Bladder cancer prognosis prediction model and bladder cancer prognosis marker

By combining the intratumor microbiome and programmed cell death-related genes, we construct predictive models and markers for bladder cancer prognosis, the problem of difficulty in predicting bladder cancer prognosis in the prior art is solved, and effective evaluation of the prognosis of bladder cancer patients is achieved. Identification of high-risk populations provides important clinical guidance.

CN118942714APending Publication Date: 2024-11-12SHANDONG YINFENG LIFE SCIENCE RESEARCH INSTITUTE +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410955031.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The prior art is difficult to effectively predict the prognosis of bladder cancer, especially muscle-invasive bladder cancer has few early symptoms, fast progress, poor prognosis, and lack of effective biomarkers to guide treatment decisions.

Method used

By combining differentially expressed genes related to intratumor microbiome and programmed cell death, predictive models and prognostic markers for bladder cancer prognosis were constructed, and data were obtained from the TCGA database using machine learning methods, and 10 differentially expressed microbiomes and 16 were screened out. Genetic markers, risk scoring models are constructed to independently predict the prognosis of bladder cancer.

Benefits of technology

This model can effectively predict the survival time and prognosis of bladder cancer patients, provide independent predictors, help identify high-risk groups and guide treatment decisions, significantly improving the accuracy and clinical guidance of bladder cancer prognosis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942714A_ABST
    Figure CN118942714A_ABST
Patent Text Reader

Abstract

The invention discloses a prediction model for prognosis of bladder cancer. The prediction model is a calculation formula as follows: RiskScore is equal to Luteibacter * (-0.4499) + Lachnostostridium * (-0.2013) + Microvirgula * 0.30) + Olyya * 0.0835 + Actinomadura * 0.1519 + Pilimeria * 0.2154 + Dermatophauls * 0.2904 + SLC39A5 * (-0.1735) + MBOAT < 2 > * 0.3151 + NAPSA * (-0.0799) + MAP < 2 > * 0.1621 + UCHL < 1 > * 0.045 + BDNF * 0.0113 + ATP6V1B1 * 0.042 + PEX < 5 > L * 0.1839. The The invention can be used for prognosis evaluation of bladder cancer patients, and has important significance on prognosis diagnosis, evaluation, treatment and the like of bladder cancer.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a prediction model for bladder cancer prognosis and a bladder cancer prognosis marker, belonging to the technical field of tumor prediction models and the technical field of gene detection. Background Art

[0002] Bladder cancer (BC) is one of the most commonly diagnosed cancer types in the world, with high morbidity and mortality. Bladder cancer is usually divided into muscle invasive bladder cancer (MIBC), non-muscle invasive bladder cancer (NMIBC) and distant metastatic tumors. About 70% of BC occurs in the urothelial cells in the innermost layer of the bladder. Although the recurrence rate of NMIBC is quite high (40% to 80%), its lethality is lower and easier to treat, and the tumor is usually removed through the urethra. The incidence of MIBC, in which the tumor grows into the bladder muscle wall, is about 30%, and it is very easy to metastasize. MIBC has few early symptoms, rapid progression, and poor prognosis. The 5-year survival rate of stage 3 is less than 50%. Despite radical cystectomy and pelvic lymph node dissection, about half of MIBC patients develop metastatic disease within two years of diagnosis. Therefore, it is still necessary to identify new and effective biomarkers.

[0003] Recent studies have shown that microorganisms are believed to affect the occurrence, development, metastasis and treatment response of different tumors, especially closely related to the tumor microenvironment and immune response. Many basic studies have shown that the microbiome in tumors affects the progression, metastasis, prognosis and immunotherapy of tumor patients by regulating oxidative stress, toll-like receptor-mediated immune response and tumor cell metabolism, involving signaling pathways such as mTOR, STAT3, Wnt, and MAPK. Therefore, microbial-tumor interactions make the microbiome an effective biomarker and therapeutic target for a variety of tumors.

[0004] The discovery of microbiota in several tumors that were initially considered sterile has strengthened the concept of the intratumoral microbiota, and emerging evidence has confirmed that the intratumoral microbiota plays a key role in the carcinogenic behavior of pancreatic, lung, nasopharyngeal, and breast cancers. In bladder cancer, studies using RNA sequencing data have revealed the microbial composition of the BC tissue microenvironment, and the bladder microbiome has been associated with the epithelial-mesenchymal transition in muscle-invasive urothelial bladder cancer. However, the correlation of the intratumoral microbiome with the clinicopathological characteristics, prognosis, and tumor microenvironment heterogeneity of BC remains unreported.

