Method for screening urothelial carcinoma by using methylated biomarker

Abstract

The invention discloses application of at least one reagent in preparation of a kit for detecting urothelial carcinoma, the at least one reagent comprises a primer or a probe for detecting methylation of at least one target gene, and the at least one target gene comprises TMEM106A, CA3, ZIC4 or a combination thereof; the kit carries out detection by a method comprising the following steps: (a) providing an in vitro sample taken from an individual; (b) determining the gene methylation degree of the at least one target gene in the in vitro sample; and (c) comparing the methylation degree of the at least one target gene in the step (b) with the methylation degree of a control group gene, and when the methylation degree of the at least one target gene is higher than that of the control group gene, judging that the individual suffers from the urothelial carcinoma at high risk.

Description

Technical Field

[0001] The present invention relates to a method for detecting urothelial carcinoma by combining artificial intelligence and epigenetic genomics. Background Art

[0002] DNA methylation modification

[0003] Epigenetics, also known as "expression genetics", "epigenetics", "exogenous genetics", "epigenetics", "exogenetic genetics", "exogenetic genetics", etc., refers to the heritable changes in gene function without changes in the DNA sequence of the gene, which ultimately lead to phenotypic changes. The most common epigenetic modification is DNA methylation modification. DNA methylation modification is mainly in a specific DNA region, where a methyl group (-CH3) is added to the carbon atom structure at the 5' end of cytosine. If high DNA methylation occurs in the promoter region, gene expression will be restricted. High methylation of promoters often leads to the non-expression of tumor suppressor genes and has been reported in many cancers. Therefore, DNA methylation biomarkers are technologies with potential for development in cancer detection, diagnosis and prognosis assessment. The present invention combines artificial intelligence and epigenetic genomics to develop a methylation detection kit suitable for Chinese people to assist in the diagnosis or prediction, monitoring and treatment of urothelial carcinoma.

[0004] Current status of detection of urothelial carcinoma

[0005] Urothelial carcinoma (UC) is a malignant tumor that occurs in all urothelial tissues, including bladder cancer, renal pelvis cancer, ureter cancer, and urethral cancer. Therefore, urothelial carcinoma can involve the bladder or occur in other parts of the body. In Asia, urothelial carcinoma is mostly bladder cancer, but upper tract urothelial carcinoma that occurs in the renal pelvis and ureter is also very common, accounting for about 40%. The probability of Asians suffering from upper tract urothelial cancer is much higher than that in Western countries, and the overall incidence has been increasing year by year. Early urothelial cancer may have no symptoms or present with painless hematuria. Because it is painless, patients are less alert and easily ignored, delaying medical treatment; and other indications, such as urinary tract stones, prostatic hypertrophy, and urinary tract infection, will have symptoms of painless hematuria, so patients may ignore the warning signs and fail to diagnose early.

[0006] Most urothelial carcinomas have a 50% risk of recurrence even after surgery, so early screening and postoperative follow-up monitoring are equally important.

[0007] The definitive diagnosis of urothelial carcinoma is mainly based on urinary endoscopy and tissue biopsy. Auxiliary diagnostic tools include urine cytology and imaging tests such as intravenous pyelography (IVP) and computed tomography urography (CTU). However, cystoscopy is required for the diagnosis of bladder cancer, but this type of invasive examination often makes patients feel uncomfortable. For bladder cancer patients with a high recurrence rate and who must be followed up regularly, their acceptance is relatively low.

[0008] Therefore, developing detection technology for urine samples with high sensitivity and low invasiveness is an important topic in the development of detection technology in the field of urothelial carcinoma. According to statistics, there will be 95,000 new diagnoses of urothelial carcinoma in the Chinese population in 2022. Among them, urothelial carcinoma originating from the renal pelvis and ureter is called upper tract urothelial carcinoma (UTUC). According to data from European and American populations, the incidence of bladder cancer is about 90%, while upper tract urothelial carcinoma accounts for about 10%. Therefore, the development of methylation analysis methods based on urine or other liquid biological samples (Liquid Biopsy) can reduce harm to patients and improve convenience, and has extremely high clinical application value and market prospects.

