A method for evaluating the risk of venous thromboembolism in a Chinese population based on a VTE-PRS-15 model

Abstract

The present application relates to the field of medical diagnosis, and specifically provides a risk assessment model for venous thromboembolism (VTE) of Chinese population and application thereof. The model is based on the genetic characteristics of Chinese population, and contains 15 genetic variables (SNPs) related to the pathophysiological mechanism of VTE, covering five aspects of anticoagulation system, coagulation system, fibrinolysis system, platelet system and vascular endothelial system. The model assesses the genetic risk of VTE of patients by calculating VTE-PRS-15 value, and determines the comprehensive risk grade of patients in combination with Caprini clinical assessment scale. The present application also provides a corresponding genetic polymorphism detection kit. According to clinical data verification, the model improves the accuracy and practicability of VTE risk assessment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical diagnosis, and in particular, the present application relates to a risk assessment model for venous thromboembolism in Chinese population and application thereof. BACKGROUND

[0002] Venous thromboembolism (VTE) refers to the abnormal coagulation of blood in the vein, which causes complete or incomplete obstruction of blood vessels, and is a venous return disorder disease, including deep vein thrombosis (DVT) and pulmonary thromboembolism (PTE). VTE presents the characteristics of occult onset, atypical clinical symptoms, easy misdiagnosis, high mortality and disability rate, etc., and is the third leading cause of death in the circulatory system after myocardial infarction and stroke. DVT can occur in veins of various parts of the body, and lower extremity veins are the most common. Acute thrombus detachment of DVT can be complicated with pulmonary embolism (PE), and the mortality rate is as high as 30%. About 80% of DVT patients have no obvious symptoms, and the mortality and disability rate is high, which is called "silent killer".

[0003] From the etiological point of view, VTE can be divided into three categories: familial genetic VTE, VTE with complex disease characteristics, and secondary VTE. Among them, VTE with complex disease characteristics is the most common VTE in clinical practice, which is determined by multiple genetic factors and acquired factors or states. Risk assessment is the premise and basis of VTE standard prevention, which is to quantify the possibility of VTE occurrence in the process of VTE occurrence. At present, the commonly used VTE risk assessment tools recommended by the Guidelines for Prevention and Treatment of Venous Thromboembolism in Hospitals (2022), Chinese Guidelines for Prevention and Treatment of Thrombosis (2018), Chinese Guidelines for Prevention of Venous Thromboembolism in Major Orthopedic Surgery (2016), Chinese Guidelines for Prevention of Venous Thromboembolism in Trauma Orthopedic Surgery Patients (2021) and other guidelines are the Caprini scale introduced from abroad, which contains 37 risk factor items, involving patient's basic information, disease history, surgical history, medication history and genetic factors, etc.

[0004] Caprini scale is a comprehensive assessment scale of genetic risk factors and clinical risk factors, which has been widely used in global VTE risk assessment. Many studies and clinical practices have confirmed its value in VTE risk assessment and prevention decision guidance. However, the genetic factors F5 Leiden mutation and F2 G20210A mutation in the scale are more common in European and American populations, and the mutation rate in Chinese population is almost zero. Therefore, when Caprini scale is applied in China, the genetic risk factors can only be calculated as zero, and only clinical risk factors are included. The genetic factors that contribute 40-60% to VTE occurrence are missing. Because it cannot provide complete and race-appropriate prediction results, the accuracy of clinical VTE risk assessment is affected, and the standard assessment of the incidence risk of all potential VTE patients in the clinic has not yet reached the expected level.

[0005] Studies have confirmed that there is no dominant gene or major locus similar to F5 Leiden mutation and F2 G20210A mutation in Chinese population. The most common in China is the VTE with multiple gene representation and complex disease characteristics, which is controlled by multiple genes and multiple sites in terms of genetics, and affects different pathophysiological mechanisms of VTE occurrence from different aspects. The effect of a single or a few gene sites is weak, and it is difficult to accurately assess the genetic risk of VTE. Therefore, it is necessary to use the internationally accepted polygenic risk score (PRS) modeling technology to construct the VTE-PRS model of Chinese genetic characteristics, quantify the cumulative effect of multiple gene polymorphisms, calculate the multiple gene variation information into the score of measuring the susceptibility of individuals to VTE, assess the genetic risk of VTE occurrence, and conduct clinical comprehensive risk assessment combined with Caprini risk assessment results.

[0006] Although there are many relevant research results of PRS technology in China at present, three key technical problems in clinical application have not been effectively solved. One is that most models need to detect dozens or even hundreds of SNPs, which has problems of high cost, high cost of clinical verification and registration, and great difficulty in clinical popularization. On the one hand, the SNP needs to be scientifically simplified, and on the other hand, the pathophysiological mechanism of VTE should be clearly explained from the molecular level to ensure that the clinical application process has both economic practicality and scientific risk assessment. Two is that most existing studies are based on simulated data to construct training set and validation set for modeling research, and lack of verification of the effectiveness and prediction performance stability of the model based on real clinical case data, and solving the key technical index of risk stratification threshold based on real clinical data. Three is that clinical common VTE is a complex disease caused by genetic factors and non-genetic factors (such as environmental factors, acquired factors), so the PRS model must be combined with Caprini scale for comprehensive risk assessment in clinical practical application, and the method and risk determination standard for solving comprehensive risk assessment are still a difficult problem. SUMMARY

[0007] In order to solve the limitations existing in the prior art, the present application develops a prediction model suitable for risk assessment of venous thromboembolism in Chinese population, and designs a genetic polymorphism detection kit and a method of comprehensive risk assessment combined with Caprini based on the model. The prediction performance and clinical value of the model are verified by collecting 2038 patients real clinical data by prospective case-control study, which is simple in clinical operation and economical and practical, and solves the problem of application landing of multi-gene risk score (PRS) in clinical application of global precision medicine for the first time. Specifically, the present application provides the following technical solutions:

