A biomarker for hepatocellular carcinoma typing and application thereof

Abstract

The application provides a biomarker for hepatocellular carcinoma typing and application thereof, and the biomarker is IL1B and HMGB2; the biomarker can be used for distinguishing a high proliferation subtype of hepatocellular carcinoma from a low proliferation subtype.

Description

Technical Field

[0001] This invention relates to the field of biotechnology, specifically to a biomarker for hepatocellular carcinoma typing and its application. Background Technology

[0002] Hepatocellular carcinoma (HCC) is a common malignant tumor that seriously threatens human health, and its high degree of tumor heterogeneity has become a core bottleneck restricting clinical efficacy. This heterogeneity is not only reflected in the differences in mutation profiles at the genomic level, but also in the spatiotemporal diversity of transcriptome, epigenetics, and metabolic pathways, making it difficult for patients to universally benefit from a single targeted therapy. Further complicating matters, during tumor evolution, the tumor microenvironment undergoes continuous dynamic remodeling—from immune surveillance to immune escape, from an inflammatory microenvironment to an immunosuppressive microenvironment. These changes profoundly affect patients' clinical outcomes and prognoses.

[0003] In current technologies, molecular subtyping biomarkers for hepatocellular carcinoma (HCC) are mainly derived from genomic or transcriptomic bulk sequencing data, neglecting the cellular composition heterogeneity and spatial distribution characteristics of the tumor microenvironment. Furthermore, existing subtyping systems often overlook the association between microenvironment subtypes and tumor cell functional states (such as proliferation, invasion, and metabolic reprogramming), resulting in an inability to effectively identify high-risk subgroups driving tumor malignant progression. In addition, although some studies have reported that certain inflammatory factors or proliferation-related genes are associated with HCC prognosis, there is a lack of specific biomarker combinations that can simultaneously reflect microenvironment characteristics and tumor malignancy, and personalized intervention strategies based on microenvironment subtypes have not been established.

[0004] Therefore, developing a molecular biomarker combination application model that considers the tumor microenvironment and spatial characteristics and can accurately identify high-risk subtypes is of significant clinical value and social necessity for overcoming current bottlenecks in liver cancer treatment, guiding personalized treatment, and prolonging overall patient survival. This technology will provide new theoretical basis and practical pathways for the precise subtyping and targeted therapy of hepatocellular carcinoma. Summary of the Invention

[0005] To address the problems existing in the background art, the present invention provides a biomarker for hepatocellular carcinoma subtyping and its application. This biomarker can be used to distinguish between high-proliferative and low-proliferative subtypes of hepatocellular carcinoma and can be used for prognostic prediction of hepatocellular carcinoma.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a biomarker for hepatocellular carcinoma subtyping, wherein the biomarker is IL1B and HMGB2; the biomarker can be used to distinguish between high-proliferative and low-proliferative subtypes of hepatocellular carcinoma.

[0007] Secondly, this invention provides the application of biomarkers for hepatocellular carcinoma (HCC) classification in identifying high-proliferative and low-proliferative subtypes of HCC. It obtains gene expression data from cancer tissues of HCC patients, uses IL1B and HMGB2 as gene sets to perform ssGSEA scoring, and classifies HCC samples into high-proliferative and low-proliferative subtypes based on the scores.

[0008] Thirdly, the present invention provides the application of reagents for detecting the expression levels of the above-mentioned biomarkers used for hepatocellular carcinoma subtyping in the preparation of hepatocellular carcinoma subtype diagnostic products.

[0009] Fourthly, this invention provides the application of reagents that downregulate the expression levels of IL1B and / or HMGB2 genes in liver tissue in the preparation of drugs for treating, improving, or alleviating hepatocellular carcinoma.

[0010] Fifthly, the present invention provides a method for constructing a hepatocellular carcinoma subtyping model, comprising the following steps: S1. Obtain gene expression data from cancer tissues of hepatocellular carcinoma patients and use IL1B and HMGB2 as gene sets for ssGSEA scoring; S2. Based on the ssGSEA score, hepatocellular carcinoma samples were divided into high-proliferative and low-proliferative subtypes.

[0011] According to the above scheme, the cut-off value of the ssGSEA score in step S2 is 1.09. Samples with a score > 1.09 are high-proliferation subtypes, and samples with a score ≤ 1.09 are low-proliferation subtypes.