[0009] In addition to cytological examinations and tumor protein marker tests, many methylation gene testing products for auxiliary diagnosis or postoperative monitoring of urothelial carcinoma and bladder cancer have been developed abroad, indicating that methylation gene testing is applicable for auxiliary diagnosis and postoperative monitoring of urothelial carcinoma. However, methylation modifications are easily affected by the external environment of cells. For example, the environment and lifestyle may bring about epigenetic changes.

[0010] The present invention hopes to combine artificial intelligence and the comorbidity of specific diseases for analysis, translate potential biomarkers into liquid biological samples (blood, feces, smears, secretions or urine), and use non-invasive detection methods to achieve accurate personalized early detection and early prevention; at the same time, with the assistance of AI recommendation technology, early cancer screening will be simpler and more accurate.

[0011] The present invention integrates internationally disclosed epigenomic data sets to screen and sort methylation biomarkers, and then analyzes specific disease comorbidity information. The present invention can also combine personal health records and personal methylation test data to conduct more accurate personalized risk indicator assessment.

[0012] The present invention explores methylation biomarkers of urothelial carcinoma by integrating and analyzing urothelial carcinoma comorbidity information, disease gene database system and public whole genome methylation database, and adds annotation information of gene function. It uses AI recommendation technology to effectively combine training and screening and sort DNA methylation biomarkers, disease gene function analysis and clinical comorbidity records, so as to construct a personalized and precise urothelial carcinoma prediction system and test kit.

[0013] The present invention uses a non-invasive method to collect personal samples, such as urine, blood, feces, smears, secretions and other biological fluid samples, and personal comorbidity information as a reference to construct an accurate screening method for methylation biomarker combinations of urothelial carcinoma. Summary of the invention

[0014] The present invention provides a use of at least one reagent in preparing a kit for detecting urothelial carcinoma, wherein the at least one reagent comprises a primer or a probe for detecting methylation of at least one target gene, wherein the at least one target gene comprises TMEM106A, CA3, ZIC4 or a combination thereof; and the kit performs detection by a method comprising the following steps: (a) providing an in vitro sample obtained from an individual; (b) determining the gene methylation level of the at least one target gene in the in vitro sample; and (c) comparing the methylation level of the at least one target gene in step (b) with that of a control group gene, wherein when the methylation level of the at least one target gene is higher than that of the control group gene, the individual is judged to have a high risk of suffering from urothelial carcinoma.

[0015] In one embodiment, the at least one target gene further comprises OTX1, TFAP2B, ZNF154, MEIS1, GALR1, FEZF2, SIM2, SOX17, or a combination thereof.

[0016] In one embodiment, the calculation formula for determining the methylation degree of the target gene is: the target gene ΔCt=|the target gene Ct-the control gene Ct|.

[0017] In one embodiment, the in vitro sample comprises urine, feces, blood, saliva or other body fluids.

[0018] In one embodiment, the urothelial cancer is bladder cancer, ureteral cancer, urethral cancer, or renal pelvis cancer.

[0019] In one embodiment, the control group gene is COL2A1, MYH14, PRHOXNB, GAPDH or ACTB.

[0020] The present invention further provides a method for constructing a prediction model for detecting that an individual suffers from urothelial carcinoma, the method comprising the following steps: (a) providing an in vitro sample of the individual; (b) detecting the methylation level of a target gene in the in vitro sample, wherein the target gene comprises TMEM106A, CA3, ZIC4 or a combination thereof; (c) detecting the methylation level of a control group gene, wherein the control group gene comprises COL2A1, MYH14, PRHOXNB, GAPDH or ACTB; (d) comparing the methylation levels of the target gene detected in steps (b) and (c) with those of the control group gene, using a calculation formula, wherein the calculation formula is: the target gene ΔCt = | the target gene Ct - the control group gene Ct |, as input data for the prediction model for the individual suffering from urothelial carcinoma; and (e) constructing the prediction model using the methylation difference ΔCt between the target gene and the control group gene as a basis and at least one machine learning algorithm, thereby providing a detection result of the individual suffering from urothelial carcinoma.

