Liver cancer prognosis marker and application thereof

Abstract

The invention discloses a prognostic marker for liver cancer and application of the prognostic marker. The prognostic marker for liver cancer comprises TXNRD1, HAO1, C1S, OPTN and SNX6. The invention provides a group of liver cancer prognosis markers, and a liver cancer patient prognosis prediction model is constructed based on the liver cancer prognosis markers. The liver cancer prognosis marker is obtained by synthesizing gene expression level, single cell analysis and space transcriptome analysis, and the diversity and large sample size of a data set improve the reliability and universality of the liver cancer prognosis marker and a liver cancer patient prognosis prediction model.

Description

Technical Field

[0001] The invention belongs to the technical field of bioinformatics, and specifically relates to a liver cancer prognosis marker and application thereof. Background Art

[0002] In recent years, immunotherapy, especially immune checkpoint inhibitors such as PD-1 / PD-L1 inhibitors, has shown great potential in the treatment of various tumors, bringing new hope for the treatment of liver cancer. However, the liver cancer microenvironment has unique immunosuppressive properties, resulting in a low response rate to immunotherapy and limited clinical benefits. Therefore, it is of great significance to deeply understand the characteristics of the liver cancer immune microenvironment and discover markers that can predict the efficacy of immunotherapy to improve the level of liver cancer immunotherapy.

[0003] The development of multi-omics technologies such as genomics, transcriptomics, and proteomics has provided a new way to analyze the tumor immune microenvironment. Using new technologies such as single-cell sequencing and spatial transcriptome sequencing, tumor heterogeneity and immune cell infiltration can be studied at the single-cell level and spatial dimension, and new immunotherapy-related markers are expected to be discovered. Integrated multi-omics data analysis can also reveal the regulatory relationship between different omics and deepen the understanding of the immune regulatory mechanism of liver cancer. In addition, the combination of omics big data and machine learning methods provides new ideas for building efficient and accurate tumor diagnosis and prognosis prediction models. The diagnostic and prognostic models constructed based on key molecular markers are expected to achieve patient stratification and personalized treatment, and improve the accuracy of liver cancer immunotherapy. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0005] To this end, the present invention proposes a liver cancer prognosis marker, which can be used to construct a prognosis prediction model for liver cancer patients, and further used to predict the efficacy of liver cancer immunotherapy, thereby promoting the precision of liver cancer immunotherapy.

[0006] The present invention also proposes a method for constructing a prognosis prediction model for liver cancer patients.

[0007] The invention also provides a liver cancer patient prognosis prediction system.

[0008] The invention also provides an electronic device.

[0009] The present invention also provides a storage medium.

[0010] According to a first aspect of the present invention, a liver cancer prognosis marker is proposed, wherein the liver cancer biomarker comprises TXNRD1, HAO1, C1S, OPTN and SNX6.

[0011] According to a second aspect of the present invention, a method for constructing a prognosis prediction model for liver cancer patients is proposed, and the construction method comprises the following steps:

[0012] S1: Obtain a gene expression dataset of cancer tissues from patients with liver cancer who have received immunotherapy, and screen for differentially expressed genes between patients who respond to and do not respond to immunotherapy;

[0013] S2: Obtain a gene expression dataset of cancer tissues of patients with liver cancer who responded to immunotherapy before and after treatment, and screen for differentially expressed genes between cancer tissues before and after immunotherapy;

[0014] S3: comparing the differentially expressed genes obtained in step S1 and step S2 to obtain genes that are differentially expressed in both sets of gene expression data sets, i.e., immune response-related genes;

[0015] S4: obtaining a gene expression dataset of cancer tissues of liver cancer patients, performing univariate and multivariate logistic regression analysis on the immune response-related genes obtained in step S3 in the dataset to determine differentially expressed genes associated with the prognosis of liver cancer patients, and establishing a risk model;

[0016] The differentially expressed genes associated with liver cancer include TXNRD1, HAO1, C1S, OPTN and SNX6.

[0017] In some embodiments of the present invention, the gene expression dataset of cancer tissue of liver cancer patients after immunotherapy in step S1 includes a spatial transcriptome dataset.

