Biomarker for diagnosing hepatocellular carcinoma and application thereof

Abstract

The invention belongs to the technical field of medical detection, and discloses a biomarker for diagnosing hepatocellular carcinoma and application of the biomarker. The biomarker comprises a combination of a coding nucleic acid of alpha-1-microglobulin, a coding nucleic acid of vitronectin, a coding nucleic acid of apolipoprotein A1, a coding nucleic acid of a fibrinogen gamma chain, a coding nucleic acid of a fibrinogen alpha chain and a coding nucleic acid of albumin; the coding nucleic acid comprises mRNA or cDNA. According to the invention, CD147 positive extracellular vesicles are purified from plasma, PCR detection is carried out on mRNA of an internally loaded hepatocellular carcinoma biomarker, and an HCC Score diagnosis model constructed by taking a detection result of the biomarker as a characteristic has relatively high diagnosis sensitivity and specificity on hepatocellular carcinoma, has great potential in early screening of liver cancer, and can be used for diagnosis of hepatocellular carcinoma. And the kit has important practical significance on early diagnosis of hepatocellular carcinoma.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical detection, and relates to a biomarker for diagnosing hepatocellular carcinoma and application thereof. BACKGROUND

[0002] Hepatocellular carcinoma accounts for 75-85% of primary liver cancer, which is an invasive tumor and has a relatively occult onset. Less than 30% of liver cancer patients are suitable for radical treatment at the first diagnosis, and most patients are in the middle and advanced stages of liver cancer at the time of diagnosis. Middle and advanced hepatocellular carcinoma is mainly treated by radiotherapy, chemotherapy and targeted drugs, but its clinical efficacy still needs to be improved. At present, the 5-year survival rate of liver cancer patients in China is only 14.1%, one of the reasons is that liver cancer cells are prone to multi-drug resistance to chemotherapy drugs and targeted drugs, ultimately leading to poor prognosis of patients. Therefore, early detection of hepatocellular carcinoma is crucial to improve the survival rate of patients. The current clinical diagnostic methods mainly include imaging and serum marker detection, and the diagnostic effect is not satisfactory, especially in the early stage.

[0003] Therefore, exploring sensitive and specific early diagnosis strategies for hepatocellular carcinoma has become the focus of current research. In recent years, extracellular vesicles as a nanoscale double-membrane vesicle secreted by cells have gradually become a promising diagnostic tool in liquid biopsy due to their high stability, rich biological markers such as proteins, lipids and nucleic acids, and the ability to reflect the pathological state of the source cells. Studies have found that the expression profile of extracellular vesicles in the body fluid of hepatocellular carcinoma patients is significantly different from that of healthy people, especially the combination of extracellular vesicle membrane proteins and nucleic acids shows good performance in distinguishing early hepatocellular carcinoma patients from high-risk populations (such as chronic hepatitis B or cirrhosis patients). Therefore, based on the high-throughput detection platform of extracellular vesicle multi-omics markers, it is expected to realize the non-invasive, sensitive and early diagnosis of hepatocellular carcinoma and promote the application of precision medicine in the prevention and control of liver cancer.

[0004] CN112280857A discloses a biomarker for diagnosing hepatocellular carcinoma, which is NDUFB3 protein or mRNA thereof. By detecting the expression level of NDUFB3 protein or mRNA thereof in clinical hepatocellular carcinoma tissue samples, compared with the normal liver tissue adjacent to the cancer, the expression of NDUFB3 protein or mRNA thereof in hepatocellular carcinoma tissue samples is significantly lower, so NDUFB3 protein or mRNA thereof can be used as a basis for early clinical diagnosis of hepatocellular carcinoma.

[0005] In summary, developing new biomarkers related to early hepatocellular carcinoma and corresponding detection methods, and expanding the diagnostic techniques for early hepatocellular carcinoma, are of great significance for the early detection of hepatocellular carcinoma. SUMMARY

[0006] In view of the deficiencies of the prior art and actual needs, the present application provides a biomarker for diagnosing hepatocellular carcinoma and application thereof, which is obtained by integrating hepatocellular carcinoma tissue transcriptome information in a TCGA database, hepatocellular carcinoma tissue transcriptome information in a GEO database and hepatocellular carcinoma cell line transcriptome information in a CCLE database, and has high diagnostic sensitivity and specificity for hepatocellular carcinoma.

[0007] To achieve the object of the present application, the present application adopts the following technical solutions:

[0008] In a first aspect, the present application provides a biomarker for diagnosing hepatocellular carcinoma, which comprises a combination of coding nucleic acids of alpha-1-microglobulin, vitronectin, apolipoprotein A1, fibrinogen gamma chain, fibrinogen alpha chain and albumin; and the coding nucleic acids comprise mRNA or cDNA.