[0021] In one embodiment, the at least one machine learning algorithm is random forest (RF), support vector machine (SVM) or any machine learning core technology.

[0022] In one embodiment, a large amount of ΔCt data or clinical data can be used together with an applicable algorithm to enable machine learning.

[0023] In one embodiment, the target gene further comprises OTX1, TFAP2B, ZNF154, MEIS1, GALR1, FEZF2, SIM2, SOX17 or a combination thereof.

[0024] In one embodiment, the in vitro sample comprises urine, feces, blood, saliva or other body fluids.

[0025] In one embodiment, the urothelial cancer is bladder cancer, ureteral cancer, urethral cancer, or renal pelvis cancer.

[0026] The present invention further provides the use of a reagent in preparing a kit for detecting urothelial carcinoma, wherein the reagent is a primer or probe combination for detecting methylation of gene A, gene B and gene C, wherein the A gene is OTX1, TFAP2B, ZNF154, MEIS1, FEZF2, SIM2, SOX17 or ZIC4, wherein the B gene is TMEM132 or GALR1, and wherein the C gene is CA3; and the kit is detected by a method comprising the following steps: (a) providing an in vitro sample obtained from an individual; (b) determining the gene methylation level of the at least one target gene in the in vitro sample; and (c) comparing the methylation level of the at least one target gene in step (b) with that of a control group gene, wherein when the methylation level of the at least one target gene is higher than that of the control group gene, the individual is judged to have a high risk of suffering from urothelial carcinoma.

[0027] The present invention further provides a methylation biomarker screening method for screening urothelial carcinoma, comprising: (1) collecting human genome methylation data of patients with urothelial carcinoma and healthy people from at least one human genome database, wherein the human genome methylation data comprises methylation sites, methylation levels, and disease gene group information corresponding to related comorbid diseases; (2) standardizing the human genome methylation data, extracting significantly differentially expressed methylation differential sites, and selecting highly differentially expressed methylated gene groups; (3) collecting urothelial carcinoma-related comorbidities and urothelial carcinoma-related comorbid disease gene groups from at least one disease comorbidity database; (4) cross-comparing the highly differentially expressed methylated gene group with the urothelial carcinoma-related comorbid disease gene group and screening out multiple candidate biomarkers; (5) performing gene function clustering analysis of the multiple candidate biomarkers in the GO database and the KEGG database to screen out biomarker function clusters; and (6) inputting the biomarker function clusters into a machine learning algorithm or recommendation system and performing a methylation biomarker importance ranking analysis to obtain at least one methylation biomarker related to urothelial carcinoma.

[0028] In one embodiment, the methylation biomarker associated with urothelial carcinoma comprises OTX1, TFAP2B, ZNF154, TMEM106A, MEIS1, GALR1, FEZF2, CA3, SIM2, ZIC4, SOX17 or a combination thereof. In one embodiment, the step (2) further comprises performing differential methylation site analysis (DMP analyze), t-test, variance analysis and gene expression differential difference on the human gene methylation data.

[0029] In one embodiment, the step (3) is to collect the comorbid disease gene group associated with urothelial carcinoma using data mining technology.

[0030] In another embodiment, the whole genome methylation database comprises a publicly published paper database, The Cancer Genome Atlas (TCGA), GEO (Gene Expression Omnibus), iMATHYL or MathBank 3.0.

[0031] In one embodiment, the machine learning algorithm is a support vector machine (SVM) or any machine learning core technology.

[0032] In one embodiment, the recommendation system is a gene-based collaborative filtering system or any recommendation system algorithm.

[0033] In one embodiment, the urothelial cancer is bladder cancer, ureteral cancer, urethral cancer or renal pelvis cancer. In a preferred embodiment, the urothelial cancer is bladder cancer.