[0018] In some embodiments of the present invention, the immunotherapy in step S1 and step S2 is immune checkpoint inhibitor therapy.

[0019] In some embodiments of the present invention, the immune checkpoint inhibitor therapy is selected from one or more of cabozantinib, nivolumab, pembrolizumab, tesimumab and durvalumab.

[0020] In some embodiments of the present invention, the risk model is a logistic regression model.

[0021] In some embodiments of the present invention, the calculation formula of the risk model is: risk score = 0.0146 × TXNRD1 expression level - 0.0014 × C1S expression level + 0.0114 × OPTN expression level + 0.0304 × SNX6 expression level - 0.0019 × HAO1 expression level.

[0022] In some embodiments of the present invention, the TXNRD1 expression level, HAO1 expression level, C1S expression level, OPTN expression level and SNX6 expression level are all values ​​obtained by global scaling and normalization of gene expression levels using a function.

[0023] In some embodiments of the present invention, the global scaling normalization uses a NormalizeData function.

[0024] In some embodiments of the present invention, when the risk score is greater than or equal to 0, it is determined as a high risk, and when the risk score is less than 0, it is determined as a low risk.

[0025] According to a third aspect of the present invention, a liver cancer patient prognosis prediction system is proposed, and the liver cancer patient prognosis prediction system comprises:

[0026] Input unit: used to input the data to be evaluated;

[0027] Analysis unit: a liver cancer patient prognosis prediction model constructed by the construction method described in the second aspect of the present invention, used to analyze the data to be evaluated in the input unit;

[0028] Evaluation unit: used to display the risk score of the prognostic prediction model for liver cancer patients;

[0029] When the risk score is greater than or equal to 0, it is determined as a high risk, and when the risk score is less than 0, it is determined as a low risk.

[0030] According to a fourth aspect of the present invention, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the construction method described in the second aspect of the present invention is implemented.

[0031] According to a fifth aspect of the present invention, a storage medium is provided, wherein the storage medium stores processor-executable instructions, and the processor-executable instructions, when executed by the processor, are used to execute the construction method described in the second aspect of the present invention.

[0032] The present invention has at least the following beneficial effects:

[0033] The present invention provides a group of liver cancer prognostic markers, and constructs a liver cancer patient prognosis prediction model based on these liver cancer prognostic markers; the liver cancer prognostic markers are obtained by comprehensive gene expression level, single cell analysis and spatial transcriptome analysis, and the diversity of the data set and the large sample size improve the reliability and universality of the liver cancer prognostic markers and the liver cancer patient prognosis prediction model;

[0034] The liver cancer patient prognosis prediction system provided by the present invention has shown the potential to predict survival risk in different populations in an independent validation data set, has good stability and universality, and provides a valuable reference for clinical decision-making and individualized treatment.

[0035] The liver cancer prognostic markers and liver cancer patient prognosis prediction model provided by the present invention not only deepen our understanding of the mechanism of liver cancer immunotherapy, but also provide new ideas for screening patients suitable for immunotherapy and predicting treatment effects. In addition, combining the prognostic markers provided by the present invention with existing biomarkers and clinical indicators is expected to develop more accurate diagnostic and prognostic assessment tools, thereby promoting the precision of liver cancer immunotherapy.

[0036] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments, wherein:

[0038] Figure 1 This is a schematic diagram of the distribution of various types of cells in the spatial transcriptome data in Example 1 of the present invention;

[0039] Figure 2 The results of the analysis of differentially expressed genes between the immunotherapy response group and the non-response group in the spatial transcriptome data in Example 1 of the present invention are shown in Figure A; wherein Figure A is a malignant cell obtained by merging 7 sample sources using the Harmony algorithm, Figure B is a heat map of the expression patterns of differentially expressed genes in the two groups of samples, Figure C is a GO gene enrichment result of the up-regulated genes in the response group, and Figure D is a GO gene enrichment result of the down-regulated genes in the response group;

[0040] Figure 3 It is a graph of differentially expressed gene analysis results between samples before and after immunotherapy in single-cell transcriptome data in Example 1 of the present invention; wherein, Figure A is the single-cell annotation results of H68a (before treatment) and H68b (after treatment) samples, Figure B is the GO gene enrichment results of up-regulated genes in samples after treatment, and Figure C is the GO gene enrichment results of down-regulated genes in samples after treatment;