[0008] In a first aspect, the present application provides a method for evaluating the risk of venous thromboembolism (VTE) in Chinese population based on a multi-gene risk score (VTE-PRS-15), comprising the following steps: (1) genotyping 15 SNPs of SERPINC1 rs2227589, PROC rs146922325, PROC rs199469469, PROS1 rs6795524, THBD rs16984852, APOH rs52797880, F11 rs2289252, F11 rs2036914, F2 rs3136516, FGG rs2066865, PAI-1 rs1799762, ABO rs8176719, MTHFR rs1801133, NOS3 rs1799983 and HIVEP1 rs169713; (2) calculating the VTE-PRS-15 value by the VTE-PRS-15 model formula; (3) determining the genetic risk of VTE of the patient by using the VTE-PRS-15 risk threshold (Cut-off value) calculated based on the K-means clustering algorithm; and (4) evaluating and determining the comprehensive risk of VTE of the patient in combination with the clinical risk (Caprini score result).

[0009] In an embodiment, the genetic polymorphism detection employs PCR-fluorescent probe technology.

[0010] In an embodiment, the calculation formula of the VTE-PRS-15 value is as follows: wherein i = 1 to 15, βi represents the effect value (i.e. Ln of odds ratio [OR]) of the i th genetic locus, and Gi represents the copy number (0, 1 or 2) of the risk allele.

[0011] In an embodiment, the K-means clustering algorithm is used to determine the risk threshold (Cut-off value) of the VTE-PRS-15 value: (1) the Cut-off value of the VTE-PRS-15 using binary classification result is 2.335. When VTE-PRS-15 ≥ 2.335, it is a genetic high risk, and when VTE-PRS-15 < 2.335, it is a genetic low risk; (2) the Cut-off values of the VTE-PRS-15 using three classification results are 2.230 and 1.730, respectively. When VTE-PRS-15 ≥ 2.230, it is a genetic high risk, when 1.730 < VTE-PRS-15 < 2.230, it is a genetic medium risk, and VTE-PRS-15 ≤ 1.730, it is a genetic low risk.

[0012] In a second aspect of the present application, a comprehensive evaluation method for evaluating the VTE risk assessment of Chinese population is provided, which comprises the following steps: (1) detecting the genetic polymorphism of the patient according to the above method, and calculating the VTE-PRS-15 value; (2) simultaneously performing clinical risk assessment by using Caprini risk assessment scale; and (3) comprehensively determining the patient as high risk, medium risk or low risk based on the VTE-PRS-15 value and Caprini score.

[0013] In an embodiment, when the Caprini scale assessment is high risk or medium risk, and the VTE-PRS-15 value is greater than or equal to 2.335, it is determined as high risk; when the Caprini scale assessment is medium risk or low risk, and the VTE-PRS-15 value is less than 2.335, it is determined as low risk, and the rest is determined as medium risk.

[0014] In an embodiment, when the VTE-PRS-15 value is greater than or equal to 2.230, it is determined as high risk; when the Caprini scale assessment is high risk and 1.730 < VTE-PRS-15 value < 2.230, it is determined as medium risk; and the rest is determined as low risk.

[0015] In a third aspect of the present application, a genetic polymorphism detection kit for VTE risk assessment of Chinese population is provided, which comprises specific primers and fluorescent probes for the 15 SNPs.

[0016] In an embodiment, the specific primers and probes comprise the sequences shown in SEQ ID NO. 1 to SEQ ID NO. 63.

[0017] In a fourth aspect of the present application, the use of the above method in the preparation of a diagnostic reagent or product for evaluating the genetic risk of venous thromboembolism of an individual is provided.

[0018] In a fifth aspect, the present application provides a method for constructing a risk assessment model for venous thromboembolism (VTE) in Chinese population, which comprises the following steps: (1) collecting clinical data of target population, and constructing a case group and a control group; (2) performing genetic polymorphism detection on the case group and the control group respectively, and detecting 15 genetic loci of SERPINC1 rs2227589, PROC rs146922325, PROC rs199469469, PROS1 rs6795524, THBD rs16984852, APOH rs52797880, F11 rs2289252, F11 rs2036914, F2 rs3136516, FGG rs2066865, PAI-1 rs1799762, ABO rs8176719, MTHFR rs1801133, NOS3 rs1799983 and HIVEP1 rs169713; (3) calculating VTE-PRS-15 value of each sample based on the detection results; and (4) analyzing the calculated VTE-PRS-15 value by using statistical methods, and determining a threshold value for risk assessment.

[0019] In an embodiment, the genetic polymorphism detection refers to using PCR fluorescent probe technology.

[0020] In an embodiment, the calculation formula of the VTE-PRS-15 value is as follows: wherein i = 1 to 15, βi represents an effect value (i.e. Ln of odds ratio [OR]) of the i th genetic locus, and G i represents the copy number (0, 1 or 2) of the risk allele.

[0021] In an embodiment, K-means clustering analysis is used to determine the risk assessment threshold value of the VTE-PRS-15 value, and to divide the genetic risk into three levels of high genetic risk, medium genetic risk and low genetic risk.