[0009] The present application is based on large-scale data integration for feature gene screening, and can mine hepatocellular carcinoma biomarkers more reliably and with stronger generalization ability than single-database transcriptome sequencing data. The present application purifies CD147-positive extracellular vesicles from plasma and performs PCR detection on the mRNA of the hepatocellular carcinoma biomarkers carried therein, and constructs a HCC Score diagnostic model based on the detection results of the biomarkers, which has high diagnostic sensitivity and specificity for hepatocellular carcinoma and has great potential in screening early-stage liver cancer and is of great practical significance for early diagnosis of hepatocellular carcinoma.

[0010] In a second aspect, the present application provides use of the biomarker for diagnosing hepatocellular carcinoma and / or its detection reagent in preparation of a hepatocellular carcinoma diagnostic product.

[0011] In a third aspect, the present application provides a kit for diagnosing hepatocellular carcinoma, which comprises reagents for detecting the presence or expression level of the biomarker for diagnosing hepatocellular carcinoma according to the first aspect.

[0012] Preferably, the reagents comprise primers and / or probes for detecting the biomarker for diagnosing hepatocellular carcinoma according to the first aspect.

[0013] Preferably, the kit further comprises a carrier of the biomarker for diagnosing hepatocellular carcinoma according to the first aspect and a carrier purification reagent.

[0014] Preferably, the carrier comprises CD147-positive extracellular vesicles.

[0015] Preferably, the carrier purification reagent comprises silica spheres or magnetic beads modified with antibodies or aptamers targeting CD147.

[0016] In a fourth aspect, the present invention provides a method for constructing a diagnostic model for hepatocellular carcinoma, the method comprising the following steps:

[0017] (1) Integration of samples from different databases: Integration of transcriptome sequencing data of multiple samples of hepatocellular carcinoma tissues and hepatocellular carcinoma cell lines, including data sets of hepatocellular carcinoma tissues and hepatocellular carcinoma cell lines;

[0018] (2) Biomarker screening for hepatocellular carcinoma: Based on the data set described in step (1), expression analysis is performed to obtain highly expressed genes in hepatocellular carcinoma tissues and hepatocellular carcinoma cell lines, and genes with low expression in immune cells are screened among the highly expressed genes to obtain the biomarkers described in the first aspect;

[0019] (3) Sample testing: using purification reagents to isolate CD147-positive extracellular vesicles from the subject's plasma, lyse the extracellular vesicles, extract RNA, and reverse transcribe it into cDNA. PCR is used to quantitatively detect the mRNA expression level of the biomarker described in the first aspect;

[0020] (4) Model construction: Binary logistic regression analysis was performed on the mRNA expression level data of hepatocellular carcinoma biomarkers in clinical samples of early hepatocellular carcinoma with cirrhosis and in cirrhosis samples to obtain a hepatocellular carcinoma diagnostic model. The hepatocellular carcinoma positive test was determined based on the model output variable, the hepatocellular score (HCC Score). The judgment standard was: when the HCC Score value was greater than -0.6, it was determined to be hepatocellular carcinoma positive.

[0021] In a fifth aspect, the present invention provides a hepatocellular carcinoma diagnostic model, which is constructed by the method for constructing a hepatocellular carcinoma diagnostic model described in the fourth aspect.

[0022] Preferably, the input variable of the hepatocellular carcinoma diagnostic model is the mRNA or cDNA expression level of the biomarker for diagnosing hepatocellular carcinoma described in the first aspect, and the output variable of the hepatocellular carcinoma diagnostic model is the subject's hepatocellular score HCC Score.

[0023] Preferably, the calculation formula for the subject's hepatocyte score is as shown in formula (1):

[0024] HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1).

[0025] wherein, the HCC Score is a hepatocyte score of the subject, the AMBP is an expression level of alpha-1-microglobulin encoding nucleic acid, the ALB is an expression level of albumin encoding nucleic acid, the VTN is an expression level of vitronectin encoding nucleic acid, the APOA1 is an expression level of apolipoprotein A1 encoding nucleic acid, the FGG is an expression level of fibrinogen gamma chain encoding nucleic acid, and the FGA is an expression level of fibrinogen alpha chain encoding nucleic acid; the encoding nucleic acid includes mRNA or cDNA.