[0041] Figure 4 It is a graph of the analysis results of differentially expressed genes between samples before and after immunotherapy in the single-cell transcriptome data in Example 1 of the present invention; wherein, Figure A is a heat map of the expression patterns of differentially expressed genes in the two groups of samples, and Figure B is the expression levels of the two genes MT1E and SPP1 in the samples before and after treatment;

[0042] Figure 5This is a graph showing the expression difference of MT1E in spatial transcriptome data in Example 1 of the present invention;

[0043] Figure 6 This is a graph showing the expression difference of SPP1 in spatial transcriptome data in Example 1 of the present invention;

[0044] Figure 7 The PCA dimensionality reduction analysis results of the four data sets used in Examples 1 and 2 of the present invention are shown in Figure 1; Figure A is GSE14520, Figure B is GSE36376, Figure C is GSE25097, and Figure D is PXD006512;

[0045] Figure 8 The results of using LASSO regression to screen liver cancer prognostic markers in Example 1 of the present invention are shown in Figure A; Figure A shows the change of the LASSO regression coefficient of the GSE14520 data set with the penalty parameter, Figure B shows the result of cross-validation to select the optimal model; Figure C shows the top 12 stable prognosis-related genes in 100 repeated modelings;

[0046] Fig. 9 This is a graph showing the results of the correlation analysis between TCGA prognosis data and the expression levels of liver cancer prognosis markers in Example 2 of the present invention;

[0047] Fig.10 This is a graph showing the results of the correlation analysis between TCGA prognosis data and the expression levels of liver cancer prognosis markers in Example 2 of the present invention;

[0048] Fig.11 The results of the association analysis between liver cancer prognostic markers and immune microenvironment scores in Example 2 of the present invention are shown in Figure A; Figure A is a heat map of the correlation between liver cancer prognostic markers and Stromal score, Immune score and ESTIMATE score, and Figure B is a heat map of the correlation between liver cancer prognostic markers and 24 immunotherapy response-related scores;

[0049] Fig.12 The Kaplan-Meier survival curve diagram of the prognosis prediction model for liver cancer patients verified by four data sets in Example 4 of the present invention; wherein, Figure A is TCGA data, Figure B is ICGC data, Figure C is GSE14520 data set, and Figure D is GSE76427;

[0050] Fig.13The results of the correlation analysis between the prognosis prediction model for liver cancer patients in Example 4 of the present invention and the TCGA clinical data are shown in Figure A; wherein, Figure A shows the patient's survival status, Figure B shows the primary tumor staging, Figure C shows the lymph node involvement staging, Figure D shows the distant metastasis staging, Figure E shows the Stage staging, Figure F shows the Grade grading, Figure G shows whether the patient carries hepatitis B or hepatitis C, and Figure H shows the patient's age;

[0051] Fig.14 It is a calibration curve diagram drawn based on three data sets in Example 4 of the present invention; wherein, Figure A is TCGA data, Figure B is ICGC data, and Figure C is GSE14520 data set;

[0052] Fig.15 The ROC curves of 1-year, 3-year and 5-year survival periods drawn based on the data set GSE76427 in four embodiments of the present invention;

[0053] Fig.16 The statistical result diagram of analyzing the expression levels of five liver cancer prognostic markers and the risk scores predicted by the model in Example 4 of the present invention; wherein, Figure A is TCGA data, Figure B is ICGC data, and Figure C is GSE14520 data set;

[0054] Fig.17 This is a result diagram of validating the risk score of the liver cancer prognosis prediction model based on self-collected clinical liver cancer samples in Example 4 of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the embodiments to clearly and completely describe the concept of the present invention and the technical effects produced, so as to fully understand the purpose, characteristics and effects of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention.

[0056] In the description of the present invention, "several" means more than one, "many" means more than two, "greater than", "less than", "exceed", etc. are understood to exclude the number itself, and "above", "below", "within", etc. are understood to include the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0057] The following will be combined with the embodiments to clearly and completely describe the concept of the present invention and the technical effects produced, so as to fully understand the purpose, characteristics and effects of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention.