[0005] Programmed cell death (PCD), also known as regulated cell death, refers to a specific form of cell death that can be controlled by a variety of biomacromolecules. For decades, PCD has been considered a key mechanism controlling the complex dynamics of tumor occurrence and progression. The ability of tumor cells to evade or resist PCD has a profound impact on their unrestricted proliferation and malignant development. Disrupting the complex process of PCD metabolic pathways can lead to the accumulation of genetically damaged or abnormal cells, thereby promoting their persistence and uncontrolled proliferation, ultimately leading to the formation of tumor masses. Summary of the invention

[0006] In view of the above-mentioned prior art, the present invention provides a prediction model for bladder cancer prognosis and its application, and a group of bladder cancer prognosis markers and their application.

[0007] The present invention is achieved through the following technical solutions:

[0008] A prediction model for bladder cancer prognosis is calculated as follows:

[0009] RiskScore=Luteibacter×(-0.4499)+Lachnoclostridium×(-0.2013)+Microvirgula×0.30)+Olleya×0.0835+Actinomadura×0.1519+Pilimelia×0.2154+D ermatophilus×0.2904+SLC39A5×(-0.1735)+MBOAT2×0.3151+NAPSA×(-0.0799)+MAP2×0.1621+UCHL1×0.045+BDNF×0.0113+ATP6V1B1×0.042+PEX5L×0.1839;

[0010] Among them, RiskScore is the risk score of bladder cancer samples, Luteibacter, Lachnoclostridium, Microvirgula, Olleya, Actinomadura, Pilimelia and Dermatophilus represent the abundance of corresponding microorganisms; SLC39A5, MBOAT2, NAPSA, MAP2, UCHL1, BDNF, ATP6V1B1 and PEX5L represent the expression levels of corresponding genes (expression matrix values); -0.4499, -0.2013, 0.307, 0.0835, 0.1519, 0.2154, 0.2904, -0.1735, 0.3151, -0.0799, 0.1621, 0.045, 0.0113, 0.042 and 0.1839 are the regression coefficients of corresponding markers.

[0011] The application of the prediction model for bladder cancer prognosis in bladder cancer prognosis prediction and in the preparation of bladder cancer prognosis prediction products.

[0012] A group of prognostic markers for bladder cancer includes 7 marker microorganisms and 8 marker genes. The genera of the 7 marker microorganisms are Luteibacter, Lachnoclostridium, Microvirgula, Olleya, Actinomadura, Pilimelia, and Dermatophilus; the names of the 8 marker genes are SLC39A5, MBOAT2, NAPSA, MAP2, UCHL1, BDNF, ATP6V1B1, and PEX5L.

[0013] The application of the bladder cancer prognosis marker in the preparation of a bladder cancer prognosis prediction product.

[0014] Furthermore, the bladder cancer prognosis prediction product is a detection reagent or kit for bladder cancer prognosis markers; the detection reagent or kit for bladder cancer prognosis markers includes: a reagent or kit for detecting the abundance of 7 marker microorganisms in bladder cancer prognosis markers, and a reagent or kit for detecting the expression levels of 8 marker genes in bladder cancer prognosis markers.

[0015] The use of a reagent or kit for detecting a bladder cancer prognosis marker in the preparation of a bladder cancer prognosis prediction product, wherein the reagent or kit for detecting a bladder cancer prognosis marker comprises: a reagent or kit for detecting the abundance of 7 marker microorganisms in the bladder cancer prognosis marker, and a reagent or kit for detecting the expression levels of 8 marker genes in the bladder cancer prognosis marker.

[0016] The bladder cancer prognosis prediction model and bladder cancer prognosis marker of the present invention can be used for the prognosis evaluation of bladder cancer patients. The present invention has important significance for the prognosis diagnosis, evaluation and treatment of bladder cancer.

[0017] The present invention obtains data of 403 bladder cancer patients from the TCGA database through cBioPortal, and uses machine learning methods to identify microbiomes associated with prognosis. Then, the differentially expressed genes (DEGs) related to programmed cell death and the microbiomes are combined to construct a tumor prognostic risk scoring model and a multi-parameter nomogram model. The 10 differentially expressed microbiomes obtained by machine learning can independently predict the prognosis of bladder cancer. In addition, the risk scoring model constructed by microbial markers and gene markers can effectively predict tumor prognosis. In this cohort study, the intratumoral microbiome is a powerful prognostic tool for bladder cancer patients, providing potential guidance for treatment decisions for patients with different levels of risk of malignant progression.