[0034] In one embodiment, the importance ranking analysis of step (6) is to perform a weight coefficient ranking analysis on the biomarkers, wherein the weight coefficient is the correlation between the biomarker and urothelial carcinoma, which is achieved by respectively detecting the methylation expression of the biomarkers in the biomarker functional cluster in step (5) in urothelial carcinoma patients and patients without urothelial carcinoma, and obtaining the difference between the mean value or median of the biomarker in urothelial carcinoma patients and patients without urothelial carcinoma, and the difference between the mean value or median of the biomarker in urothelial carcinoma patients and patients without urothelial carcinoma is the weight coefficient of the candidate biomarker.

[0035] In another embodiment, the at least one methylation biomarker associated with urothelial carcinoma comprises OTX1, TFAP2B, ZNF154, TMEM106A, MEIS1, GALR1, FEZF2, CA3, SIM2, ZIC4, and SOX17; or the at least one biomarker is selected for combination in each gene function group according to its gene function annotation; or at least one biomarker is recommended in each gene function group according to personal comorbidity information.

[0036] The present invention further provides a method for predicting whether an individual has at least one methylation biomarker for urothelial cancer, comprising: (1) performing disease gene analysis on the subject's comorbidity information to obtain a highly correlated gene group; (2) placing the highly correlated gene group and the methylation biomarker group associated with urothelial cancer into a machine algorithm; and (3) the machine algorithm forms a specific personalized gene combination through a gene-based collaborative filtering mechanism to accurately predict the methylation biomarker, wherein the at least one methylation biomarker associated with urothelial cancer comprises TMEM106A, CA3, ZIC4 or a combination thereof.

[0037] In one embodiment, the at least one methylation biomarker associated with urothelial cancer further comprises OTX1, TFAP2B, ZNF154, MEIS1, GALR1, FEZF2, SIM2, SOX17, or a combination thereof.

[0038] In one embodiment, the specific personalized gene panel comprises 2 to 5 of the methylation biomarkers associated with urothelial carcinoma.

[0039] In one embodiment, the urothelial cancer is bladder cancer, ureteral cancer, urethral cancer or renal pelvis cancer. In a preferred embodiment, the urothelial cancer is bladder cancer.

[0040] The present invention also provides a kit for implementing the above detection method, wherein the kit may include the above parts or parts for implementing the above method, and the kit includes primers or probes for methylation of at least one target gene and instructions for use.

[0041] Other features and advantages of the present invention will be revealed from the following non-limiting detailed description and examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Methylation biomarker screening for bladder cancer.

[0043] Figure 2 Biomarker validation process.

[0044] Figure 3 This is the procedure for clinical sample operation.

[0045] Figure 4 The results of gene testing for 11 methylation biomarkers of urothelial carcinoma. Abbreviations: Normal: healthy individuals; BLC: patients with urothelial carcinoma.

[0046] Figure 5These are the clinical test results of TMEM106A+CA3+ZIC4.

[0047] Figure 6 This is the screening process for prediction methods.

[0048] Figure 7 This is the screening result of the control group gene.

[0049] Figure 8 Screening algorithm for diagnosing patients with painless hematuria. DETAILED DESCRIPTION

[0050] The following examples are non-limiting and merely represent various aspects and features of the present invention.

[0051] Example 1: Screening of methylation biomarkers using urothelial carcinoma as an example

[0052] This test combines DNA methylation profile information and comorbidity analysis to complete the screening of disease gene groups with significant differential methylation biomarkers and comorbidity-related disease in the whole genome of urothelial carcinoma, such as Figure 1 shown.

[0053] The steps of screening the whole genome of urothelial carcinoma for significantly differentially methylated biomarkers and disease gene groups associated with comorbidities include:

[0054] (1) Collect CpG site information of human whole-genome methylation from databases such as The Cancer Genome Atlas (TCGA), GEO (Gene Expression Omnibus), iMATHYL, or MathBank 3.0, and identify CpG site information associated with urothelial carcinoma and with significant β-value changes.