[0058] When a numerical range is disclosed herein, the above range is considered to be continuous and includes the minimum and maximum values ​​of the range, as well as every value between such minimum and maximum values. Further, when a range refers to an integer, every integer between the minimum and maximum values ​​of the range is included. In addition, when multiple ranges are provided to describe features or characteristics, the ranges can be combined. In other words, unless otherwise indicated, all ranges disclosed herein should be understood to include any and all subranges included therein.

[0059] Unless otherwise specified, the reagents, methods and equipment used in the present invention are conventional reagents, methods and equipment in the art.

[0060] Example 1 Screening of liver cancer prognostic markers

[0061] In this example, a set of liver cancer prognostic markers was obtained by integrating and screening the spatial transcriptome dataset, the single-cell transcriptome dataset and the gene expression dataset. The dataset information used is shown in Table 1;

[0062] Table 1 Dataset information used for screening of prognostic markers for liver cancer

[0063]

[0064] The specific screening methods are as follows:

[0065] 1. Identification of gene sets that are responsive to immunotherapy based on spatial transcriptome data:

[0066] 1) Preprocessing of spatial transcriptome data:

[0067] In order to exclude the influence of other non-malignant cells on the results, the spots of the spatial transcriptome were first screened, and only those spots annotated as malignant tumor cells were retained. The immune infiltration areas in the local microenvironment of the tumor were identified for subsequent screening of differentially expressed genes. The screening results are as follows: Figure 1 As shown; among them, HCC01-HCC04 are samples that respond to immunotherapy, and HCC05-HCC07 are samples that do not respond to immunotherapy;

[0068] 2) Analysis of spatial transcriptome data:

[0069] The data of GSE238264 were normalized using the SCTransform method of the R package 'Seurat'. This method can not only standardize the data, but also detect highly variable genes and store the results in a specific assay (Spatial). The Harmony method was used to integrate the SCT data of 7 patient samples. This method effectively corrected the technical differences between samples and ensured the comparability of the data. The Harmony algorithm was used to unify the data of 4 samples that responded to immunotherapy and 3 samples that did not respond to immunotherapy into the same reference space (e.g. Figure 2 A);

[0070] On this basis, we further identified 566 genes with significant differences in expression between patients who responded to immunotherapy and those who did not respond to immunotherapy, namely the gene set of immunotherapy response based on spatial transcriptome (e.g. Figure 2 In the above immunotherapy response gene set, 534 (94.34%) genes were up-regulated in the response group, and only 32 (5.65%) genes were down-regulated.

[0071] Functional enrichment analysis was performed on the above 566 immunotherapy-responsive genes, and the results were as follows: Figure 2 C and Figure 2 As shown in D;

[0072] Depend on Figure 2 C shows that the up-regulated genes are enriched in biological processes such as leukocyte-mediated immune response, cell adhesion, and T cell activation; Figure 2 D shows that the down-regulated genes are significantly enriched in pathways such as humoral immune response and complement activation.

[0073] 2. Identify gene sets that respond to immunotherapy based on single-cell transcriptome data:

[0074] 1) Preprocessing of single-cell transcriptome data:

[0075] The R package 'Seurat' was used to process the data in 10X Genomics format. After the data were imported, quality control was performed, including detecting the UMI count, gene count, and proportion of mitochondrial genes for each cell to identify and remove low-quality cells;

[0076] Then, the NormalizeData function was used to normalize the data and perform global scaling normalization on the gene expression levels of each cell. The 'FindVariableFeatures' method was used to identify highly variable genes (HVGs), which is also an important step in single-cell preprocessing and helps to highlight biological signals in subsequent analysis.

[0077] Finally, use tSNE or UMAP method to perform linear or nonlinear dimensionality reduction on the data to visualize the results.

