Biomarker combination for liver cancer diagnosis and application thereof

Abstract

The invention provides a method and a kit for diagnosing or identifying whether an individual suffers from liver cancer or not by detecting a metabolic biomarker, and the metabolic biomarker comprises 1-stearoyl-2-arachidonyl-sn-glycerol, NG, S-glycerol, S-glycerol, N-glycerol, N-glycerol, N-glycerol, N-glycerol, N-glycerol, N-glycerol, N-glycerol, N-glycerol and N-glycerol. The compound is selected from one or more of N, N-dimethyl-L-arginine, N, N-dimethyl-L-arginine (ADMA), taurochenodeoxycholic acid, proline-valine (Pro-Val) and 1-oleoyl-L-alpha-lysophosphatidic acid.

Description

Invention Field

[0001] This application relates to the medical field. Specifically, this application provides a kit for diagnosing or identifying liver cancer in an individual and its related uses. Background of the Invention

[0002] Liver cancer is a malignant tumor that seriously endangers human health, originating from cellular mutations in the liver. It primarily originates from hepatocytes and intrahepatic bile duct cells, and its incidence and mortality rates are quite high globally, ranking as the fourth most common type of cancer after lung cancer, colorectal cancer, and stomach cancer. Even more seriously, malignant tumors from other parts of the body can metastasize to the liver, forming liver metastases, further complicating the treatment of liver cancer. The disease cycle of liver cancer is relatively long, ranging from several months to several years. During this process, the disease progresses gradually, and symptoms gradually worsen. However, liver cancer often has no obvious symptoms in its early stages, or only presents with some non-specific signs, such as mild abdominal pain and indigestion, making it difficult to detect in its early stages. Therefore, many patients are only diagnosed when the disease has progressed to an advanced stage, significantly impacting treatment effectiveness and prognosis.

[0003] Given the high incidence and mortality rates of liver cancer, as well as the difficulty in early detection, early detection and diagnosis are of paramount importance. Early detection and diagnosis allow us to provide patients with more timely and effective treatment, significantly improving treatment outcomes and survival rates. Therefore, we need to actively seek and develop new methods that can effectively predict or diagnose liver cancer.

[0004] While existing screening methods such as ultrasound, CT, and MRI are helpful in diagnosing liver cancer to some extent, they still have limitations, such as low sensitivity and specificity. Therefore, developing new biomarkers for liver cancer diagnosis or risk assessment is particularly important. These biomarkers may originate from specific genes, proteins, or metabolites of liver cancer cells, and they can reflect the occurrence, development, and metastasis of liver cancer, thus providing a more accurate and reliable basis for early diagnosis and risk assessment of liver cancer. Invention Overview

[0005] On one hand, this application provides a kit for diagnosing or identifying liver cancer in an individual, comprising reagents for detecting the levels of metabolites in a sample from said individual, said metabolites comprising one or more of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine (ADMA), taurine chenodeoxycholic acid, proline-valine (Pro-Val), and 1-oleoyl-L-α-lysophosphatidylcholine.

[0006] In some implementations, the liver cancer is hepatocellular carcinoma.

[0007] In some embodiments, the metabolite is selected from two or more of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidic acid.

[0008] In a preferred embodiment, the metabolite is a combination of 1-stearoyl-2-arachidonicyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophosphatidic acid.

[0009] In some embodiments, the sample is selected from serum, feces, plasma, or whole blood. In a preferred embodiment, the sample is serum or feces. In a more preferred embodiment, the sample is serum.

[0010] In some embodiments, the kit is a kit for performing LC-MS / MS analysis, which contains mass spectrometry standards for detecting the metabolites.

[0011] On the other hand, this application provides the use of reagents for detecting the level of metabolites in samples from an individual in the preparation of kits or drugs for diagnosing or identifying liver cancer in said individual, said metabolites comprising one or more of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophosphatidic acid.

[0012] In some implementations, if the level of the metabolite in the sample changes compared to the level in a control sample, it indicates that the individual has or may have liver cancer.