[0018] This invention is the first to evaluate the clinical significance of the presence of intratumoral microbiota in BC patients, and to predict prognosis in combination with programmed cell death genes. In the study of the present invention, the RNA-Seq expression profile obtained from the Cancer Genome Atlas (TCGA) provides an opportunity to study the relationship between intratumoral microbiota, gene expression and BC prognosis. TCGA data of microbiome data and clinical characteristics are used to evaluate the role of intratumoral microbiota in BC prognosis. The study found that the intratumoral microbiota can be used as an independent predictor of BC survival prognosis, and the RS model containing microbiome and PCD genes was systematically evaluated, highlighting the important role of the microbiome in BC prognosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 : Volcano plot of the abundance and prognosis of differential microbial markers.

[0020] Figure 2 : Heat map of relative abundance of microbial markers in samples.

[0021] Figure 3 : Forest plot of univariate Cox regression results.

[0022] Figure 4 : Schematic diagram of the cumulative residual curves of four machine learning models.

[0023] Figure 5 : Schematic diagram of ROC curves of four machine learning models.

[0024] Figure 6 : Schematic diagram of microbial importance ranking of random forest model.

[0025] Figure 7 : Microbial one-factor Cox forest plot of random forest model.

[0026] Figure 8 : PCD gene distribution ring diagram.

[0027] Fig. 9 : KEGG enrichment analysis diagram of gene markers.

[0028] Fig.10 :GO enrichment analysis diagram of gene markers.

[0029] Fig.11 : LASSO regression coefficient plot.

[0030] Fig.12 : LASSO regression parameter plot.

[0031] Fig.13 : Schematic diagram of the survival curve of the training set.

[0032] Fig.14: Schematic diagram of the survival curve of the test set.

[0033] Fig.15 : Schematic diagram of the time-ROC curve of the training set.

[0034] Fig.16 : Schematic diagram of the time-ROC curve of the test set.

[0035] Fig.17 : Scatter plot of riskscore and survival status. DETAILED DESCRIPTION

[0036] The present invention will be further described below in conjunction with the embodiments. However, the scope of the present invention is not limited to the following embodiments. It will be appreciated by those skilled in the art that various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention.

[0037] The instruments, reagents, and materials involved in the following examples, unless otherwise specified, are all conventional instruments, reagents, and materials already available in the prior art and can be obtained through regular commercial channels. The experimental methods and detection methods involved in the following examples, unless otherwise specified, are all conventional experimental methods and detection methods already available in the prior art.

[0038] Example 1 Prediction model for bladder cancer prognosis

[0039] 1. Data Acquisition

[0040] The RNA expression profile, clinical information, and intratumoral microbiome characteristics of BC were obtained from the “Bladder Urothelial Carcinoma (TCGA, PanCancer Atlas)” dataset of the cbioportal database, the URL is as follows: https: / / cbioportal-datahub.s3.amazonaws.com / blca_tcga_pan_can_atlas_2018.tar.gz.

[0041] 2. Screening for microbial markers significantly associated with survival

[0042] 2.1 Analysis of differential abundance of microorganisms

[0043] After excluding patients with incomplete clinical information, 403 samples remained. Patients were grouped according to the median of their survival time (OS_MONTHS). Those with OS_MONTHS ≥ the median were classified as "high survivors" (HS), and those with OS_MONTHS < the median were classified as "low survivors" (LS). The patient ID lists of the "high survivors" and "low survivors" tumor groups were uploaded to the cbioportal platform. The "Compare" function under Groups can compare the microbiome characteristics between the groups. The original number of microorganisms was 1407, and 1048 microorganisms remained after decontamination. After screening the microorganisms with a difference in microbial abundance pvalue < 0.05 between the two groups of patients, 73 microorganisms were retained. The volcano plot of the abundance of differential microbial markers is shown in the figure below. Figure 1 As shown in the figure, the relative abundance heat map of microbial markers in the samples is shown in Figure 2 shown.

[0044] 2.2 Univariate Cox analysis of microbial markers

[0045] The R package "survival" was used to perform univariate Cox regression analysis on the normalized data of 73 microbial logCPM. The forest plot of the univariate Cox regression results is shown in the figure below. Figure 3 As shown. 47 microorganisms with pvalue < 0.05 were retained for the next step of machine learning model screening.