[0026] In the present application, the clinical detection result of the model reaches 0.9309 in AUC value, indicating that the model has extremely high discrimination ability. The sensitivity of the model is 85.71%, the specificity is 87.10%, and the overall diagnostic accuracy is 86.47%, which can effectively identify early liver cancer individuals and minimize misjudgment of liver cirrhosis patients. The performance of the model is still stable by using machine learning method combined with cross-validation strategy for model performance evaluation, the AUC value is 0.9287, the sensitivity is increased to 88.17%, the specificity is 84.42%, and the diagnostic accuracy is still maintained at 86.47%, which is highly consistent with the performance of the logistic regression model.

[0027] In a sixth aspect, the present application provides a hepatocellular carcinoma diagnosis device, which comprises a detection unit and an evaluation unit.

[0028] The detection unit is used to perform the following steps:

[0029] The CD147 positive extracellular vesicles are separated from the plasma of the subject by using a purification reagent, the RNA is extracted after the extracellular vesicles are lysed, and the RNA is reversely transcribed into cDNA, and the mRNA or cDNA expression level of the biomarker in the first aspect is quantitatively detected by using PCR.

[0030] The evaluation unit is used to perform the following steps:

[0031] The mRNA or cDNA expression level of the biomarker detected by the detection unit is input into the hepatocellular carcinoma diagnosis model in the fifth aspect, and whether it is positive for hepatocellular carcinoma is judged according to the hepatocyte score HCC Score, and the calculation formula of the hepatocyte score HCC Score is shown in formula (1):

[0032] HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1)

[0033] Wherein, HCC Score is the score of hepatocyte of the subject, AMBP is the expression level of alpha-1-microglobulin coding nucleic acid, ALB is the expression level of albumin coding nucleic acid, VTN is the expression level of vitronectin coding nucleic acid, APOA1 is the expression level of apolipoprotein A1 coding nucleic acid, FGG is the expression level of fibrinogen gamma chain coding nucleic acid, and FGA is the expression level of fibrinogen alpha chain coding nucleic acid; the coding nucleic acid includes mRNA or cDNA; the standard for judgment is that when the value of HCC Score is greater than -0.6, it is judged as positive for hepatocellular carcinoma.

[0034] Compared with the prior art, the present application has the following beneficial effects:

[0035] (1) The present application is based on large-scale data integration for feature gene screening, and mines hepatocellular carcinoma biomarkers, which is more reliable than single database transcriptome sequencing data and has stronger generalization ability.

[0036] (2) The present application further confirms the effectiveness of six hepatocyte metabolism related genes in CD147 positive extracellular vesicles as early hepatocellular carcinoma diagnosis biomarkers in large sample verification, and establishes a high performance prediction model based on logistic regression and machine learning algorithm, which provides a method reference for clinical diagnosis of hepatocellular carcinoma. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 It is a flowchart for establishing a diagnosis model;

[0038] Figure 2 It is a heat map of gene expression in 376 hepatocellular carcinoma tissues in TCGA database, and the top 100 ordering from ALB to LBP;

[0039] Figure 3 It is a Wayne diagram of candidate genes in four data sets in GEO database;

[0040] Figure 4 It is a cross comparison chart of candidate genes in seven hepatocellular carcinoma cell lines in CCLE database;

[0041] Figure 5Data graph for clinical sample detection, wherein A is a six-gene expression heat map of 93 liver cirrhosis samples and 77 early hepatocellular carcinoma samples; B is a box plot comparing the liver cirrhosis group and the hepatocellular carcinoma group according to the output HCC Score value; C is an ROC curve evaluation of the hepatocellular carcinoma diagnosis model according to the output HCC Score value, and the AUC value reaches 0.9309; D is a confusion matrix for verifying the effect of the hepatocellular carcinoma diagnosis model, the sensitivity of the model is 85.71%, the specificity is 87.10%, and the overall diagnostic accuracy is 86.47%; and E is a model performance evaluation using a machine learning method combined with a cross-validation strategy, the model performs stably, the AUC value is 0.9287, the sensitivity is improved to 88.17%, the specificity is 84.42%, and the diagnostic accuracy is still maintained at 86.47%. DETAILED DESCRIPTION

[0042] To further illustrate the technical means adopted by the present application and its effects, the present application will be further described below in conjunction with the embodiments and drawings. It can be understood that the specific embodiments described herein are only used to explain the present application, and not to limit the present application.

[0043] If a specific technology or condition is not specified in the embodiments, it is performed according to the technology or condition described in the literature in the art, or according to the product manual. If the reagent or instrument used is not specified by the manufacturer, it is a conventional product that can be commercially available through a regular channel.

