is used to analyze and determine useful indicator factors, establish a "determined indicator" and a positive early warning model for HR-HPV based on the corresponding multivariate analysis algorithm, and determine the early warning threshold or standard.

2. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 1, characterized in that: The indicator correlation analysis module includes blood routine, urine biology, liver function and liver enzymes, kidney function, blood lipids and blood sugar and other basic indicators: The blood test, urine biomarkers, liver function and liver enzymes, kidney function, blood lipids and blood sugar and other basic indicators include: There are 24 blood routine indicators, including white blood cell count, neutrophil percentage, lymphocyte percentage, monocyte percentage, eosinophil percentage, basophil percentage, absolute neutrophil count, absolute lymphocyte count, absolute monocyte count, absolute eosinophil count, absolute basophil count, red blood cell count, hemoglobin concentration, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red blood cell distribution width, platelet count, hematocrit, mean platelet volume and platelet volume distribution width, as well as neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio; There are 14 urine routine indicators, including urine occult blood, urine bilirubin, urine urobilinogen, urine ketones, urine protein, urine nitrite, urine glucose, urine pH, urine specific gravity, urine routine vitamin C, leukocyte esterase, urine microscopic examination of white blood cells, urine microscopic examination of red blood cells, and urine casts; Liver function and 7 liver enzymes, specifically alanine aminotransferase, total bilirubin, direct bilirubin, indirect bilirubin, aspartate aminotransferase, γ-glutamyl transferase, and alkaline phosphatase; There are three types of renal function indicators, specifically creatinine, urea nitrogen, and uric acid; There are five blood lipid and blood sugar indicators, namely total cholesterol, triglycerides, fasting blood sugar, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol; Other basic indicators: age, height, weight, body mass index, systolic blood pressure, diastolic blood pressure, pulse; The quantitative relationship includes correlation or distribution difference and effect intensity, and the HR-HPV includes 17 genotypes, specifically HPV16, HPV18, HPV31, HPV33, HPV35, HPV39, HPV45, HPV51, HPV52, HPV53, HPV56, HPV58, HPV59, HPV66, HPV68, HPV73, and HPV82.

3. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 1, characterized in that: The correlation analysis algorithm includes the Pearson correlation coefficient algorithm, and its formula is: ; in, is the Pearson correlation coefficient, 、 The two variables are observations, 、 are the means of the two variables, is the number of observations.

4. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 1, characterized in that: The “determined index” is an index with high correlation and high effect strength; the multivariate analysis algorithm includes a logistic regression algorithm, and its formula is: ； in, is the predicted probability of HR-HPV positivity, is the intercept term, is the regression coefficient, is the value of the "determined indicator"; the multivariate analysis algorithm also includes a support vector machine and a random forest, and multiple algorithms can be used in combination or in a comprehensive manner.

5. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 1, characterized in that: When the model building module determines the warning threshold or standard, it draws the receiver operating characteristic curve, calculates the area under the curve, and selects the value corresponding to the maximum Youden index as the warning threshold. The calculation formula of the Youden index is: Youden index = sensitivity + specificity - 1, where sensitivity is the true positive rate and specificity is the true negative rate.

6. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 5, characterized in that: It also includes a model verification, evaluation and optimization module, which is used to evaluate the accuracy, stability and extrapolation reliability of the model through internal and external verification, and continuously adjust and optimize; the internal verification is the verification of modeling data, and is carried out at the same time as modeling, using the cross-validation method to divide the modeling data into Equal parts, each time 1 copy of data modeling, 1 copy of data verification, repeat The external verification includes verification of past data and verification of newly collected physical examination data, and the verification of newly collected physical examination data is further divided into verification of known test result data and verification of unknown result data.

7. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 6, characterized in that: When evaluating the accuracy of the model, the model verification, evaluation and optimization module uses indicators such as accuracy, precision, and recall. The accuracy is calculated as follows: accuracy = (true positive + true negative) / (true positive + false positive + true negative + false negative); the precision is calculated as follows: precision = true positive / (true positive + false positive); and the recall is calculated as follows: recall = true positive / (true positive + false negative).

8. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 6, characterized in that: It also includes a model application and long-term observation module, which is used to apply the optimized model to people undergoing physical examinations, etc. Specifically, the routine blood and urine biological indicator data of the physical examination population are obtained through the system interface, and after inputting the data into the model, the HR-HPV positive warning results are obtained and fed back to the relevant medical personnel; at the same time, long-term observation is carried out on the application population to record the degree of consistency between the actual HR-HPV infection situation and the warning results.

9. The accurate early warning auxiliary diagnosis model system based on high-risk HPV positivity according to claim 8, characterized in that: The model application and long-term observation module is also used to regularly feed back data to the model verification, evaluation and optimization module based on long-term observation results to further optimize the model. The optimization content includes adjusting the selection of "determined indicators" and updating the parameters of the multivariate analysis algorithm.

Images