[0022] Compared with the prior art, the present application has the following beneficial technical effects:

[0023] (1) The present application is based on the principles of evidence-based medicine, statistics and genomics, and 15 SNPs of five systems (anticoagulation system, coagulation system, fibrinolysis system, platelet system and vascular endothelial system) covering the pathophysiological mechanism of VTE are scientifically screened, 2038 cases of case-control data are prospectively collected, it is verified that the VTE-PRS-15 value is an independent risk factor for VTE, and the effectiveness of the VTE-PRS-15 model for genetic risk stratification of VTE is verified, which fills the gap of Chinese population characteristics genetic factors in VTE clinical risk assessment, and improves the interpretability and preventive guidance of the detection results.

[0024] (2) The application mines the best risk threshold of the VTE-PRS-15 model and the VTE-PRS-15B+Caprini and VTE-PRS-15T+Caprini comprehensive risk judgment standards based on real clinical data, and verifies the effectiveness, safety and reliability of the VTE-PRS-15B+Caprini and VTE-PRS-15T+Caprini through real clinical data.

[0025] (3) The application overcomes the core pain points in the prior art such as gene site redundancy, lack of data verification, disconnection between clinical application and promotion, and high promotion cost, and provides a safe, effective and innovative tool for precise prevention and control of VTE. At the same time, the application considers the clinical practice, and under the premise of maintaining the use of the Caprini scale, the risk assessment accuracy and safety of the original Caprini scale are greatly improved by adding the genetic evaluation result of the VTE-PRS-15 model, and for the first time, the genetic risk factor (VTE-PRS-15) is systematically integrated and judged with the commonly used risk prediction model (Caprini), providing a more comprehensive, more convenient and more reliable VTE risk assessment tool for the clinic. When applied in the clinic, it reduces the economic burden of patients, improves the standard prevention rate of VTE, and reduces the incidence of VTE. BRIEF DESCRIPTION OF DRAWINGS

[0026] The accompanying drawings are included to provide a further understanding of the application, and constitute a part of the specification, illustrate the application together with the embodiments of the application, and are used to explain the application, and do not constitute a limitation on the application. In the drawings:

[0027] Figure 1 is a VTE-PRS-15 value distribution graph of the case group and the control group;

[0028] Figure 2 is a ROC curve of the VTE-PRS-15;

[0029] Figure 3 is a ROC curve of the VTE-PRS-15 model 10 times resampling;

[0030] Figure 4 is a ROC curve of the Caprini scale in the target data. DETAILED DESCRIPTION

[0031] The preferred embodiments of the application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the application, and do not limit the application.

[0032] The target data for this example was constructed using a prospective case-control study (National Medical Research Registry Information System Project Registration Number: MR-61-23-019694) to collect clinical data from 18 multi-center hospitals. The target data for the model study was established.

[0033] Example 1: Prospective Case-Control Study

[0034] Research Method: Prospective case-control study is a clinical research method that combines the characteristics of prospective cohort study and traditional case-control study in epidemiological research. The core is to determine the study groups (case group and control group) first, and then prospectively collect or observe the data related to the outcome to explore the association between exposure factors and disease (or other outcomes).

[0035] Sample size calculation: Based on the number of model VTE-PRS-15 variables (SNPs) 15, the sample size was calculated. There are 3 genotypes for each SNP, according to the "20 times of the number of variables as the sample size meets the most stringent empirical rule", the sample size calculation formula for case group and control group: 15 (SNP) x 3 (genotype) x 20 = 900 cases. Considering that the sample collection failure rate is usually 10-30% and the complexity of VTE clinical evaluation scale Caprini, according to the sample collection failure rate of 20%, the calculation is: 900 + 900 x 20% = 1080 cases. Therefore, 2200 cases of multi-center clinical samples were collected, including 1100 cases of case group and 1100 cases of control group, to ensure that the sample size for statistical analysis is not less than 1800 cases, including not less than 900 cases of case group and not less than 900 cases of control group.

[0036] Outcome indicator: Deep vein thrombosis (DVT).

[0037] DVT diagnosis method and standard: In this study, color Doppler ultrasound examination of both lower extremity veins was used as the diagnostic standard for lower extremity DVT. For DVT with high clinical probability, negative or uncertain results of color Doppler ultrasound of both lower extremities, magnetic resonance angiography or angiography can be further performed according to the relevant guidelines.

[0038] Observation period and observation indicators: From admission to 3 months after surgical treatment. The observation indicators include Caprini score table related indicators, genetic information (VTE-PRS-15 value), laboratory indicators, imaging examination results, etc.

[0039] Data grouping: Patients diagnosed with DVT during the observation period were included in the case group; patients without DVT during the observation period were included in the control group.

[0040] Inclusion criteria: age 18-80 years; orthopedic inpatients, hospitalization time ≥3d; surgical patients or perioperative bone trauma patients; fracture patients are fresh fracture patients; all have completed double lower limb venous color Doppler ultrasound examination; no contraindication to anticoagulation; agree to participate in the study and sign the informed consent form (provided by the patient himself or his legal guardian). Exclusion criteria: patients with pathological fractures; patients with cirrhosis, liver cancer or hepatectomy; patients with long-term bed rest, hemiplegia and other limited activities before admission; patients in the terminal stage of various diseases with expected survival time <1 year; combined with renal dysfunction or liver dysfunction; combined with hematological diseases or coagulation dysfunction; previously suffered from thrombotic diseases and are receiving treatment; patients with poor compliance, or known mental illness and intellectual disability; patients who do not receive regular anticoagulant prophylaxis.

[0041] Sample collection process: For orthopedic (VTE high-risk high-incidence department) inpatients, patients who meet the inclusion and exclusion criteria sign the informed consent form. The inclusion case group occurs DVT during hospitalization, and whole blood samples are collected for VTE-PRS-15 value detection, and the clinical observation form information is filled in. The whole blood sample is collected for VTE-PRS-15 value detection before discharge for patients who do not occur DVT during hospitalization, and the clinical observation form information is filled in, and followed up for three months after discharge. The follow-up group occurs DVT during follow-up, and the control group does not occur DVT.