[0044] The present application is based on the hepatocellular carcinoma tissue and hepatocellular carcinoma cell line transcriptome sequencing data available in the prior art, and the hepatocellular carcinoma specific marker screening is performed, and a diagnosis model is further constructed, and a flowchart is shown as Figure 1 .

[0045] In an embodiment of the present application, a kit for diagnosing hepatocellular carcinoma is provided, which comprises reagents for detecting the presence or expression level of the biomarkers for diagnosing hepatocellular carcinoma described in the present application, such as AMBP primers and probes (Hs00155697_m1), VTN primers and probes (Hs00169863_m1), APOA1 primers and probes (Hs00163641_m1), FGG primers and probes (Hs00241037_m1), FGA primers and probes (Hs00241027_m1), ALB primers and probes (Hs00609411_m1), and antibodies or aptamer-modified silica spheres or magnetic beads targeting CD147.

[0046] In an embodiment of the present application, a method for constructing a hepatocellular carcinoma diagnosis model is provided, which comprises the following steps:

[0047] (1) Different database sample integration: integrate hepatocellular carcinoma tissue and hepatocellular carcinoma cell line transcriptome sequencing sample data, which includes hepatocellular carcinoma tissue and hepatocellular carcinoma cell line data set;

[0048] (2) Biomarker screening of hepatocellular carcinoma: based on the data set of step (1), expression analysis is performed to obtain highly expressed genes in hepatocellular carcinoma tissue and hepatocellular carcinoma cell lines, and among the obtained highly expressed genes, genes with low expression in immune cells are screened to obtain a biomarker combination of alpha-1-microglobulin encoding nucleic acid, vitronectin encoding nucleic acid, apolipoprotein A1 encoding nucleic acid, fibrinogen gamma chain encoding nucleic acid, fibrinogen alpha chain encoding nucleic acid and albumin encoding nucleic acid;

[0049] (3) Sample detection: using a purification reagent to separate CD147 positive extracellular vesicles from the plasma of the subject, extracting RNA after lysing the extracellular vesicles, and reverse transcribing it into cDNA, and using PCR to quantitatively detect the mRNA expression levels of the alpha-1-microglobulin encoding nucleic acid, vitronectin encoding nucleic acid, apolipoprotein A1 encoding nucleic acid, fibrinogen gamma chain encoding nucleic acid, fibrinogen alpha chain encoding nucleic acid and albumin encoding nucleic acid;

[0050] (4) Model construction: performing binary logistic regression analysis on the mRNA expression level data of alpha-1-microglobulin encoding nucleic acid, vitronectin encoding nucleic acid, apolipoprotein A1 encoding nucleic acid, fibrinogen gamma chain encoding nucleic acid, fibrinogen alpha chain encoding nucleic acid and albumin encoding nucleic acid in early hepatocellular carcinoma with cirrhosis samples and cirrhosis samples in clinical samples to obtain a hepatocellular carcinoma diagnosis model, and according to the model output variable subject liver cell score HCC Score to determine whether it is hepatocellular carcinoma positive, the judgment standard is: when the HCC Score value is greater than -0.6, it is judged as hepatocellular carcinoma positive.

[0051] In a specific embodiment of the present application, a hepatocellular carcinoma diagnosis model is provided, which is constructed by the method for constructing a hepatocellular carcinoma diagnosis model.

[0052] Preferably, the input variable of the hepatocellular carcinoma diagnosis model is the expression level of the combination of alpha-1-microglobulin encoding nucleic acid, vitronectin encoding nucleic acid, apolipoprotein A1 encoding nucleic acid, fibrinogen gamma chain encoding nucleic acid, fibrinogen alpha chain encoding nucleic acid and albumin encoding nucleic acid, and the output variable of the hepatocellular carcinoma diagnosis model is the subject liver cell score HCC Score. The calculation formula of the subject liver cell score is shown in formula (1):

[0053] HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1).

[0054] wherein the HCC Score is a hepatocyte score of the subject, the AMBP is an expression level of an alpha-1-microglobulin encoding nucleic acid, the ALB is an expression level of an albumin encoding nucleic acid, the VTN is an expression level of a vitronectin encoding nucleic acid, the APOA1 is an expression level of an apolipoprotein A1 encoding nucleic acid, the FGG is an expression level of a fibrinogen gamma chain encoding nucleic acid, and the FGA is an expression level of a fibrinogen alpha chain encoding nucleic acid; the encoding nucleic acid includes mRNA or cDNA.

[0055] In one embodiment of the present application, a hepatocellular carcinoma diagnosis device is provided, which comprises a detection unit and an evaluation unit.