[0013] In a specific implementation, the metabolites are a combination of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline valine, and 1-oleoyl-L-α-lysophosphatidyl acid. If the levels of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, and taurine chenodeoxycholic acid in a sample from an individual are higher than those in a control sample, and the levels of proline valine and 1-oleoyl-L-α-lysophosphatidyl acid in a sample from the individual are lower than those in a control sample, then the individual has or may have liver cancer.

[0014] In a specific implementation, a value is calculated based on the following logistic regression formula and compared with a threshold. A value greater than the threshold indicates that the individual has or may have liver cancer: logit(P) = log(P / (1-P)) = 1.986 + 2.508 × taurine chenodeoxycholic acid concentration – 4.499 × 1-oleoyl-L-α-lysophosphatidyl acid concentration – 2.088 × NG,NG-dimethyl-L-arginine concentration – 1.762 × 1-stearoyl-2-arachidonicyl-sn-glycerol concentration + 0.191 × prolyl-valine concentration. In some implementations, the threshold is set to 0.68.

[0015] Furthermore, this application provides a method for detecting metabolite levels in a sample, comprising treating the sample to be tested with the reagents in the aforementioned kit. In a specific embodiment, the sample is selected from serum, plasma, whole blood, or feces, preferably serum or feces, more preferably serum. In a specific embodiment, the metabolite comprises one or more of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline valine, and 1-oleoyl-L-α-lysophosphatidic acid.

[0016] In some embodiments, the method for detecting the levels of the metabolites in a sample includes contacting the sample to be tested with a detection reagent and then performing LC-MS / MS analysis. In a preferred embodiment, the metabolites to be detected are a combination of five metabolites: 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine. Brief description of the attached figures

[0017] Figure 1 The study flowchart shows the collection of plasma and stool samples from liver cancer and healthy subjects, followed by untargeted metabolomics analysis. Subsequent analyses identified functional metabolites and diagnostic biomarkers for liver cancer.

[0018] Figure 2 The changes in serum metabolites in hepatocellular carcinoma are shown. (A) PLS-DA and OPLS-DA analyses were used to assess the distribution of serum metabolites in different subject groups. (B) Z-score heatmap of significantly altered plasma metabolites. Metabolites enriched in hepatocellular carcinoma are marked in red, and metabolites decreased in hepatocellular carcinoma are marked in blue.

[0019] Figure 3The changes in fecal metabolites in hepatocellular carcinoma (HCC) are shown. (A) PLS-DA and OPLS-DA analyses were used to assess the distribution of fecal metabolites in different subject groups. (B) Z-score heatmap of significantly altered fecal metabolites. Metabolites enriched in HCC are marked in red, and metabolites decreased in HCC are marked in blue.

[0020] Figure 4 The study revealed functional pathways of serum and fecal metabolites that were significantly altered in hepatocellular carcinoma (HCC). Functional pathways of serum metabolites (A) and fecal metabolites (B) that were significantly altered in HCC were identified through metabolite set enrichment overview analysis.

[0021] Figure 5 Functional validation of key metabolites in hepatocellular carcinoma (HCC) is shown. (AC) Cell viability assays (A), apoptosis assays (B), and cell cycle analysis (C) of two human HCC cell lines (Hep3B and Huh7) treated with γ-L-glutamyl-L-valine. (D) Representative images and cell viability assay results of organ samples from HCC patients treated with γ-L-glutamyl-L-valine. (EG) Cell viability assays (E), apoptosis assays (F), and cell cycle analysis (G) of two human HCC cell lines treated with butyl lactate. (H) Representative images and cell viability assay results of organ samples from HCC patients treated with butyl lactate.

[0022] Figure 6 The results of serum biomarker combinations in diagnosing hepatocellular carcinoma (HCC) are presented. (A) The results of serum biomarker combinations in diagnosing HCC in the discovery and validation cohorts were analyzed using a random forest analysis of non-target metabolites. (B) The results of serum biomarker combinations in diagnosing HCC in the discovery cohort were analyzed using targeted metabolites. (C) The prognostic value of serum biomarker combinations was analyzed using targeted metabolites.