[0055] (2) Then, data normalization technology (BMIQ) was used to analyze the expression of different methylation probes (DMPs), and ChAMP (Chip Analysis Methylation Pipeline) was used to conduct a comprehensive analysis of urothelial carcinoma patients (experimental group) and non-urothelial carcinoma patients (control group). t-test, analysis of variance (ANOVA), and gene expression difference were performed to extract methylation differential sites with significant differential expression in the whole genome, and select highly differentially expressed methylated gene groups.

[0056] (3) Collect urothelial carcinoma-related comorbidities and urothelial carcinoma-related comorbid disease gene groups from at least a disease comorbidity database or a publicly published paper database.

[0057] (4) The highly differentially expressed methylated gene group was cross-compared with the comorbid disease gene group related to the urothelial carcinoma-related comorbidity using the modified Jaccard index to screen out 101 candidate biomarkers (36 highly methylated genes and 65 hypomethylated genes). The 101 candidate biomarkers were then subjected to gene function clustering analysis using the GO database and the KEGG database to screen out biomarker function clusters, such as Figure 2 shown.

[0058] (5) The biomarker functional cluster was clustered using TCGA data (whole exome sequencing genomes of 419 bladder cancer tissues and 21 normal tissues, DNA methylation genome data, and related comorbidity gene data) to complete the support vector machine (SVM) machine learning training model, and the methylation biomarker importance ranking analysis was performed to obtain the top 20 biomarkers with significant differential methylation and their gene functional clustering information, among which the biomarkers with significant differential methylation included: OTX1, TFAP2B, ZNF154, TMEM106A, MEIS1, GALR1, FEZF2, CA3, SIM2, ZIC4, SOX17, LITD1, NKX2-8, GCM2, KCNA3, ALX3, FOXD3, EOMES, TRH, and ONECUT1.

[0059] (6) Finally, through the collection of clinical samples and experimental verification, the effectiveness of the system and biomarker screening, as well as the selection of the most representative biomarkers based on weight ranking and different functional grouping, was confirmed. The most representative biomarkers included the top 11 methylation biomarkers: OTX1, TFAP2B, ZNF154, TMEM106A, MEIS1, GALR1, FEZF2, CA3, SIM2, ZIC4 and SOX17.

[0060] Example 2: Clinical Experimental Verification

[0061] After parameter screening, the top 11 target genes annotated in three different gene function groups (two of which had an accuracy (ACC) > 0.9) were used in a validation experiment with 60 clinical urine samples (30 urothelial carcinoma positive samples and 30 negative samples). The clinical urine samples were mainly from urothelial carcinoma patients (experimental group) and non-urothelial carcinoma patients (control group). Figure 3The clinical sample operation procedure is to first extract the deoxyribonucleic acid (DNA) or DNA fragments from the urine sample; then use the principle that bisulfite can convert cytosine on DNA into uracil, but methylated cytosine cannot be converted into uracil to extract bisulfite-treated DNA (BS-DNA); use the principle of real-time polymerase chain reaction (qPCR) with specific primers and hydrolysis probes Detect the methylation level of the target gene.

[0062] The sensitivity, specificity and significance of the 11 methylation biomarkers for detecting urothelial carcinoma are shown in Table 1.

[0063] Table 1. Sensitivity, specificity and p-values ​​of 11 methylation biomarkers Methylation biomarkers Sensitivity Specificity p-value OTX1 80％ 65％ 0.0002 TMEM106A 80％ 55％ <0.0001 FEZF2 80％ 76％ <0.0001 ZIC4 87％ 93％ <0.0001 TFA2B 83％ 69％ <0.0001 MEIS1 87％ 21％ 0.0216 CA3 87％ 34％ 0.0001 SOX17 83％ 79％ <0.0001 ZNF154 83％ 66％ <0.0001 GALR1 80％ 83％ <0.0001 SIM2 80％ 76％ <0.0001

[0064] like Figure 4 As shown in the results, among the 11 methylation biomarker gene detection results, 8 methylation biomarkers (TFAP2B, ZNF154, TMEM106A, GALR1, FEZF2, SIM2, ZIC4, SOX17) can be used alone to effectively distinguish urothelial carcinoma from non-urothelial carcinoma using urine samples (p value <0.0001). Compared with traditional urine cytology (sensitivity of 38%), the methylation biomarkers screened by the present invention have higher sensitivity, and the experimental operation is more convenient and can be standardized.