[0056] The detection unit is configured to perform the following steps:

[0057] The CD147 positive extracellular vesicles are separated from the plasma of the subject using a purification reagent, the RNA is extracted after the extracellular vesicles are lysed, and the RNA is reversely transcribed into cDNA, and the mRNA or cDNA expression levels of the alpha-1-microglobulin encoding nucleic acid, the vitronectin encoding nucleic acid, the apolipoprotein A1 encoding nucleic acid, the fibrinogen gamma chain encoding nucleic acid, the fibrinogen alpha chain encoding nucleic acid, and the albumin encoding nucleic acid are quantitatively detected using PCR.

[0058] The evaluation unit is configured to perform the following steps:

[0059] The mRNA or cDNA expression levels of the biomarkers detected by the detection unit are input into a hepatocellular carcinoma diagnosis model, and whether the subject is positive for hepatocellular carcinoma is determined according to the hepatocyte score HCC Score, and the calculation formula of the hepatocyte score HCC Score is shown in Formula (1):

[0060] HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1).

[0061] Among them, HCC Score is the subject's hepatocellular score, AMBP is the expression level of the nucleic acid encoding α-1-microglobulin, ALB is the expression level of the nucleic acid encoding albumin, VTN is the expression level of the nucleic acid encoding vitronectin, APOA1 is the expression level of the nucleic acid encoding apolipoprotein A1, FGG is the expression level of the nucleic acid encoding fibrinogen γ chain, and FGA is the expression level of the nucleic acid encoding fibrinogen α chain; the encoding nucleic acid includes mRNA or cDNA; the judgment standard is: when the HCC Score value is greater than -0.6, it is judged as positive for hepatocellular carcinoma.

[0062] Example 1

[0063] This example screens hepatocellular carcinoma-specific biomarkers.

[0064] Based on publicly available transcriptome sequencing data of hepatocellular carcinoma tissues and hepatocellular carcinoma cell lines: The tumor transcriptome data in the TCGA database includes 376 hepatocellular carcinoma tissues, which were obtained using the Illumina HiSeq 2000 RNA sequencing platform. Gene expression levels were converted to transcripts per million (TPM) through standardization. The genes with the highest expression in hepatocellular carcinoma tissues were screened, such as Figure 2 As shown in the figure, the top 100 genes with the highest expression levels were selected from ALB to LBP. Four high-quality expression profile datasets related to hepatocellular carcinoma were selected from the GEO database (GSE164760, GSE63898, GSE56140, GSE25097, etc.). These datasets all contain control information of hepatocellular carcinoma tissues and corresponding non-tumor tissues. The top 200 genes with the highest expression in hepatocellular carcinoma tissues were screened out. The "pairwise intersection" strategy was adopted, that is, the highly expressed gene sets screened out by each of the four datasets were intersected pairwise, and candidate genes that appeared in two or more datasets were extracted, as shown in the figure. Figure 3The intersection part of GEO and TCGA will be further analyzed, and the representative genes with stable high expression characteristics in hepatocellular carcinoma tissues will be optimized. In order to further verify the expression stability and representativeness of the candidate genes in hepatocellular carcinoma at the cellular level, the CCLE (Cancer Cell Line Encyclopedia) database is introduced to analyze the model of tumor cell lines derived from hepatocellular carcinoma. Seven representative hepatocellular carcinoma cell lines are screened from the CCLE database. These cell lines are widely used in basic and translational medical research related to hepatocellular carcinoma, covering different molecular characteristics and biological behaviors, and have good representativeness and experimental operability. The gene expression matrix data of the 7 hepatocellular carcinoma cell lines are extracted, and the gene expression of each cell line is sorted respectively to screen the top 200 highly expressed genes as a preliminary candidate gene set that may be enriched in expression at the cellular level. In order to improve the robustness of the data, the top 200 highly expressed genes screened from each of the 7 cell lines are cross-compared in a "two-by-two intersection" manner, and the genes commonly highly expressed in multiple cell lines are extracted as the final candidate genes, such as Figure 5The core purpose of this strategy is to identify metabolic-related genes that are stably expressed in multiple liver cancer cell models to exclude expression bias due to the special background of a single cell line. Based on the integration of TCGA, GEO and CCLE multi-platform, multi-level candidate gene screening results, the human white blood cell transcriptome dataset (such as from the Human Protein Atlas data) was further introduced, and the model analyzed the expression of candidate genes in major immune cell types (including T cells, B cells, monocytes, NK cells, granulocytes, etc.). For genes with obvious expression in white blood cells, strict exclusion was carried out to maximize the avoidance of false positive results or signal interference in subsequent plasma extracellular vesicle detection. After a series of bioinformatics processes of comprehensive screening and layer-by-layer exclusion: screening of metabolic-related genes highly expressed in hepatocellular carcinoma tissues in the TCGA database, taking the top 200 highly expressed differential genes in four independent data sets in GEO and intersecting each other, screening the top 200 genes expressed in 7 representative hepatocellular carcinoma cell lines in CCLE and intersecting each other, reverse screening of white blood cell expression profile data and removal of interfering genes, and only selecting liver cell metabolism-related genes. Finally, 6 candidate genes were determined, which were the coding nucleic acid of alpha-1-microglobulin, the coding nucleic acid of vitronectin, the coding nucleic acid of apolipoprotein A1, the coding nucleic acid of fibrinogen gamma chain, the coding nucleic acid of fibrinogen alpha chain and the coding nucleic acid of albumin (AMBP, VTN, APOA1, FGG, FGA and ALB), which were highly expressed in hepatocellular carcinoma tissues and cell models, and were lowly expressed or not expressed in peripheral white blood cells, with good specificity, stability and detectability.