[0042] This study collected 2620 clinical samples from February 2023 to March 2025. Eliminate cases that do not meet the inclusion and exclusion criteria during follow-up, such as follow-up dropouts, missing B-ultrasound diagnosis results, and hospitalization time less than three days, etc. A total of 2038 cases of target data that meet the inclusion and exclusion criteria for statistical analysis were included, including 1013 cases in the case group and 1025 cases in the control group, meeting the sample size of 900 cases in the case group and control group required for research analysis.

[0043] The target data is divided into case group data and control group data, which is composed of six parts of patient general information data, Caprini score table index data, laboratory index examination data, imaging data, genetic testing data, and DVT prevention data.

[0044] Based on our previous research results "A set of biomarkers, kits and applications for genetic risk prediction of venous thromboembolism in Chinese Han population (ZL2022101197350)" and the academic paper published in the international renowned vascular surgery specialty journal JVasc Surg Venous Lymphat Disord. 2024 Jan;12(1):101666 "Construction and optimization of a polygenic risk model for venous thromboembolism in the Chinese population", in which according to the screening principles and classification standards of disease-related gene variation sites in the Chinese population, genetic variables conforming to the characteristics of the Chinese population were screened, and the stepwise forward selection method was used to optimize the model variables (detection sites). Starting from the SNP with the highest OR value (PROC, rs146922325) among the 53 SNPs, SNPs were added to the PRS in order from high to low, and the AUC of PRS was calculated after each SNP was added. With the addition of SNPs, the AUC of PRS rapidly increased, and until the 10th SNP was included in PRS, the AUC no longer significantly improved, and finally 10 SNPs related to the risk of VTE occurrence and verified by repeated verification were determined (PROC gene rs146922325 (C>T), PROC gene rs199469469 (AAG->), THBD gene rs16984852 (C>A), ABO gene rs8176719 (->C), FGG gene rs2066865 (G>A), PAI-1 gene rs1799762 (4G>5G), APOH gene rs8178847 (A>G), F11 gene rs2289252 (C>T), F11 gene rs2036914 (T>C), MTHFR gene rs1801133 (G>A)).

[0045] Then we analyzed the molecular mechanism of disease occurrence from the system level, combined with the influence of genetic factors on the pathophysiological mechanism of VTE, and optimized the model from the perspective of "genomics". From the previous research results of the patent and published papers, 5 important SNPs related to the pathophysiology of VTE were supplemented, which are SERPINC1 rs2227589 (C>T), PROS1 rs6795524 (A>G), F2 rs3136516 (G>A), HIVEP1 rs169713 (T>C) and NOS3 rs1799983 (T>G).

[0046] At the same time, considering the problem of primer and probe synthesis in the development of detection reagents, the APOH gene rs8178847(A>G) is replaced by the completely linked APOH gene rs52797880(A>G) with the same function.

[0047] Based on the above 15 sites, we constructed the VTE-PRS-15 model, the formula is as follows: Where: i represents the ith SNP (i = 1, 2,..., 15); βi represents the effect value of the ith SNP (i.e. Ln of odds ratio [OR]); Gi represents the risk allele copy number of the ith SNP, represented by (0, 1 or 2) respectively.

[0048] The 15 VTE-related SNPs in the VTE-PRS-15 model comprehensively cover the five major pathophysiological mechanisms of VTE occurrence, not only maintaining the characteristics of optimization and simplicity, but also improving the coverage of the model detection target for VTE genetic risk factors and clinical applicability, which can provide the basis for explaining the causes of disease and precise prevention from the molecular level. The specific information of the 15 SNPs and their pathophysiological mechanism classification are shown in Table 1.

[0049] Table 1 Information of 15 SNPs in VTE-PRS-15 and their pathophysiological mechanism classification

[0050]

[0051] Based on the VTE-PRS-15 model, 15 VTE-related SNPs were determined, and 15 sets of specific primers and probes were designed using PCR-fluorescent probe technology (Table 2):

[0052] Table 2 Primer and probe sequences of 15 VTE-related SNPs

[0053]

[0054]

[0055] Example 2 Determination of VTE-PRS-15 model threshold

[0056] The threshold (cut-off value) of VTE-PRS-15 model in risk assessment is the key conversion node from "continuous" to "discrete" and "quantitative" to "qualitative", and its core value is to simplify complex data into a decision-making classification result, which is a key technical index for clinical application of the model. In this embodiment, K-means clustering algorithm was used to calculate and optimize the best threshold of VTE-PRS-15 model risk stratification based on the true results of the target data.

[0057] 2.1 K-means clustering algorithm:

[0058] S1 Determine ideal points: a. 100% DVT identification rate in high risk (1013 cases); b. 0 non-DVT identification rate in high risk; c. 100% non-DVT identification rate in low risk (1025 cases); d. 0 DVT identification rate in low risk.

[0059] S2 Data loading and integration: The program reads the file line by line, extracts the values in the first four columns of each line, which are the number of DVT patients in high risk, the number of non-DVT patients in high risk, the number of non-DVT patients in low risk, and the number of DVT patients in low risk, respectively, forming a four-dimensional data point. In order to find the original source of the data at the end, the program creates a custom ClusterPoint object to store four-dimensional numerical data (point) and the original row number (originalRowNum) where it is located. All these ClusterPoint objects are added to a total list to form a complete and analyzable data set.