[0065] Example 3: Optimal biomarker complementary combination

[0066] In this embodiment, the similarity of two genes is calculated according to the GO annotation of the genes, and a matrix is ​​used for hierarchical grouping. The 11 methylation biomarkers can be divided into group A, group B and group C by using Ward's method, as shown in Table 2. The higher the similarity of gene functions in the same group, the less complementary they are. When performing the optimal biomarker combination, selecting a combination of different groups will obtain better prediction results.

[0067] Table 2. Clustering results of 11 methylation biomarkers Methylation biomarkers Grouping OTX1 A TFAP2B A ZNF154 A TMEM132D B MEIS1 A GALR1 B FEZF2 A CA3 C SIM2 A ZIC4 A SOX17 A

[0068] Example 4: Methylation Biomarkers

[0069] The 11 methylation biomarkers were analyzed using different machine learning models. Three biomarkers, OTX1, ZNF154 and CA3, were selected from the 11 biomarkers as a combination and ranked according to the F1 score. Different optimal combinations were obtained. The models used included support vector machine, random forest, XGBoost tree model, K-nearest neighbors and logistic regression. The relevant results are shown in Table 3 below.

[0070] Table 3. Results from different machine learning models

[0071] Example 5: Personalized prediction system

[0072] Clinical sample verification has shown that a single methylation biomarker has limited effect in distinguishing urothelial carcinoma from non-urothelial carcinoma. Therefore, through gene function clustering and optimal gene combination analysis, any two, any three, any four or any five methylation biomarkers are combined with each other, and different gene function clustering annotations are added to obtain the most effective multiple combination of methylation biomarkers.

[0073] In one set of experiments, if the three methylation biomarkers of TMEM106A, CA3, and ZIC4 were selected, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the three methylation biomarkers of TMEM106A, CA3, and ZIC4 were 86.67%, 89.66%, 89.66%, 86.67%, and 88.13%, respectively, based on the validation results of 59 clinical samples (30 positive samples and 29 negative samples). Figure 5 As shown, the AUC score analysis and ROC curve diagram of the combination of three methylation biomarkers TMEM106A, CA3, and ZIC4, the AUC score is 0.89.

[0074] In another set of experimental cases, if the three methylation biomarkers OTX1, ZNF154, and CA3 were selected, the sensitivity was 75.00%, the specificity was 94.12%, the positive predictive value (PPV) was 85.71%, the negative predictive value (NPV) was 88.89%, and the accuracy was 88.00% based on the validation results of 25 clinical samples (8 positive samples and 17 negative samples).

[0075] In another set of experimental cases, if the three methylation biomarkers OTX1, GALR1, and ZIC4 were selected in combination, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 80 clinical samples (44 positive samples and 36 negative samples) were 93.18%, 86.11%, 89.13%, 91.189%, and 90.00%, respectively.

[0076] The present invention also establishes a method for predicting whether an individual has at least one methylation biomarker for urothelial carcinoma, which comprises: (1) performing disease gene analysis on the subject's comorbidity information to obtain highly correlated genes; (2) placing the highly correlated gene group and the methylation biomarker group related to urothelial carcinoma into a machine algorithm; and (3) the machine algorithm uses a gene-based collaborative filtering mechanism (Gene-Based Collaborative Filtering) to form a specific personalized gene combination with the highly correlated genes and the methylation biomarker for urothelial carcinoma in Example 1 to accurately predict the methylation biomarker.

[0077] In this embodiment, the machine algorithm is trained using the screening method described in Example 1.