[0046] 2.3 Microbial marker machine learning model screening

[0047] Four machine learning models were built and tested using the R packages “caret” and “DALEX”. Cumulative residual curves and ROC curves were plotted using random forest (RF), support vector machine (SVM), extreme gradient boosting (XGB), and generalized linear model (GLM). The cumulative residual curves are shown in Figure 4 As shown, the ROC curve is Figure 5 As shown by Figure 4 , Figure 5 It can be seen that the RF model is the most effective machine learning model. The importance of microorganisms in the random forest model was ranked, and the results are as follows: Figure 6 As shown in the figure, the top ten important microorganisms were selected. In order to verify the value of the microbiome as an independent prognostic indicator for BC patients, a univariate Cox regression analysis was performed, as shown in the figure. Figure 7 As shown, the microbiome was confirmed to be an important risk factor (pvalue<0.001), among which 3 genera were negatively correlated with the risk of death (more in the HS group, called "good bacteria"), and 7 genera were positively correlated with the risk of death (more in the LS group, called "bad bacteria").

[0048] 3. Screening of genetic markers significantly associated with survival

[0049] The gene expression matrix of 403 samples was sorted. The number of genes in the original gene expression matrix was 20,511. 17,459 genes expressed in more than half of the samples were retained. The R package "edgeR" was used for differential gene analysis. The threshold was set to p value < 0.05 and |log2FC|> 0.5. At this time, there were 1,266 differentially expressed genes. After intersection with the PCD gene set, the survival curve was used to screen the genes with p value < 0.05, and 16 key differential genes related to PCD in BC were obtained. The classified ring diagram was visualized for the programmed cell death mechanism to which they belonged, as shown in the following figure: Figure 8 As shown in the figure, the KEGG enrichment analysis of gene markers is shown in Fig. 9 As shown in the figure, the GO enrichment analysis of gene markers is shown in Fig.10 As shown, GO and KEGG enrichment analysis results showed that these 16 key differential genes were involved in various carcinogenesis-related pathways.

[0050] 4. Build and validate risk scoring models

[0051] A prognostic scoring system for BC was established using 26 variables, including 10 microbial markers (screened in 2 above) and 16 gene markers (screened in 3 above) as predictors. Patients were randomly divided into a training set (n=282) and a test set (n=121) at a ratio of 7:3. LASSO regression analysis was performed on the training set. The LASSO regression coefficients are shown in the figure below. Fig.11 As shown, the LASSO regression parameter diagram is as follows Fig.12 As shown, the best model was determined based on the model coefficient graph and parameter graph, and this method generated a prediction model consisting of 7 microbiome groups and 8 genes, which had good prediction accuracy for the survival time of BC patients (C-index = 0.74).

[0052] Table 1 shows the coefficient values ​​of each variable in the LASSO regression analysis.

[0053] Table 1

[0054]

[0055]

[0056] The risk score for each sample is calculated based on the following formula:

[0057] Riskscore i=∑Ni(expresion / abundance of variables j*coefβj);

[0058] Among them, expression / abundance of variables represents the gene expression of the corresponding predictor or the abundance of the microbial marker, coefβ represents the regression coefficient of the corresponding predictor in the lasso regression result, Riskscore represents the sum of the gene expression of the predictor or the abundance of the microbial marker in each sample multiplied by the corresponding regression coefficient, i represents the sample, and j represents the marker.

[0059] Substituting each regression coefficient into the calculation formula, we get the following prediction model:

[0060] RiskScore=Luteibacter×(-0.4499)+Lachnoclostridium×(-0.2013)+Microvirgula×0.30)+Olleya×0.0835+Actinomadura×0.1519+Pilimelia×0.2154+D ermatophilus×0.2904+SLC39A5×(-0.1735)+MBOAT2×0.3151+NAPSA×(-0.0799)+MAP2×0.1621+UCHL1×0.045+BDNF×0.0113+ATP6V1B1×0.042+PEX5L×0.1839;

[0061] Among them, RiskScore is the risk score of each bladder cancer sample, Luteibacter, Lachnoclostridium, Microvirgula, Olleya, Actinomadura, Pilimelia, Dermatophilus are the names of microorganisms; SLC39A5, MBOAT2, NAPSA, MAP2, UCHL1, BDNF, ATP6V1B1, PEX5L are the names of genes; -0.4499, -0.2013, 0.307, 0.0835, 0.1519, 0.2154, 0.2904, -0.1735, 0.3151, -0.0799, 0.1621, 0.045, 0.0113, 0.042, 0.1839 are the regression coefficients of the corresponding markers.