[0012] Sixthly, the present invention provides a hepatocellular carcinoma classification model, which is constructed by the above-described method.

[0013] In a seventh aspect, the present invention provides a method for constructing a model for predicting the prognosis of hepatocellular carcinoma, comprising the following steps: Step 1. Obtain the expression levels of IL1B and HMGB2 genes in cancer tissues of hepatocellular carcinoma patients; Step 2. Based on the expression levels of IL1B and HMGB2 genes and prognostic data, construct a Cox regression model to calculate the patient's prognostic risk score.

[0014] Based on the above scheme, the specific model for predicting the prognosis of patients with hepatocellular carcinoma is as follows:

[0015]

[0016] Eighthly, the present invention provides a model for predicting the prognosis of hepatocellular carcinoma, which is constructed by the above-described method.

[0017] The beneficial effects of this invention are as follows: This invention identifies two key genes, IL1B and HMGB2, whose expression levels are significantly increased in the high-proliferative subtype of hepatocellular carcinoma. These two genes, IL1B and HMGB2, can serve as biomarkers for hepatocellular carcinoma subtyping, distinguishing between high-proliferative and low-proliferative subtypes of hepatocellular carcinoma, and can be used for prognosis prediction. At the same time, they provide potential targets for postoperative treatment and recurrence intervention in hepatocellular carcinoma patients. Attached Figure Description

[0018] Figure 1 The following are the results of hepatocellular carcinoma subtype identification in Example 1 of this invention: A: UMAP diagram of identified cell subtypes; B: The top rockfall diagram shows the relative changes in the area under the CDF curve when different k values ​​are selected in the consensus clustering analysis, and the bottom heatmap shows the clustering of the samples; C: The heatmap shows the relative abundance of tumor microenvironment cell subtypes under consensus clustering based on 130 hepatocellular carcinoma samples; D: The KM curve shows the association between specific subtype scores and overall survival of liver cancer patients (from publicly available data from TCGA-LIHC).

[0019] Figure 2 The following are the tumor status identification results for hepatocellular carcinoma subtypes in Example 2 of this invention: A: UMAP map of identified tumor cells and bile duct-like epithelial cells; B: Heatmap showing the average score of tumor-related features for each cancer cell subtype; C: Stacked plot showing the proportion of cancer cell subtypes in the four tumor microenvironment subtype groups; D: KM curve showing the association between specific cancer cell subtype scores and overall survival of liver cancer patients (from publicly available data from TCGA-LIHC).

[0020] Figure 3 The following are the results of biomarker identification for hepatocellular carcinoma subtypes in this invention: A: This comprehensive heatmap shows the results of NicheNet analysis on macrophages and proliferating cancer cells. The first part of the comprehensive heatmap shows the Pearson coefficient of macrophage ligands; a higher coefficient indicates that the ligand has a strong ability to regulate the Ca_TOP2A target gene. The second part shows the regulatory potential of the target gene. The third part shows the expression of the target gene in different cancer cell subtypes, where TF represents transcription factor. B: The dot plot shows the expression of key ligands in different cell types. C: The dot plot shows the expression of IL1B and HMGB2 in four tumor microenvironment subtype groups. D: The distribution of tumor-associated macrophages with high / low IL1B expression and cancer cells with high / low HMGB2 expression in three hepatocellular carcinoma samples. The heatmap shows the KL divergence changes of tumor-associated macrophages with high / low IL1B expression and cancer cells with high / low HMGB2 expression in three hepatocellular carcinoma samples; a high KL divergence indicates stronger spatial co-localization.

[0021] Figure 4 The results of identifying high-proliferative subtypes of hepatocellular carcinoma using the combined IL1B and HMGB2 model in this embodiment of the invention are as follows: A: Violin box plot shows that patients were subtyped using IL1B and HMGB2 combined with ssGSEA scores, with a cut-off value set at 1.0943; B: Violin box plot shows that the high-proliferative subtype has a significantly activated proliferation signal.

[0022] Figure 5 This is the performance evaluation result of the joint IL1B and HMGB2 model in Embodiment 5 of the present invention on the training set and validation set.