[0060] S3 Data standardization: The program first calculates the mean (mean) and standard deviation (standard deviation) of each dimension (i.e. each column) in the data set. Then the program iterates through each data point and applies the standardization formula to each value in it: New value = (original value - mean) / standard deviation. P' = [(x1-μ1) / σ1, (x2-μ2) / σ2, (x3-μ3) / σ3, (x4-μ4) / σ4], where μ is the mean of each column and σ is the standard deviation of each column.

[0061] S4 K-means clustering:

[0062] The algorithm uses a common heuristic rule k = sqrt(n / 2) (where n is the total number of data points) to automatically determine the number of clusters k to be formed.

[0063] The program uses an optimized K-means initialization method KMeansPlusPlusClusterer to intelligently select the initial cluster centers (centroids) so that they are as dispersed from each other as possible.

[0064] The K-means algorithm starts iteration, assigns each data point to the nearest cluster center, and recalculates the center of each cluster (i.e. takes the average of all data points in the cluster as the new cluster center). By repeatedly repeating the above two steps, until the cluster assignment no longer changes, all data points are divided into k different clusters, each cluster is composed of a group of data points that are close to each other in the feature space.

[0065] S5 Optimal data point precise positioning: a. The algorithm first calculates the distance between the center point of each cluster and the preset "ideal point", finds the cluster whose center point is closest to the "ideal point", and this cluster is called "optimal cluster"; b. The algorithm directly searches within the "optimal cluster", traverses all ClusterPoint objects in this cluster, calculates the distance between each point and the "ideal point", and finally determines the point with the smallest distance as the "optimal data point".

[0066] S6 Result output: The program outputs the detailed information of the "optimal data point" found. From the metadata saved in the ClusterPoint object, the program not only displays the four-dimensional numerical value of this optimal point, but also accurately provides the row number (originalRowNum) where this data point was originally found, so that the analysis results have traceability and practical application value.

[0067] 2.2 VTE-PRS-15 model threshold and risk stratification based on binary classification results

[0068] The VTE-PRS-15 value range calculated according to the target data is 0.680-5.780, and a threshold (Cut-off value) is set in this range. If the value is greater than or equal to the Cut-off value, it is considered to be a genetic high-risk, and if the value is less than the Cut-off value, it is considered to be a genetic low-risk.

[0069] According to the "Guidelines for the Quality Evaluation and Management of Prevention and Treatment of Venous Thromboembolism in Hospitals (2022 Edition)", the Caprini score results are divided into three risk levels: high risk, medium risk and low risk. Caprini score ≥ 5 points is high risk, Caprini score 3-4 points is medium risk, and Caprini score 1-2 points is low risk.

[0070] The three-class risk assessment results of the Caprini scale and the two-class risk assessment results of VTE-PRS-15 are used for comprehensive risk assessment. We name this comprehensive assessment model "VTE-PRS-15B+Caprini", where B represents binary classification (Binary Classification). The assessment results are defined as three risk levels: high risk, medium risk, and low risk. According to the risk assessment criteria listed in Table 3, four groups of results are obtained.

[0071] Table 3 Comprehensive risk assessment and results of each group of VTE-PRS-15B+Caprini

[0072]

[0073] The VTE-PRS-15 value calculated based on the target data is in the data interval from 0.680 to 5.780, with a step interval of 0.005. The Cut-off value corresponding to each incremental step is calculated, and then the DVT and non-DVT patient cases in the high, medium, and low risk categories are counted according to the four sets of comprehensive risk assessment results obtained.

[0074] The VTE-PRS-15B+Caprini comprehensive risk assessment automation data processing program is written, and the program traverses a total of 4085 groups of data, with each group of results corresponding to a Cut-off value. Four criteria are set: a. The number of DVT patients predicted as high risk should be higher than the Caprini scale prediction result; b. The number of non-DVT patients predicted as low risk should be higher than the Caprini scale prediction result; c. The number of non-DVT patients predicted as high risk should be less than the Caprini scale prediction result; d. The number of DVT patients predicted as low risk should be less than the Caprini scale prediction result. The computer program calculation results show that there are 67 groups of results that meet the above four criteria, and all are the second group of evaluation results in Table 3.

[0075] For the 67 groups of results, the K-means clustering method is used to calculate the optimal data point, which corresponds to a VTE-PRS-15 binary classification Cut-off value of 2.335. When VTE-PRS-15 is greater than or equal to 2.335, it is considered to be a genetic high risk, and when VTE-PRS-15 is less than 2.335, it is considered to be a genetic low risk.

[0076] 2.3 VTE-PRS-15 model threshold and risk stratification based on three classification results

[0077] Within the VTE-PRS-15 value interval of 0.680-5.780 calculated based on the target data, two threshold values are set to divide the VTE-PRS-15 risk assessment results into three types: genetic high risk, genetic medium risk, and genetic low risk.

[0078] According to the "Guidelines for Quality Evaluation and Management of Venous Thromboembolism Prevention in Hospitals (2022 Edition)", the Caprini score results are divided into three risk levels: high risk, medium risk, and low risk. Caprini score ≥ 5 points is considered high risk, Caprini score 3-4 points is considered medium risk, and Caprini score 1-2 points is considered low risk.

[0079] The Caprini scale three-class risk assessment results and the VTE-PRS-15 three-class risk assessment results were combined for comprehensive risk assessment, and this comprehensive assessment mode was named "VTE-PRS-15T+Caprini", where T represents tertiary classification, and the assessment results are defined as three risk levels of high risk, medium risk, and low risk, and 432 groups of results are obtained according to the risk assessment criteria listed in Table 4.

[0080] Table 4 Comprehensive risk assessment criteria and results of each group of VTE-PRS-15T+Caprini

[0081]

[0082] Based on the target data, the VTE-PRS-15 value was calculated from the data interval of 0.680 to 5.780 with a step size of 0.01, and the Cut-off value corresponding to each incremental step was calculated. Then, according to the 432 groups of comprehensive risk assessment results, the number of DVT and non-DVT patients in the high, medium, and low risk categories was counted.