[0078] 2) Analysis of single-cell transcriptome data:

[0079] Since the H68 patient in the single-cell dataset GSE151530 responded to immunotherapy, only two samples of the H68 patient were retained, one from before treatment (H68a) and one from after treatment (H68b). The results after dimensionality reduction analysis are as follows: Figure 3 As shown in A;

[0080] By comparing H68a and H68b, 593 significantly differentially expressed genes were found, i.e., the gene set of immunotherapy response based on single-cell transcriptome; among them, 406 genes were upregulated after treatment, and 187 genes were downregulated after treatment;

[0081] GO functional enrichment analysis showed that the up-regulated genes were mainly enriched in the biological processes of terpenoid, fatty acid and alcohol metabolism (such as Figure 3 B), while the down-regulated genes were mainly enriched in biological processes such as cytoplasmic translation, ribosome biogenesis and ATP synthesis (such as Figure 3 C);

[0082] Comparing the gene set of immunotherapy response identified in the single-cell data with the gene set of immunotherapy response identified in the spatial transcriptome data, we found that 51 genes showed significant differential expression in both datasets (e.g. Figure 4 These common differentially expressed genes include SPP1 and MT1E. The differential expression of SPP1 and MT1E was verified in spatial transcriptome data and single-cell transcriptome data. The results are shown in Figure 2. Figure 4 B. Figure 5 and Figure 6 Among them, MT1E was the most significantly differentially expressed gene in the spatial transcriptome, while SPP1 was significantly upregulated in the malignant cells of the four responding samples in the spatial transcriptome.

[0083] 3. Screening of prognostic markers for liver cancer based on gene sets that respond to immunotherapy:

[0084] The GSE14520 dataset involved in this example contains 247 cancer tissue samples and 239 adjacent normal samples;

[0085] This example uses the PCA method to show the structure of the above data set (such as Figure 7 shown), from Figure 7The displayed results show that there is a clear separation between cancer tissue and adjacent normal samples, indicating that the data in the dataset is representative.

[0086] 1) Data preprocessing: remove samples of benign liver diseases (such as cirrhosis) and only keep cancer tissue samples.

[0087] 2) LASSO regression analysis to screen prognostic markers for liver cancer:

[0088] In order to screen suitable genes from the immune response-related gene set as liver cancer prognosis markers, this example uses the LASSO algorithm for feature selection.

[0089] In this example, the R software package glmnet was used to implement LASSO analysis. The aforementioned 51 immunotherapy response gene sets were screened based on the GSE14520 dataset. The family parameter in the glmnet method was set to "cox" to meet the needs of survival analysis. After each cox regression model was constructed and the optimal lambda value was selected through cross-validation, a set of key genes related to prognosis (the coefficient was not 0) was obtained. In order to ensure the robustness of the genes screened by the model, the above process was repeated 100 times, and then the frequency of genes appearing in the 100 models was counted, and the genes with the highest frequency were selected. The LASSO regression coefficient changes with the penalty parameter as shown in Figure 8 A; the result of cross-validation to select the optimal model is shown in Figure 8 As shown in B; the top 12 genes that stably appeared in 100 repeated modeling are as follows Figure 8 As shown in C; through this process, five key genes TXNRD1, HAO1, C1S, OPTN and SNX6 that play an important role in predicting the prognosis of HCC patients were successfully identified in the spatial transcriptome data GSE238264 and the single-cell transcriptome data GSE151530, namely the liver cancer prognosis markers provided by the present invention.

[0090] Example 2 Validation of liver cancer prognostic markers

[0091] This example uses the prognostic data from the TCGA and ICGC databases to verify the liver cancer prognostic markers provided in Example 1. The data information used is shown in Table 2.

[0092] Table 2 Dataset information used for validation of liver cancer prognostic markers

[0093]

[0094] 1. Validation of liver cancer prognostic markers in TCGA data:

[0095] The TCGA HCC dataset provides very comprehensive clinical characteristics information for patients, which helps to study the relationship between genes and clinical characteristics. This example analyzes the relationship between the expression level of liver cancer prognostic markers provided in Example 1 and the status (dead or alive) of HCC patients, T / N / M stage and pathological stage based on TCGA data. The results are as follows: Fig. 9 and Fig.10 shown.

[0096] Depend on Fig. 9 It can be seen that although some analysis results were not statistically significant (p>0.05), TXNRD1, OPTN and SNX6 were positively correlated with survival status, T / N / M stage and pathological stage; while HAO1 and C1S were negatively correlated with survival status, T / N / M stage and pathological stage. The above results directly or indirectly indicate the availability and accuracy of the liver cancer prognostic markers provided in Example 1.