[0078] In this embodiment, the comorbidity information of the subject includes test data.

[0079] In this embodiment, the specific personalized gene combination is a specific personalized gene combination formed by the highly correlated gene and the top 11 methylation biomarkers: OTX1, TFAP2B, ZNF154, TMEM106A, MEIS1, GALR1, FEZF2, CA3, SIM2, ZIC4 and SOX17.

[0080] In this embodiment, the specific personalized gene combination includes 2 to 5 of the top 11 methylation biomarkers, for example, the specific personalized gene combination of the subject is OTX1+TFAP2B+ZNF154.

[0081] Figure 6 The process of predictive method for detecting methylation biomarkers of urothelial carcinoma is to combine comorbidity information and use disease gene association analysis, comorbidity analysis and DNA methylation data map (TGGA) to complete personal precise detection.

[0082] Example 6: Screening of control group genes

[0083] GAPDH or ACTB are commonly used control group genes. In this example, it was found that compared with traditional housekeeping genes, MYH14 and PRHOXNB also showed excellent methylation stability and can be used as control group genes. Figure 7 As shown, the present invention adopts the machine learning algorithm of random forest or support vector machine (SVM). In most disease groups (colorectal cancer, bladder cancer, liver cancer, lung cancer, breast cancer), MYH14 and PRHOXNB can be used as control group genes for DNA methylation analysis, which can replace traditional housekeeping genes, and provide good stability and applicability in cancer detection.

[0084] Example 7: Use scenario

[0085] The urothelial carcinoma detection method of the present invention can be used as an auxiliary diagnosis or diagnostic tool. Figure 8 As shown, when diagnosing patients with painless hematuria, patients can undergo urothelial carcinoma methylation gene testing (the gene combination consists of 2 to 5 clinically validated methylation biomarker genes combined with genes with high correlation obtained through personal comorbidity analysis). If the test result is high risk, clinical specialists are recommended to perform highly invasive cystoscopy or urinary endoscopy on the patient to confirm the disease; if the test result is asymptomatic and low-risk, clinical specialists are recommended to continue to follow up the patient and implement other symptom or disease-related treatments.

[0086] While the invention has been described and illustrated in sufficient detail for those skilled in the art to make and use it, it should be apparent that various substitutions, modifications and improvements will occur without departing from the spirit and scope of the invention.

[0087] Those skilled in the art will readily appreciate that the present invention is well adapted to carry out the above-mentioned objects and obtain the above-mentioned objects and advantages, as well as those inherent therein. The above-mentioned processes and methods for producing them represent preferred embodiments, are exemplary, and do not limit the scope of the invention. Those skilled in the art will recognize modifications and other uses thereof. Such modifications are included within the spirit of the invention and are limited by the scope of the claims.

Claims

1. Use of at least one reagent in the preparation of a kit for detecting urothelial carcinoma, wherein the at least one reagent comprises a primer or a probe for detecting methylation of at least one target gene, Wherein the at least one target gene comprises TMEM106A, CA3, ZIC4 or a combination thereof; And the kit is detected by a method comprising the following steps: (a) providing an in vitro sample obtained from an individual; (b) determining the gene methylation level of the at least one target gene in the in vitro sample; and (c) comparing the methylation level of the at least one target gene in step (b) with that of a control group gene, wherein when the methylation level of the at least one target gene is higher than that of the control group gene, the individual is judged to have a high risk of urothelial cancer.

2. The use according to claim 1, wherein the at least one target gene further comprises OTX1, TFAP2B, ZNF154, MEIS1, GALR1, FEZF2, SIM2, SOX17 or a combination thereof.

3. The use according to claim 1, wherein the in vitro sample comprises urine, feces, blood, saliva or other body fluids.

4. The use according to claim 1, wherein the urothelial cancer is bladder cancer, ureteral cancer, urethral cancer or renal pelvis cancer.

5. The use according to claim 1, wherein the control group genes are COL2A1, MYH14, PRHOXNB, GAPDH or ACTB.

Images