[0065] Example 2

[0066] This example carries out the construction and evaluation of a hepatocellular carcinoma diagnosis model.

[0067] 1. Collection and preparation of clinical data and samples

[0068] A retrospective analysis was used to collect hepatocellular carcinoma with cirrhosis patient samples and cirrhosis patient samples.

[0069] The hepatocellular carcinoma patients met the following inclusion criteria: all were early cases that could be radically treated by surgery; the pathological diagnosis was clear about the tumor pathology (Edmondson) stage; the basic information of all cases was complete; all cases were accompanied by cirrhosis; the following cases were excluded: pregnant patients, reproductive embryonic tumors, combined with other organ malignancies, severe infectious diseases, and severe diseases of other important organs (such as heart, lung, kidney, etc.).

[0070] Liver cirrhosis group inclusion criteria (control group): clinically, imaging or liver biopsy confirmed diagnosis of liver cirrhosis, and no evidence of hepatocellular carcinoma or other liver tumors; the case basic information, clinical examination results and follow-up data are complete; no liver transplantation or other liver resection treatment within 6 months before enrollment; exclude the following cases: pregnant patients, reproductive embryonic tumors, combined with other organ malignancies, severe infectious diseases, and severe other important organ diseases (such as heart, lung, kidney, etc.).

[0071] 2. Isolation of CD147-positive extracellular vesicles from blood samples of subjects and RNA extraction

[0072] Collect 4 mL of peripheral venous blood from the subject in an EDTA vacuum anticoagulant blood collection tube. After sampling, invert the sampling tube 5 times, and treat at 4°C for 24 hours. To avoid contamination by epithelial cells, all blood samples are non-first tube peripheral blood. Place the blood collection tube in a horizontal centrifuge at 300 g for 15 min to obtain plasma and lower blood cell sediment; aspirate the upper plasma, centrifuge at 4°C for 15 min at 2500 g, and discard the sediment. Use the purification reagent to isolate CD147-positive extracellular vesicles from the subject's plasma, use QIAzol lysis solution to lyse the extracellular vesicles, and release the RNA; use the Qiagen miRNeasy Micro Kit to extract RNA and reverse transcribe the RNA into cDNA template by reverse transcription kit.

[0073] 3. Quantification of the 6 mRNA genes screened in Example 1 using ddPCR

[0074] According to the mRNA coding gene sequences of the target genes (AMBP, ALB, APOA1, FGG, FGA and VTN) screened in Example 1, order the target gene primers and probes (Thermo Fisher); detect the target gene mRNA molecules in the sample by QX200 Auto DG Droplet Digital PCR (Droplet Digital PCR, ddPCR) model, and obtain the mRNA expression of the target gene in each sample. The ddPCR reaction system volume is 20 μL, which is composed of ddPCR probe premix (without dUTP), enzyme-free sterile water, cDNA template, target gene primers and probes. After PCR amplification of the target gene cDNA template, the microdroplet reader (QX200 Droplet Reader) is used for quantitative detection of the target gene mRNAs in the sample (transcripts / μL), and the mRNA expression of the captured target gene in each sample is calculated.