[0097] Depend on Fig.10 It can be seen that although these five genes have no statistical significance with some clinical characteristics, they have a correlation trend, that is, the expression levels of TXNRD1, OPTN and SNX6 are positively correlated with the pathological stage, while the expression levels of OPTN and SNX6 are negatively correlated with the pathological stage. The above results directly or indirectly indicate the availability and accuracy of the liver cancer prognostic markers provided in Example 1.

[0098] 2. Correlation analysis between prognostic markers of liver cancer and immune microenvironment:

[0099] The ESTIMATE algorithm can provide researchers with scores for tumor purity, stromal cell levels, and immune cell infiltration levels in tumor tissues based on expression data. This example analyzes the correlation between TXNRD1, HAO1, C1S, OPTN, and SNX6 and the Stromal score, Immune score, and ESTIMATE score. The results are as follows: Fig.11 As shown in A.

[0100] Depend on Fig.11 A shows that TXNRD1, SNY6 and OPTN are significantly positively correlated with Stromal score, Immune score and ESTIMATE score, while HAO1 and C1S are negatively correlated. The above results indicate that these genes may play an important role in regulating the infiltration of stromal cells and immune cells in the tumor microenvironment.

[0101] Subsequently, this example evaluated the relationship between TXNRD1, HAO1, C1S, OPTN and SNX6 and 24 immunotherapy response-related signatures (the 24 immunotherapy response-related signature data are from reference Thorsson, Vésteinn et al. "The Immune Landscape of Cancer." Immunity vol. 48, 4 (2018): 812-830.e14.doi: 10.1016 / j.immuni.2018.03.023). The results are as follows: Fig.11 As shown in B.

[0102] Depend on Fig.11 B shows that the expression of TXNRD1, OPTN and SNX6 is significantly positively correlated with these signature scores, while HAO1 and C1S are significantly negatively correlated. The above results reveal that these genes may play an important role in regulating the tumor immune microenvironment and affecting the effect of immunotherapy.

[0103] Example 3 Establishment of a liver cancer prognosis prediction model

[0104] This example establishes a liver cancer prognosis prediction model based on the liver cancer prognosis markers and the data set GSE14520 provided in Example 1. The specific method is as follows:

[0105] In this example, a Cox proportional hazard regression model was constructed based on the liver cancer prognostic biomarkers screened in Example 1; the R software package glmnet was used to implement LASSO analysis, and the cv.glmnet function was used for cross-validation, which automatically performed LASSO regression and fitted multiple models on a series of regularization parameters λ; the performance of the model under different λ values ​​can be evaluated through cross-validation, and the λ that minimizes the mean square error (MSE) is selected as the optimal regularization parameter; finally, after determining the optimal λ value, the glmnet function was used again, this time specifying the optimal λ value obtained by cross-validation to obtain the final LASSO regression model; the risk score calculation formula of the obtained liver cancer prognosis prediction model is as follows:

[0106] HCC prognostic risk score = 0.0146 × TXNRD1 expression - 0.0014 × C1S expression + 0.0114 × OPTN expression + 0.0304 × SNX6 expression - 0.0019 × HAO1 expression.

[0107] When the risk score is greater than or equal to 0, it is judged as high risk, that is, poor prognosis; when the risk score is less than 0, it is judged as low risk, that is, good prognosis. The expression levels of the above five genes are all data after global normalization by the NormalizeData function.

[0108] Example 4 Validation of liver cancer prognosis prediction model

[0109] Based on the liver cancer prognosis prediction model provided in Example 3, this example uses the dataset GSE14520 and clinical data to verify it. The information of the data used is shown in Table 3.

[0110] Table 3 Data used to validate the liver cancer risk assessment model

[0111]

[0112] In order to evaluate the predictive performance of the model, this example calculated the C-index and drew a calibration curve. In addition, the patients were divided into two groups, high-risk and low-risk, and the Kaplan-Meier survival curve and log-rank test were used to compare the survival of the two groups of patients. The results are shown in Figure 2. Fig.12 shown.