[0083] An automatic data processing program for VTE-PRS-15T+Caprini comprehensive risk assessment was written, and the program traversed a total of 2,246,400 groups of data, with each group of results corresponding to a Cut-off value. Then, the same four criteria as the above calculation were set: a. The number of DVT patients predicted as high risk must be higher than the Caprini scale prediction result; b. The number of non-DVT patients predicted as low risk must be higher than the Caprini scale prediction result; c. The number of non-DVT patients predicted as high risk must be less than the Caprini scale prediction result; d. The number of DVT patients predicted as low risk must be less than the Caprini scale prediction result. The computer program calculation results showed that there were 14529 groups of results that met the above four criteria.

[0084] For the 14529 groups of results, the K-means clustering method was used to calculate the optimal data point, and the VTE-PRS-15 three-class Cut-off values corresponding to this data point were 2.230 and 1.730, respectively. When VTE-PRS-15≥2.230, it is genetically high risk, when 1.730

[0085] Example 3 Confirmation of the determination criteria for the comprehensive risk assessment of the VTE-PRS-15 model combined with the Caprini scale

[0086] Based on the real clinical data collected from the prospective case-control study, the K-means clustering algorithm was used to calculate the optimal threshold of the two classification results of VTE-PRS-15 model, and the risk judgment criteria of VTE-PRS-15B+Caprini comprehensive evaluation and VTE-PRS-15T+Caprini comprehensive evaluation were obtained as shown in Table 5 and Table 6. Among them, VTE-PRS-15B refers to the genetic risk assessment using two classification results, and VTE-PRS-15T refers to the genetic risk assessment using three classification results.

[0087] Table 5 Risk judgment criteria of VTE-PRS-15B+Caprini comprehensive evaluation

[0088]

[0089] Table 6 Risk judgment criteria of VTE-PRS-15T+Caprini comprehensive evaluation

[0090] Example 4 Verification of clinical application of VTE-PRS-15 model

[0091] 4.1 Performance and verification of genetic risk prediction of VTE-PRS-15

[0092] ①According to the detection results of the VTE-related 15 SNPs of the case group and the control group samples in the target data, the VTE-PRS-15 value of each sample was calculated.

[0093] ②The mean value of VTE-PRS-15 of the two groups of patients was calculated to be 2.91 (95% CI 2.88-2.95) with a standard deviation of 0.72, the maximum value was 5.78, and the minimum value was 0.68; the mean value of VTE-PRS-15 of the control group was 2.66 (95% CI 2.61-2.69) with a standard deviation of 0.66, the minimum value was 0.68, the maximum value was 4.53, and the median was 2.64; the mean value of VTE-PRS-15 of the case group was 3.18 (95% CI 3.13-3.22) with a standard deviation of 0.69, the minimum value was 1.29, the maximum value was 5.78, and the median was 3.08.

[0094] ③In order to intuitively show the distribution of VTE-PRS-15 scores in the two groups of people, based on the VTE-PRS-15 values in the study population from the minimum value 0.68 to the maximum value 5.78 (full range 5.10), the group interval was set to 0.2, and the number of cases and controls in different score intervals was calculated. The bar chart was drawn with VTE-PRS-15 value as abscissa and individual distribution ratio as ordinate (as shown in Figure 2). Figure 2), the results showed that the proportion of the number of cases in each interval increased with the increase of VTE-PRS-15 value.

[0095] ④Independent sample t test was used to evaluate the correlation between VTE-PRS-15 value and DVT, the results showed that the mean of VTE-PRS-15 in the case group was significantly higher than that in the control group (3.18±0.69 vs 2.66±0.67, p<0.001), the mean difference between the two groups was 0.52 (95% CI: 0.46~0.58). Levene's test of homogeneity of variance showed that there was no significant difference in variance between the two groups (F=1.120, p=0.290), supporting the t test results under the assumption of equal variance. The high t value (t=17.38) and the confidence interval not crossing zero further verified the robustness of the results. The results showed that VTE-PRS-15 can be used as an effective indicator for genetic risk stratification of DVT. The results of independent sample t test are shown in Table 7.

[0096] Table 7 Independent sample t test

[0097]

[0098] ⑤Taking VTE-PRS-15 value as a continuous independent variable and the occurrence of DVT as a binary dependent variable, the area under the receiver operating characteristic curve (AUC) was used to evaluate the predictive discrimination ability of the model. The results showed that the AUC of the model was 0.702 (95% confidence interval: 0.680-0.725). The ROC curve is as follows Figure 3 .

[0099] Conclusion: The AUC of VTE-PRS-15 is 0.702 (95% CI: 0.680-0.725, P<0.001), which indicates that it has good predictive ability for VTE risk (AUC>0.7), which is significantly better than random classification (AUC=0.5).

[0100] ⑥Bootstrap method to verify the discrimination of VTE-PRS-15

[0101] Bootstrap method in clinical prediction model validation method was used to write Bootstrap resampling software program, using model development target data (2038 cases of prospective case-control study), through resampling with replacement, a Bootstrap resampling validation set with the same sample size (1013 cases of case group and 1025 cases of control group) was constructed to verify the discrimination (AUC) of VTE-PRS-15 model, then the process was repeated 10 times, that is, 10 validation sets were constructed, the AUC of the model in each validation set was calculated, and the mean was calculated to evaluate the stability of VTE-PRS-15 model performance.