[0075] Among them, the target gene primers and probes of nucleic acids are commercial products, as shown in Table 1 below:

[0076] Table 1

[0077]

[0078] 4. Construction and validation of an early diagnosis model for hepatocellular carcinoma

[0079] Based on gene expression data, draw gene expression heat map, such as Figure 5 As shown in Figure A, six genes—encoding nucleic acid for α-1-microglobulin, vitronectin, apolipoprotein A1, fibrinogen γ chain, fibrinogen α chain, and albumin—were significantly differentially expressed between the early-stage hepatocellular carcinoma group and the cirrhosis group, demonstrating good classification potential and consistent expression trends. Furthermore, a binary logistic regression model was constructed using gene expression data from these 170 samples for the diagnosis and prediction of early-stage hepatocellular carcinoma. The input variables include the mRNA expression levels of the nucleic acid encoding α-1-microglobulin, the nucleic acid encoding vitronectin, the nucleic acid encoding apolipoprotein A1, the nucleic acid encoding fibrinogen γ chain, the nucleic acid encoding fibrinogen α chain, and the nucleic acid encoding albumin. The output variable is the subject's score HCC Score, HCC Score = (0.255×AMBP) + (0.517×ALB) + (0.819×VTN) - (0.19×APOA1) - (0.435×FGG) + (0.684×FGA) - 4.534 Formula (1);

[0080] Wherein, HCC Score is the hepatocyte score of the subject, AMBP is the expression level of the nucleic acid encoding α-1-microglobulin, ALB is the expression level of the nucleic acid encoding albumin, VTN is the expression level of the nucleic acid encoding vitronectin, APOA1 is the expression level of the nucleic acid encoding apolipoprotein A1, FGG is the expression level of the nucleic acid encoding fibrinogen γ chain, and FGA is the expression level of the nucleic acid encoding fibrinogen α chain. Figure 5 As shown in Figure B. The diagnostic performance of this model is excellent, as shown in Figure 5 As shown in Figure C, the ROC curve based on the HCC Score value is used to evaluate the diagnosis model for hepatocellular carcinoma, and the AUC value reaches 0.9309. In the confusion matrix used to verify the effectiveness of the diagnosis model for hepatocellular carcinoma, the model has a sensitivity of 85.71%, a specificity of 87.10%, and an overall diagnostic accuracy of 86.47%. Figure 5As shown in the middle D figure, the early liver cancer individual can be effectively identified, and the misjudgment of the liver cirrhosis patient is minimized. In order to further verify the generalization ability and stability of the model, the model performance is evaluated by using the machine learning method combined with the cross-validation strategy, and the performance of the model is still stable, such as Figure 5 As shown in the middle E figure, the AUC value is 0.9287, the sensitivity is increased to 88.17%, the specificity is 84.42%, and the diagnostic accuracy is 86.47%.

[0081] In summary, the liver cancer biomarker is mined in the present application, and the liver cancer diagnosis model is further constructed by detecting clinical data and based on binary logistic regression and machine learning algorithm, so that the liver cancer can be accurately diagnosed from the liver cirrhosis control.

[0082] The detailed method of the present application is illustrated by the above embodiments, but the present application is not limited to the above detailed method, that is, it does not mean that the present application must rely on the above detailed method to be implemented. It should be understood by those skilled in the art that any improvement of the present application, equivalent replacement of each raw material of the product of the present application, addition of auxiliary ingredients, selection of specific methods, etc. fall within the protection scope and disclosure scope of the present application.

Claims

1. A biomarker for diagnosing hepatocellular carcinoma, characterized by, The biomarker comprises a combination of a coding nucleic acid of alpha-1-microglobulin, a coding nucleic acid of vitronectin, a coding nucleic acid of apolipoprotein A1, a coding nucleic acid of fibrinogen gamma chain, a coding nucleic acid of fibrinogen alpha chain and a coding nucleic acid of albumin; the coding nucleic acid comprises mRNA or cDNA.

2. Use of the biomarker and / or its detection reagent for diagnosing hepatocellular carcinoma in the preparation of a hepatocellular carcinoma diagnosis product.

3. A kit for the diagnosis of hepatocellular carcinoma, characterized by, The kit comprises a reagent for detecting the presence or expression level of the biomarker for diagnosing hepatocellular carcinoma according to claim 1.

4. The kit for diagnosis of hepatocellular carcinoma according to claim 3, characterized by, The reagent comprises primers and / or probes for detecting the biomarker for diagnosing hepatocellular carcinoma according to claim 1.

5. The kit for the diagnosis of hepatocellular carcinoma according to claim 3 or 4, characterized in that, The kit further comprises a carrier of the biomarker for diagnosing hepatocellular carcinoma according to claim 1 and a carrier purification reagent.