[0113] Depend on Fig.12 It can be seen that the liver cancer prognosis prediction model provided in Example 3 has a C-index of 0.68 [9% CI: 0.63-0.73, p < 0.001] in the TCGA data set, indicating that it has good predictive ability; the patients were divided into high-risk and low-risk groups, and the Kaplan-Meier survival analysis showed that there was a significant difference in the survival curves of the two groups of patients (p < 0.001); further, this embodiment was independently verified on the ICGC data set, and the C-index of the model on the ICGC data set was 0.70 [95% CI: 0.62-0. 78, p<0.001], and there were also significant differences in the survival curves of patients in the high-risk and low-risk groups (p=0.0043); this indicates that the prognostic prediction model constructed by the present invention has good generalization ability; the C-index of the prognostic prediction model on GSE14520 is 0.62 [95% CI: 0.56-0.68, p<0.001], and there were also significant differences in the survival curves of patients in the high-risk and low-risk groups (p=0.047); the survival curves of patients in the high-risk and low-risk groups of the prognostic prediction model in GSE76427 also showed significant differences (p=0.013).

[0114] Fig.12The results further confirmed that the liver cancer patient prognosis prediction model constructed in Example 3 has good stability and universality. Through the verification of three data sets, the liver cancer patient prognosis prediction model provided by the present invention shows the potential to predict the survival risk of HCC patients in different populations, providing a valuable reference for clinical decision-making and individualized treatment. This example also verifies the practical significance of this diagnostic model in clinical practice. The risk score calculated based on the liver cancer patient prognosis prediction model is analyzed with the clinical characteristics of different patients from the TCGA database. The results are as follows: Fig.13 As shown. Fig.13 It can be seen that the risk score calculated by the prognostic prediction model for liver cancer patients was significantly increased in deceased HCC patients (dead group), and the prognostic risk score was also positively correlated with T stage, AJCC Stage and grade.

[0115] In order to further evaluate the prediction accuracy of the model, a calibration curve was drawn in this embodiment, and the results obtained are as follows Fig.14 As shown. Fig.14 It can be seen that in the three data sets of TCGA, ICGC and GSE14520, the predicted 1-year survival rate is well consistent with the actually observed survival rate, and the consistency between 2-year and 3-year survival rates also has a good trend, indicating that the prognostic model provided by the present invention has high accuracy in short-term prediction.

[0116] This example also analyzes the AUC values ​​of the liver cancer prognosis prediction model at 1 year, 3 years, and 5 years based on the data set GSE76427. The results are as follows: Fig.15 As shown. Fig.15 It can be seen that the liver cancer prognosis prediction model provided by the present invention has an AUC greater than 0.5 at 1 year, 3 years and 5 years, showing excellent classification performance.

[0117] This example also analyzes the expression differences of 5 liver cancer prognostic markers in high-risk and low-risk groups. The results are as follows: Fig.16 As shown. Fig.16 It can be seen that TXNRD1, OPTN and SNX6 were significantly upregulated in the high-risk group (the expression level of the SNX6 gene was missing in the GSE14520 dataset), while HAO1 and C1S were significantly downregulated in the high-risk group. These differences in expression patterns may reflect the different roles of these genes in the progression of HCC, providing clues for further understanding the molecular mechanisms of liver cancer.

[0118] Finally, this example also collected 80 liver cancer surgical samples from Shenzhen People's Hospital for experimental verification, extracted total RNA from the liver cancer samples and performed subsequent reverse transcription and qPCR detection, and performed model verification based on the expression levels of TXNRD1, HAO1, C1S, OPTN and SNX6 and the liver cancer prognosis evaluation model calculation formula provided in Example 3. The results are as follows: Fig.17 shown by Fig.17 It can be seen that the liver cancer prognosis evaluation model constructed by the present invention has a good evaluation effect on liver cancer (the high-risk group indicates a poor prognosis, P=0.0081); the combination of entity experiments and big data analysis further verifies the reliability of this prediction model and has certain clinical value.

[0119] Example 5 A liver cancer risk assessment system

[0120] This embodiment provides a liver cancer patient prognosis prediction system, which specifically includes the following four units:

[0121] Input unit: used to input the data to be evaluated;

[0122] An analysis unit: analyzing the data to be evaluated in the input unit based on the liver cancer patient prognosis prediction model established in Example 3;

[0123] Evaluation unit: used to display the risk score of the prognostic prediction model for liver cancer patients;

[0124] When the risk score is greater than or equal to 0, it is determined as a high risk, and when the risk score is less than 0, it is determined as a low risk.