[0023] Figure 6 The following are in vivo and in vitro experimental results of the biomarkers of this invention: A: Comparison of subcutaneous tumor size between mice in the anti-IL-1β treatment group and the IgG control group; B: Tumor proliferation curves showing tumor growth in mice treated with anti-IL-1β or immunoglobulin G; C: Bar graphs showing the proportion of macrophages with high IL-1β expression in the anti-IL-1β treatment group and the IgG control group, with 7 samples in each group; D: Western blot results showing that increased IL-1β concentration can upregulate the expression level of HMGB2 in Hep3B / HuH-7 liver cancer cells; E: Comparison of subcutaneous tumor size in mouse models inoculated with wild-type and shHMGB2 tumor cells after treatment with PBS and IL-1β; F: Tumor proliferation curves showing the tumor growth rate in different groups. Detailed Implementation

[0024] The principles and features of the present invention are described below with reference to the accompanying drawings and specific embodiments. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0025] In this invention, the inventors integrated single-cell transcriptome data from approximately 179 human livers and performed omics data analysis to discover four subtypes of hepatocellular carcinoma microenvironments and identify IL1B. + Macrophages and HMGB2 + A highly proliferative subtype of hepatocellular carcinoma dominated by tumor cells. Through high-resolution cell interaction network analysis and multi-level experimental validation, IL1B and HMGB2 genes were established as dual characteristic biomarkers for this highly proliferative subtype, and their potential as therapeutic targets was also confirmed. This finding provides a new theoretical framework and practical pathway for the development of precise subtyping and targeted therapy strategies for hepatocellular carcinoma.

[0026] Example 1: Identification of Hepatocellular Carcinoma Subtypes Related to the Microenvironment Single-cell sequencing data from 179 liver samples across six public cohorts were collected and integrated. These samples included 130 hepatocellular carcinoma tumor tissue samples, 21 adjacent normal tissue samples, 17 hepatitis B samples, and 11 healthy liver control samples. Specifically, The c0 cohort data includes 82 hepatocellular carcinoma tumor tissue samples and 14 adjacent normal tissue samples. The original FASTQ data (BioProject ID: PRJCA007744) is stored at the China National Center for Bioinformation (CNCB) (https: / / www.cncb.ac.cn / ). The c1 cohort data was downloaded from the Gene Expression Omnibus (GEO) and contained 32 hepatocellular carcinoma (HCC) tumor tissue samples. The cohort data (GSE125449) for both batches c2_1 and c2_2 were downloaded from GEO and contained a total of 9 hepatocellular carcinoma (HCC) tumor tissue samples. The c3 cohort data (GSE115469) was downloaded from GEO and contains 5 healthy liver samples from deceased donors who were deemed suitable for liver transplantation. The c4 cohort data (CRA002308) comes from the Genome Sequence Archive in BIG DataCenter, Beijing Institute of Genomics (BIG), Chinese Academy of Sciences (https: / / bigd.big.ac.cn / gsa), and includes 7 hepatocellular carcinoma (HCC) tumor tissues and 7 paired adjacent normal tissue samples; The c5 cohort data (GSE182159) was downloaded from GEO and contains 6 healthy liver samples and 17 hepatitis B (HBV)-related hepatitis samples.

[0027] Samples containing only raw FASTQ data were filtered using CellRanger (v.3.0.1) and aligned with the human reference genome GRCh38. All samples were then quality-controlled using the R package Seurat, and cell filtering was performed according to the following criteria: (1) cells with 100-6000 detectable genes were retained; (2) the UMI count of each cell was less than 30,000; and (3) the proportion of mitochondrial genes was less than 50%. Next, the data were normalized and scaled. The top 2000 hypervariable genes were calculated for principal component dimensionality reduction, and 50 principal components were selected for subsequent cell clustering. The R package Harmony was used to remove batch effects between different cohorts. Cell clustering was performed using the FindNeighbors and FindClusters functions, with a resolution of 0.1 for large cell groups. The principal components and resolution settings for subsequent cell subpopulations were as follows: T and NK cells (30, 1.2); myeloid cells (10, 0.7); stromal cells (15, 0.8); B and plasma cells (10, 0.6); tumor and epithelial cells (15, 0.2). Finally, cells were visualized using uniform manifold approximation and projection (UMAP), and cell annotation was performed based on previous research and the CellMarker 2.0 database (http: / / 117.50.127.228 / CellMarker / index.html). A total of 1,003,004 high-quality cells, covering 58 cell types, were obtained. Figure 1 A). Consistent clustering of hepatocellular carcinoma tumor tissue samples ( Figure 1 B) identified four microenvironment-related subtypes ( Figure 1 C) Based on different infiltration ratios, cell types can be divided into: S1, a subtype dominated by macrophages; S2, a subtype dominated by B cells and plasma cells; S3, a subtype dominated by T cells and dendritic cells; and S4, a subtype dominated by stromal cells.