6. The kit for diagnosis of hepatocellular carcinoma according to claim 5, characterized by, The carrier comprises CD147-positive extracellular vesicles; the carrier purification reagent comprises an antibody or aptamer modified silica sphere or magnetic bead targeting CD147.

7. A method of constructing a diagnostic model for hepatocellular carcinoma, characterized by, The method comprises the following steps: (1) Different database sample integration: integrating hepatocellular carcinoma tissue and hepatocellular carcinoma cell line transcriptome sequencing sample data, which comprises a hepatocellular carcinoma tissue and hepatocellular carcinoma cell line data set; (2) Biomarker screening for hepatocellular carcinoma: based on the data set in step (1), performing expression analysis to obtain highly expressed genes in hepatocellular carcinoma tissue and hepatocellular carcinoma cell lines, and screening genes with low expression in immune cells from the obtained highly expressed genes to obtain the biomarker according to claim 1; (3) Sample detection: separating CD147-positive extracellular vesicles from the plasma of a subject using a purification reagent, extracting RNA after lysing the extracellular vesicles, and reverse transcribing the RNA into cDNA, and quantitatively detecting the mRNA expression level of the biomarker according to claim 1 using PCR; (4) Model construction: performing binary logistic regression analysis on the mRNA expression level data of the hepatocellular carcinoma biomarker in early hepatocellular carcinoma with cirrhosis samples and cirrhosis samples in clinical samples to obtain a hepatocellular carcinoma diagnosis model, and determining whether a subject is positive for hepatocellular carcinoma according to the model output variable, i.e., the subject's hepatocyte score (HCC Score), and the determination standard is that when the HCC Score value is greater than -0.6, the subject is determined to be positive for hepatocellular carcinoma.

8. A hepatocellular carcinoma diagnosis model, characterized by, The hepatocellular carcinoma diagnosis model is constructed by the method for constructing a hepatocellular carcinoma diagnosis model according to claim 7; The input variable of the hepatocellular carcinoma diagnosis model is the mRNA or cDNA expression level of the biomarker for diagnosing hepatocellular carcinoma according to claim 1, and the output variable of the hepatocellular carcinoma diagnosis model is the subject's hepatocyte score (HCC Score). 9.The hepatocellular carcinoma diagnosis model according to claim 8, characterized in that, The calculation formula of the subject's hepatocyte score is shown in formula (1): HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1) wherein, HCC Score is the hepatocyte score of the subject, AMBP is the expression level of alpha-1-microglobulin encoding nucleic acid, ALB is the expression level of albumin encoding nucleic acid, VTN is the expression level of vitronectin encoding nucleic acid, APOA1 is the expression level of apolipoprotein A1 encoding nucleic acid, FGG is the expression level of fibrinogen gamma chain encoding nucleic acid, and FGA is the expression level of fibrinogen alpha chain encoding nucleic acid; the encoding nucleic acid includes mRNA or cDNA.

10. A hepatocellular carcinoma diagnostic device, characterized by comprising: The hepatocellular carcinoma diagnosis device comprises a detection unit and an evaluation unit. The detection unit is used to perform the following steps: separating CD147 positive extracellular vesicles from the plasma of the subject by using a purification reagent, extracting RNA after lysing the extracellular vesicles, and reverse transcribing the RNA into cDNA, and quantitatively detecting the mRNA or cDNA expression level of the biomarker of claim 1 by using PCR; The evaluation unit is used to perform the following steps: inputting the biomarker mRNA or cDNA expression level detected by the detection unit into the hepatocellular carcinoma diagnosis model of claim 8 or 9, and judging whether it is positive for hepatocellular carcinoma according to the hepatocyte score HCC Score, and the calculation formula of the hepatocyte score HCC Score is shown in formula (1): HCC Score = (0.255 x AMBP) + (0.517 x ALB) + (0.819 x VTN) - (0.19 x APOA1) - (0.435 x FGG) + (0.684 x FGA) - 4.534 Formula (1) wherein, HCC Score is the hepatocyte score of the subject, AMBP is the expression level of alpha-1-microglobulin encoding nucleic acid, ALB is the expression level of albumin encoding nucleic acid, VTN is the expression level of vitronectin encoding nucleic acid, APOA1 is the expression level of apolipoprotein A1 encoding nucleic acid, FGG is the expression level of fibrinogen gamma chain encoding nucleic acid, and FGA is the expression level of fibrinogen alpha chain encoding nucleic acid; the encoding nucleic acid includes mRNA or cDNA; The standard for judgment is that when the HCC Score value is greater than -0.6, it is judged to be positive for hepatocellular carcinoma.

Images