[0102] The area under the receiver operating characteristic curve (AUC) of the VTE-PRS-15 model was plotted based on 10 resampling validation sets, and the calculation results are shown in Table 8. The results showed that the AUC interval of the VTE-PRS-15 model in the 10 resampling samples was 0.675-0.728, and the AUC mean was 0.701. Compared with the AUC of the VTE-PRS-15 model calculated based on the target data, which was 0.702, the two results were highly consistent, reflecting the stability of the model in different data sets.

[0103] Table 8 VTE-PRS-15 resampling validation AUC

[0104]

[0105] The area under the receiver operating characteristic curve (AUC) of the VTE-PRS-15 model was plotted based on 10 resampling validation sets, and the calculation results are shown in Table 8. The results showed that the AUC interval of the VTE-PRS-15 model in the 10 resampling samples was 0.675-0.728, and the AUC mean was 0.701. Compared with the AUC of the VTE-PRS-15 model calculated based on the target data, which was 0.702, the two results were highly consistent, reflecting the stability of the model in different data sets. Figure 4

[0106] 4.2 Verification of the clinical value of VTE-PRS-15

[0107] First, the VTE-PRS-15 value was used as the independent variable, and whether DVT occurred was used as the dependent variable. Single factor binary Logistic regression was used to analyze the influence of VTE-PRS-15 value on DVT (see Table 9). When the VTE-PRS-15 value increased by 1 point, the risk of DVT increased by 218.2% (OR=3.182, 95% CI: 2.740-3.694, p<0.001); with a standard deviation (SD=0.68), an increase of 1 SD corresponds to an increase of 119.8% in the risk of DVT (OR=2.198, 95% CI: 1.934-2.499). Therefore, VTE-PRS-15 is an important risk factor for the occurrence of DVT.

[0108] Table 9 Single factor binary Logistic regression analysis of the influence of VTE-PRS-15 on DVT

[0109] Further single factor analysis was performed on all risk factors of the Caprini scale in the target data, and chi-square test (χ² test) was used for inter-group comparison. All tests were set at a two-sided test level α=0.05, and P<0.05 indicated that the difference was statistically significant. The statistical results are shown in Table 10.

[0110] Table 10 Single factor analysis results of Caprini scale

[0111] ​

[0112]

[0113] As shown in Table 10, the risk factors with significant statistical difference (P < 0.001) include: age, BMI > 25, severe lung disease (such as pneumonia, course < 1 month), bedridden patients, pregnancy or postpartum, large open surgery (> 45 minutes), elective joint replacement, bedridden > 72 hours, plaster fixation, hip / pelvis or lower extremity fracture.

[0114] The orthopedic clinical risk factors with statistical significance (P < 0.001) in the above Caprini scale single factor analysis, body mass index (BMI) ≥ 25 kg / m², large open surgery (> 45 minutes), elective joint replacement, hip / pelvis or lower extremity fracture, age (continuous variable), acute spinal cord injury (course < 1 month) and VTE-PRS-15 value, are included in the multi-factor logistic regression analysis, and the analysis results are shown in Table 11.

[0115] Table 11 VTE-related multi-factor logistic regression analysis

[0116]

[0117] The results show that the correlation between VTE-PRS-15 value and DVT is still statistically significant (P < 0.001), that is, it has a significant contribution to the occurrence of DVT, proving that VTE-PRS-15 value is an independent risk factor for the occurrence of DVT (from a statistical point of view, if a variable has a significant contribution to the result in a statistical model containing known risk factors, it is called an independent risk factor or independent risk factor). VTE-PRS-15 (OR = 3.330, 95% CI: 2.836-3.911), which indicates that the risk of DVT increases by 233% for every 1 point increase in VTE-PRS-15 value.

[0118] 4.3 Verification of clinical application of VTE-PRS-15

[0119] ①Caprini scale data analysis

[0120] The Caprini scale and risk assessment standard recommended by the National Pulmonary Embolism and Deep Vein Thrombosis Prevention and Control Capability Building Project Expert Committee “Guidelines for Prevention and Control of Venous Thromboembolism in Hospitals (2022 Edition)” are used in the present application. According to the target data risk assessment results, see Table 12.

[0121] Table 12 Caprini scale risk assessment results of target data

[0122]

[0123] The total number of target data collection cases was 2038, of which 1691 were assessed as high-risk by the Caprini scale, accounting for 82.97% of the total cases. Among them, 806 cases did not occur DVT, and 885 cases occurred DVT, accounting for 52.34% of the 1691 high-risk cases. Among the 60 cases assessed as low-risk by the Caprini scale, 47 cases did not occur DVT, and 13 cases occurred DVT, accounting for 78.33% of the 60 low-risk cases.

[0124] The Caprini scale was defined as positive for high-risk and negative for medium and low-risk. The sensitivity of the Caprini scale was 87.36%, and the specificity was 21.37%. The area under the curve (AUC) of the receiver operating characteristic curve (ROC) of the Caprini scale was 0.569 (95% CI: 0.544-0.592, P<0.001), indicating that the Caprini scale had low discrimination ability for VTE risk groups in the current clinical environment. The ROC curve of the Caprini scale is shown in Figure 4 .

[0125] The use of the Caprini scale in the current domestic clinical environment is limited by factors such as racial differences, missing genetic risk factors, time-consuming and complex operations, and different levels of understanding of certain items, which affect the accuracy of the assessment.