[0028] Specifically: First, all tumor cells were excluded, and the average gene expression of each cell subtype in each sample was calculated. Then, consensus clustering analysis was performed using the R package ConsensusClusterPlus, with 1000 resampling iterations and PAM as the clustering algorithm. Finally, based on the relative change in area under the CDF curve, k=4 was selected as the final subtype classification result. To further evaluate the association between different subtypes and overall survival of HCC patients, the top 10 specific genes for each subtype were first obtained using the R package COSG based on a cosine similarity algorithm. Subsequently, based on TCGA-LIHC cohort RNA-seq data, subtype-specific gene scoring was performed on patients using the R package GSVA. The optimal cut-off value was determined using the R packages survival and survminer, and survival analysis was performed on high and low score groups for different subtypes. KM curves were plotted, and a log-rank p-value less than 0.05 was considered a significant survival difference. The results showed that the S2 and S3 subtypes were often associated with better overall survival, and both subtypes had higher T-cell infiltration; the S4 subtype did not show a statistically significant difference; only the macrophage-dominated S1 subtype was associated with worse overall survival. Figure 1 D).

[0029] Example 2: Identification of Hepatocellular Carcinoma Subtype Tumor Status Tumor cell cluster identification ( Figure 2 A) All single-cell sequencing data underwent quality control in Seurat, using the following screening criteria: (1) cells with 100 to 6000 detectable genes were retained; (2) cells with a UMI count exceeding 30,000 were excluded; and (3) cells with a mitochondrial gene proportion exceeding 50% were excluded. The data were then normalized and scaled, and principal component analysis was performed using the top 2000 highly variable genes. Fifty principal components were selected for subsequent cluster analysis, and batch effects between different cohorts were corrected using the R package Harmony. Cell clustering was performed using the FindNeighbors and FindClusters functions, with 15 principal components for tumor cells and a resolution of 0.2. Visualization was performed using Uniform Manifold Approximation Projection (UMAP), and five tumor cell subtypes (Ca_01, Ca_02, Ca_03, Ca_04, and Ca_05) and bile duct-like epithelial cells expressing KRT19 (Epi_KRT19) were identified. Subsequently, a set of key tumor functional genes was downloaded from the CancerSEA database (http: / / biocc.hrbmu.edu.cn / CancerSEA / ), and the set of key tumor functional genes was scored. Figure 2B) revealed that Ca_O5 exhibited a highly proliferative and invasive state, indicating that it was an active tumor cell. By assessing the proportion of different tumor cell types in patients with four identified microenvironment subtypes, the results showed that highly proliferating Ca_O5 was most abundant in the S1 subtype. Figure 2 C). To assess the association between different tumor cell subpopulations and the survival of liver cancer patients, the top 15 specific genes for each subtype were first obtained using the COSG package in R based on a cosine similarity algorithm. Subsequently, based on TCGA-LIHC cohort RNA-seq data, the GSVA package in R was used to score patients for subtype-specific genes. The optimal cut-off value was determined using the survival and survminer packages in R. Survival analysis was performed on high and low score groups for different subtypes, and KM curves were plotted. A log-rank p-value less than 0.05 was considered a significant survival difference. The results showed that Ca_05 tumor cells were significantly associated with worse overall survival in patients ( Figure 2 D). These results indicate the existence of a macrophage-dominated S1 subtype of hepatocellular carcinoma, in which tumor cells are in a highly proliferative state, leading to a worse prognosis.