[0062] According to the risk score, the patients were divided into HRG (high-risk group) and LRG (low-risk group). Survival analysis was performed on the high-risk and low-risk groups in the training set and the test set based on the model score. The survival curve of the training set is shown in Fig.13 As shown, the survival curve of the test set is Fig.14 As shown, the survival curve analysis showed that the prognosis of HRG was significantly worse than that of LRG in both the training set and the test set.

[0063] The time-ROC curve of the training set is as follows Fig.15 As shown in Figure 2, the areas under the ROC curve for the 1-year, 3-year, and 5-year OS rates in the training set were 0.74, 0.71, and 0.72, respectively; the time-ROC curve of the test set was as follows: Fig.16 As shown, the areas under the ROC curve of the test set were 0.73, 0.68, and 0.73, respectively, indicating the effectiveness of the model in predicting the prognosis of BS.

[0064] The scatter plot of Riskscore and survival status is as follows Fig.17 As shown, it can be seen that the higher the Riskscore, the higher the mortality rate and the shorter the patient's survival time. The final results show that the prediction of the prognostic scoring model of BC has more clinical guiding significance for the screening and treatment of high-risk groups of bladder cancer.

[0065] The above examples are provided to those skilled in the art to fully disclose and describe how to make and use the claimed embodiments, and are not intended to limit the scope of the disclosure herein. Modifications that are obvious to those skilled in the art will fall within the scope of the appended claims.

Claims

1. A prediction model for bladder cancer prognosis, It is characterized in that The calculation formula is as follows: RiskScore=Luteibacter×(-0.4499)+Lachnoclostridium×(-0.2013)+Microvirgula×0.30)+Olleya×0.0835+Actinomadura×0.1519+Pilimelia×0.2154+D ermatophilus×0.2904+SLC39A5×(-0.1735)+MBOAT2×0.3151+NAPSA×(-0.0799)+MAP2×0.1621+UCHL1×0.045+BDNF×0.0113+ATP6V1B1×0.042+PEX5L×0.1839; Among them, RiskScore is the risk score of bladder cancer samples, Luteibacter, Lachnoclostridium, Microvirgula, Olleya, Actinomadura, Pilimelia and Dermatophilus represent the abundance of corresponding microorganisms; SLC39A5, MBOAT2, NAPSA, MAP2, UCHL1, BDNF, ATP6V1B1 and PEX5L represent the expression levels of corresponding genes; -0.4499, -0.2013, 0.307, 0.0835, 0.1519, 0.2154, 0.2904, -0.1735, 0.3151, -0.0799, 0.1621, 0.045, 0.0113, 0.042, 0.1839 are the regression coefficients of corresponding markers.

2. Use of the prediction model for bladder cancer prognosis according to claim 1 in the prediction of bladder cancer prognosis, or in the preparation of a bladder cancer prognosis prediction product.

3. The use according to claim 2, Features: The bladder cancer prognosis prediction product is a detection reagent or kit for bladder cancer prognosis markers; the detection reagent or kit for bladder cancer prognosis markers includes: a reagent or kit for detecting the abundance of 7 marker microorganisms in bladder cancer prognosis markers, and a reagent or kit for detecting the expression levels of 8 marker genes in bladder cancer prognosis markers.

4. A panel of prognostic markers for bladder cancer, Features: It includes 7 marker microorganisms and 8 marker genes. The genus names of the 7 marker microorganisms are Luteibacter, Lachnoclostridium, Microvirgula, Olleya, Actinomadura, Pilimelia, and Dermatophilus; the names of the 8 marker genes are SLC39A5, MBOAT2, NAPSA, MAP2, UCHL1, BDNF, ATP6V1B1, and PEX5L.

5. Use of the bladder cancer prognosis marker according to claim 4 in the preparation of a bladder cancer prognosis prediction product.

6. The use according to claim 5, Features: The bladder cancer prognosis prediction product is a detection reagent or kit for bladder cancer prognosis markers; the detection reagent or kit for bladder cancer prognosis markers includes: a reagent or kit for detecting the abundance of 7 marker microorganisms in bladder cancer prognosis markers, and a reagent or kit for detecting the expression levels of 8 marker genes in bladder cancer prognosis markers.

7. Use of the detection reagent or kit for bladder cancer prognosis markers according to claim 4 in the preparation of a bladder cancer prognosis prediction product, It is characterized in that The detection reagent or kit for bladder cancer prognosis markers includes: a reagent or kit for detecting the abundance of 7 marker microorganisms in bladder cancer prognosis markers, and a reagent or kit for detecting the expression levels of 8 marker genes in bladder cancer prognosis markers.