[0126] ②Clinical prediction accuracy analysis after adding genetic risk factor VTE-PRS-15

[0127] Statistical analysis was performed on the prediction results of the Caprini scale, VTE-PRS-15B+Caprini, and VTE-PRS-15T+Caprini. First, high risk or high risk was defined as positive, and medium and low risk or medium and low risk were defined as negative. The sensitivity, specificity, positive predictive value, and negative predictive value of Caprini, VTE-PRS-15B+Caprini, and VTE-PRS-15T+Caprini were calculated. The calculation formula is as follows:

[0128]

[0129] Table 13 Results of risk assessment accuracy of three models

[0130]

[0131] The results showed that:

[0132] The sensitivity of Caprini scale was 87.36%, the specificity was 21.37%, the positive predictive value was 52.34%, and the negative predictive value was 63.11%; VTE-PRS-15B+Caprini mode had higher sensitivity (90.52%) while its specificity was significantly higher than other modes and Caprini scale. This mode can greatly reduce the misdiagnosis rate of traditional Caprini scale while having good DVT risk identification ability; VTE-PRS-15T+Caprini mode has higher sensitivity (93.88%) and its specificity (27.63%) is higher than Caprini scale. This mode can also reduce the misdiagnosis rate of traditional Caprini scale to some extent while having good DVT risk identification ability.

[0133] Using VTE-PRS-15B+Caprini mode, the sensitivity of Caprini scale was improved from 87.36% to 90.52%, and the specificity was improved from 21.37% to 33.85%. That is, Caprini scale will misdiagnose 79 out of 100 people as high risk, and VTE-PRS-15B+Caprini will misdiagnose 66 out of 100 people as high risk, i.e. about 13 people per 100 are misdiagnosed less. At the same time, specificity is directly related to positive predictive value (PPV). Under the condition that the prevalence rate remains unchanged, the improvement of specificity will make the PPV increase synchronously, i.e. when the model predicts "positive", the actual prevalence rate is higher, so the credibility of VTE-PRS-15B+Caprini positive result is also improved.

[0134] Further calculate the net reclassification improvement index NRI of the three models. NRI is an index for measuring the improvement of new prediction model (or new index) compared with the old model in the accuracy of risk stratification. Taking Caprini scale as the old model and VTE-PRS-15B+Caprini and VTE-PRS-15T+Caprini as the new model, the calculation formula is as follows:

[0135] NRI=(sensitivity new +specificity new ) - (sensitivity old +specificity old )

[0136] The calculation results are shown in Table 14.

[0137] Table 14 NRI of improved prediction ability of two new models

[0138]

[0139] The results show that the NRI of the two new models is greater than zero, which means that the risk stratification ability is substantially improved compared with the traditional Caprini scale, and the results of the two clinical application modes of VTE-PRS-15 are better than those of the Caprini scale alone. Among them:

[0140] The NRI of VTE-PRS-15B+Caprini compared with the traditional Caprini scale is 15.65% (95% CI, 11.73%, 19.57%; p<0.001), i.e. the proportion of "correct reclassification" increases by 15.65% compared with the traditional Caprini scale, 3.16% of DVT patients are correctly reclassified from the medium-low risk group to the high risk group, and 12.49% of non-DVT patients are correctly reclassified from the high risk group to the medium-low risk group. This new model significantly improves the old model and has significant clinical value;

[0141] The NRI of VTE-PRS-15T+Caprini compared with the traditional Caprini scale is 11.88% (95% CI, 7.57%, 16.19%; p<0.001), i.e. the proportion of "correct reclassification" increases by 11.88% compared with the traditional Caprini scale, 6.52% of DVT patients are correctly reclassified from the medium-low risk group to the high risk group, and 5.37% of non-DVT patients are correctly reclassified from the high risk group to the medium-low risk group. This new model also significantly improves the old model and has significant clinical value.

[0142] 4.4 Verification of the clinical prediction accuracy of VTE-PRS-15 after adding genetic risk factors

[0143] Bootstrap resampling was used to construct 10 validation sets in the target data, calculate the indicators of each model in each validation set, and calculate the mean to evaluate the stability of the VTE-PRS-15B+Caprini and VTE-PRS-15T+Caprini model performance. See Tables 15, 16, 17 and 18. The validation results show that in the 10 Bootstrap resampling sample sets, the comprehensive risk assessment model with genetic factors performs stably, and is highly consistent with the results obtained based on the target data. It is verified that the addition of genetic factors VTE-PRS-15 in clinical evaluation has higher risk assessment accuracy for DVT patients and non-DVT patients than the traditional Caprini scale, and the method has good stability in clinical prediction performance.

[0144] Table 15 Resampling validation results of VTE-PRS-15B+Caprini comprehensive risk assessment

[0145]

[0146] Table 16 Validation results for VTE-PRS-15T + Caprini composite risk assessment

[0147]

[0148] Table 17 Validation results for accuracy metrics of new and old model risk assessments

[0149]

[0150] Table 18 Validation results for NRI for improved predictive ability of two new models

[0151]

[0152] It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application cover modifications and variations of this application provided they come within the scope of the appended claims and their equivalents.

Claims

1. A gene polymorphism detection kit for assessing the risk of venous thromboembolism in the Chinese population, characterized in that, It includes specific primers and fluorescent probes for 15 SNPs, which consist of SERPINC1 rs2227589, PROCrs146922325, PROC rs199469469, PROS1 rs6795524, THBD rs16984852, APOH rs52797880, F11 rs2289252, F11 rs2036914, F2 rs3136516, FGG rs2066865, PAI-1 rs1799762, ABOrs8176719, MTHFR rs1801133, NOS3 rs1799983, and HIVEP1 rs169713.

2. The reagent kit according to claim 1, characterized in that, The specific primers and probes include the sequences shown in SEQ ID NO.1 to SEQ ID NO.

63.

3. The use of the kit according to claim 1 or 2 in the preparation of an assessment reagent or product for evaluating the genetic risk of venous thromboembolism in an individual in the Chinese population.

Images