[0030] Example 3: Identification Results of Key Subtype-Related Biomarkers for Hepatocellular Carcinoma To explore the potential cell-cell communication between macrophages and proliferating tumor cells in promoting tumor cell proliferation, the R package nichenetr was used to infer cell-cell communication between macrophages and proliferating tumor cells. CaO5 cells were designated as signal receivers, and all macrophage subtypes were defined as signal senders. The top 30 ligands, receptors, and their target genes were then sorted and visualized according to Pearson scores. The results showed that PDGFB and IL1B in macrophages can activate the transcription factor HMGB2 in tumor cells (…). Figure 3 A). IL1B expression is more specific in macrophages, while PDGFB is excluded because it is mainly expressed in endothelial cells. Figure 3 B). Therefore, IL1B was identified as a key marker for macrophages in the S1 subtype, and HMGB2 was identified as a key marker for tumor cells in the S1 subtype. Furthermore, analysis of the overall expression levels of different subtypes revealed that IL1B and HMGB2 had the highest expression levels in the S1 subtype. Figure 3C). To assess the spatial distribution of different cell populations, the R package CellTrek was used for spatial embedding of single-cell data. First, a sparse graph was generated using a random forest model. Then, a spot-cell similarity matrix was constructed for each individual cell, adding spatial coordinate information. The scoloc function was used to summarize the co-localization patterns of different cell subpopulations, and the Kullback-Leibler divergence (KLD) was used to assess the strength of cell co-localization. Higher KLD values ​​indicate stronger co-localization of cell subclusters. Cells with different gene expression levels were grouped according to their median expression and evaluated using CellTrek. The results showed that macrophages with high IL1B and tumor cells with high HMGB2 exhibited significant spatial co-localization (…). Figure 3 (D) indicates that the spatial interaction between these two cell types may shape the S1 subtype, and IL1B and HMGB2 can serve as key biomarkers for identifying the highly proliferating S1 subtype of hepatocellular carcinoma.

[0031] Example 4: Identification of hyperproliferative subtypes of hepatocellular carcinoma using a combined IL1B and HMGB2 model. To evaluate the performance of IL1B and HMGB2 in identifying high-proliferative subtypes of hepatocellular carcinoma (HCC) patients, tissue sequencing data from the TCGA-LIHC cohort were used. The R package GSVA was used to score patients based on a combined IL1B and HMGB2 score and ssGSEA score. The cut-off value was set at 1.0943. A high score was assigned to the high-proliferative subtype, and a low score to the low-proliferative subtype. Figure 4 A).

[0032] The formula for calculating ssGSEA is as follows:

[0033]

[0034] In short, the expression values ​​of genes in a sample are sorted, and the enrichment score of the label S is calculated based on the difference between the empirical cumulative distribution function (ECDF) of the genes corresponding to the S label (IL1B and HMGB2) and all genes. By traversing the sorted gene list, an accumulated statistic is added when a gene belonging to the S label is encountered, and the accumulated statistic is decremented when a gene not belonging to the S label is encountered, thus calculating the ECDF difference (ΔECDF) between the label S and the remaining genes. The ES is calculated by the area under the ΔECDF curve, where p is the maximum rank of the genes in the sample. Finally, the genes in the sample are randomly permuted to calculate the normalized enrichment score (NES), with n set to 1000.

[0035] To further validate the grouping results, patients were scored using the key cancer consensus proliferation signals HALLMARK_E2F_TARGETS and HALLMARK_G2M_CHECKPOINT. Figure 4 B), the results showed that the high-density group had a significantly enhanced proliferation signal (B). P = 6.2e-16 and P <2.2e-16).

[0036] Example 5: Performance Evaluation Results of the Joint Model of IL1B and HMGB2 for Prognostic Prediction To evaluate the combined prognostic predictive performance of IL1B and HMGB2 in patients with hepatocellular carcinoma, we constructed a Cox regression model for IL1B and HMGB2 and calculated a combined IL1B and HMGB2 gene expression score based on tissue sequencing data from patients using the model coefficients. The score calculation formula is as follows:

[0037]

[0038] The expression levels of IL1B and HMGB2 were normalized using TPM. Further analysis using KM curves revealed that in hepatocellular carcinoma patients in the TCGA-LIHC cohort, divided into training and validation sets at a 5:5 ratio, the combined score of IL1B and HMGB2 (cut-off value 3.5152) showed significantly better overall survival prediction performance than either IL1B or HMGB2 expression alone (training set: P = 0.0011; Validation set: P = 0.011)( Figure 5 Patients with a combined score higher than the cut-off value had worse overall survival than those with a score lower than the cut-off value, indicating that the combination of IL1B and HMGB2 biomarkers can effectively indicate the prognosis of patients with liver cancer.