[0125] Embodiment 6 An electronic device

[0126] This embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor;

[0127] When the processor executes the computer program, the method for establishing the prognosis prediction model for liver cancer patients provided in Example 3 is implemented.

[0128] Embodiment 7 A storage medium

[0129] This embodiment provides a storage medium storing instructions executable by a processor. When the instructions executable by the processor are executed by the processor, they are used to execute the method for establishing a prognosis prediction model for liver cancer patients provided in Example 3.

[0130] In summary, the present invention focuses on multi-omics data related to immunotherapy, discovers a group of new immunotherapy-related genes, and constructs a prognosis prediction model for liver cancer patients. These findings have deepened the understanding of the response mechanism of liver cancer immunotherapy, and provided new ideas for screening patients suitable for immunotherapy and predicting efficacy, which is expected to promote the precision of liver cancer immunotherapy.

[0131] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. Various changes can be made within the knowledge of ordinary technicians in the relevant technical field without departing from the purpose of the present invention. In addition, the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

Claims

1. A liver cancer biomarker, characterized in that: The liver cancer biomarkers include TXNRD1, HAO1, C1S, OPTN and SNX6.

2. A method for constructing a prognosis prediction model for liver cancer patients, characterized in that: The construction method comprises the following steps: S1: Obtain a gene expression dataset of cancer tissues from patients with liver cancer who have received immunotherapy, and screen for differentially expressed genes between patients who respond to and do not respond to immunotherapy; S2: Obtain a gene expression dataset of cancer tissues of patients with liver cancer who responded to immunotherapy before and after treatment, and screen for differentially expressed genes between cancer tissues before and after immunotherapy; S3: comparing the differentially expressed genes obtained in step S1 and step S2 to obtain genes that are differentially expressed in both sets of gene expression data sets, i.e., immune response-related genes; S4: obtaining a gene expression dataset of cancer tissues of liver cancer patients, performing univariate and multivariate logistic regression analysis on the immune response-related genes obtained in step S3 in the dataset to determine differentially expressed genes associated with the prognosis of liver cancer patients, and establishing a risk model; The differentially expressed genes associated with liver cancer include TXNRD1, HAO1, C1S, OPTN and SNX6.

3. The construction method according to claim 2, characterized in that: The gene expression dataset of cancer tissues of liver cancer patients after immunotherapy described in step S1 includes a spatial transcriptome dataset.

4. The construction method according to claim 3, characterized in that: The immunotherapy described in step S1 and step S2 is immune checkpoint inhibitor therapy.

5. The construction method according to claim 2, characterized in that: The calculation formula of the risk model is: risk score=0.0146×TXNRD1 expression level-0.0014×C1S expression level+0.0114×OPTN expression level+0.0304×SNX6 expression level-0.0019×HAO1 expression level.

6. The construction method according to claim 5, characterized in that: The TXNRD1 expression level, HAO1 expression level, C1S expression level, OPTN expression level and SNX6 expression level are all values ​​obtained by global scaling and normalization of gene expression levels using a function.

7. The construction method according to claim 6, characterized in that: When the risk score is greater than or equal to 0, it is determined as a high risk, and when the risk score is less than 0, it is determined as a low risk.

8. A liver cancer patient prognosis prediction system, characterized in that: The liver cancer patient prognosis prediction system comprises: Input unit: used to input the data to be evaluated; Analysis unit: a liver cancer patient prognosis prediction model constructed by the construction method according to any one of claims 2 to 7, used for analyzing the data to be evaluated in the input unit; Evaluation unit: used to display the risk score of the prognostic prediction model for liver cancer patients; When the risk score is greater than or equal to 0, it is determined as a high risk, and when the risk score is less than 0, it is determined as a low risk.

9. An electronic device, characterized in that: The electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the construction method as claimed in any one of claims 2 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium stores processor-executable instructions, and the processor-executable instructions are used to execute the construction method described in any one of claims 2 to 7 when executed by the processor.

Images