[0039] Example 6: Results of in vivo and in vitro experiments The role of anti-IL1B therapy in inhibiting tumor growth was investigated in a mouse experiment. Mice were divided into two groups (n=7 in each group): both groups underwent subcutaneous tumor formation using the mouse hepatocellular carcinoma cell line Hep1-6; the control group received intraperitoneal injection of IgG antibody, while the experimental group received intraperitoneal injection of anti-IL1B antibody. The administration frequency was once every two days for 17 days. The results showed that the tumors in the anti-IL1B antibody group were significantly smaller and grew more slowly. Figure 5 (A and 5B). Furthermore, flow cytometry analysis of dissected mouse tumors revealed IL1B levels in the anti-IL1B antibody group. + Macrophages were significantly inhibited ( Figure 5C). Cellular experiments revealed that the addition of recombinant IL1B protein to the human hepatocellular carcinoma cell line Hep3B significantly increased HMGB2 expression, indicating that IL1B can directly upregulate HMGB2 expression. Figure 5 D). Further experiments on nude mice investigated the effect of HMGB2 knockdown on the tumor-promoting effect of IL1B. Nude mice were divided into four groups (n=6 per group): groups one and two underwent subcutaneous tumor formation using the human hepatocellular carcinoma cell line Hep3B, while groups three and four underwent subcutaneous tumor formation using the HMGB2-knockdown Hep3B cell line. In addition, groups one and four received peritumoral injections of PBS, while groups two and three received peritumoral injections of recombinant IL1B protein. Treatment was administered every two days for 31 days. Results showed that IL1B significantly promoted tumor growth, and HMGB2 knockdown inhibited the tumor-promoting efficiency of IL1B (…). Figure 5 E).

[0040] The results indicate that IL1B and HMGB2 play important tumor-promoting roles in the proliferation of hepatocellular carcinoma. Anti-IL1B treatment or HMGB2 knockout can significantly reduce the tumor growth rate, suggesting that IL1B and HMGB2 are potential therapeutic targets for hepatocellular carcinoma.

[0041] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A biomarker for hepatocellular carcinoma subtyping, characterized in that, The biomarkers are IL1B and HMGB2; The biomarkers can be used to distinguish between high-proliferative and low-proliferative subtypes of hepatocellular carcinoma.

2. The application of the biomarker for hepatocellular carcinoma classification as described in claim 1 in identifying high-proliferative and low-proliferative subtypes of hepatocellular carcinoma, characterized in that, Gene expression data of cancer tissues from hepatocellular carcinoma patients were obtained. The ssGSEA score was performed using a gene set consisting of IL1B and HMGB2. Based on the score, hepatocellular carcinoma samples were divided into high-proliferation subtypes and low-proliferation subtypes.

3. The application of the reagent for detecting the expression level of the biomarker for hepatocellular carcinoma subtyping as described in claim 1 in the preparation of hepatocellular carcinoma subtype diagnostic products.

4. Application of reagents that downregulate the expression levels of IL1B and HMGB2 genes in liver tissue in the preparation of drugs for the treatment, improvement or relief of hepatocellular carcinoma.

5. A method for constructing a hepatocellular carcinoma classification model, characterized in that, Includes the following steps: S1. Obtain gene expression data from cancer tissues of hepatocellular carcinoma patients and use IL1B and HMGB2 as gene sets for ssGSEA scoring; S2. Based on the ssGSEA score, hepatocellular carcinoma samples were divided into high-proliferative and low-proliferative subtypes.

6. The method for constructing a hepatocellular carcinoma subtyping model according to claim 5, characterized in that, In step S2, the cut-off value of the ssGSEA score is 1.

09. Samples with a score > 1.09 are high-proliferation subtypes, and samples with a score ≤ 1.09 are low-proliferation subtypes.

7. A hepatocellular carcinoma classification model, characterized in that, It is constructed by the construction method described in claim 5 or 6.

8. A method for constructing a model to predict the prognosis of hepatocellular carcinoma, characterized in that, Includes the following steps: Step 1. Obtain the expression levels of IL1B and HMGB2 genes in cancer tissues of hepatocellular carcinoma patients; Step 2. Based on the expression levels of IL1B and HMGB2 genes and prognostic data, construct a Cox regression model to calculate the patient's prognostic risk score.

9. The method for constructing a prognostic model for hepatocellular carcinoma according to claim 8, characterized in that, The specific model for predicting the prognosis of patients with hepatocellular carcinoma is as follows: 。 10. A model for predicting the prognosis of hepatocellular carcinoma, characterized in that, It is constructed by the construction method described in claim 8 or 9.

Images