[0023] Figure 7 The results show the performance of using Lasso, support vector machine and logistic regression to validate the combination of biomarkers.

[0024] Figure 8 The results of diagnosing early and late-stage liver cancer using a combination of serum biomarkers are shown. Detailed Implementation

[0025] This application discloses the application of five specific metabolites in the diagnosis or prognosis of liver cancer. These five metabolites are 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline valine, and 1-oleoyl-L-α-lysophospholipid. The inventors established an optimal combination model of biomarkers using machine learning algorithms. This combination of five metabolites serves as a non-invasive diagnostic biomarker for liver cancer, not only distinguishing liver cancer patients from healthy controls but also predicting the prognosis of liver cancer patients.

[0026] The inventors of this application have successfully established a novel non-invasive method for diagnosing liver cancer based on the aforementioned metabolite biomarkers, which is expected to provide new strategies and ideas for the early diagnosis, prevention, or treatment of liver cancer. This novel combination of biomarkers developed by the inventors provides a reliable and efficient tool for liver cancer diagnosis.

[0027] The inventors discovered that the levels of five metabolites—1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine-chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophospholipid—in samples such as serum, plasma, or feces are closely associated with liver cancer. For example, the concentrations of three metabolites—1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, and taurine-chenodeoxycholic acid—are increased in the serum, plasma, or feces of liver cancer patients; while the concentrations of two metabolites—proline-valine and 1-oleoyl-L-α-lysophospholipid—are decreased in the serum, plasma, or feces of liver cancer patients.

[0028] The term "metabolite" as used in this article refers to substances produced during the metabolism of the body and microorganisms and found in serum, including 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine-chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine. "1-Stearyl-2-arachidonicoyl-sn-glycerol" is a diacylglycerol (DAG) containing polyunsaturated fatty acids. "NG,NG-dimethyl-L-arginine" is a derivative of arginine. "Taurine-chenodeoxycholic acid" as used in this article can also refer to common bile salts, such as sodium taurine-chenodeoxycholic acid. "Proline-valine" is a dipeptide composed of proline and valine. "1-Oleoyl-L-α-lysophosphatidylcholine" is an analogue of lysophosphatidylcholine (LPA).

[0029] In some embodiments, the kits disclosed herein contain reagents for detecting the concentration of the aforementioned metabolites in samples from an individual. In some embodiments, liquid chromatography-mass spectrometry (LC-MS) is used to analyze the metabolites in samples such as serum or fecal samples.

[0030] In a specific implementation, the kit disclosed herein can simultaneously detect multiple metabolites, including 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophospholipid.

[0031] In some embodiments, the kits disclosed herein may also include suitable standard controls. In specific embodiments, the standard controls indicate the mean values ​​of metabolites from healthy individuals without liver cancer. In some embodiments, the standard controls may be provided in the form of set values. Additionally, the kits of the present invention may provide instructions to guide users in analyzing test samples and assessing the presence, risk, or status of liver cancer in test individuals.

[0032] The inventors of this application have discovered that, compared to a single biomarker, detecting a combination of multiple biomarkers can significantly improve the sensitivity and accuracy of detection, and better predict or diagnose liver cancer.

[0033] In some embodiments, the sample from an individual includes plasma, serum, whole blood, and other suitable bodily fluids, such as feces. In a specific embodiment, the sample is serum or plasma. In other embodiments, the sample is feces.

[0034] The term "individual" as used in this application refers to mammals, including but not limited to primates, cattle, horses, pigs, sheep, goats, dogs, cats, and rodents such as rats and mice. The methods and kits of this application are applicable to biological samples derived from humans. In certain embodiments, the individual has been identified as having cancer or suspected of having cancer.

[0035] Control samples can be samples from individuals who do not have liver cancer, pooled samples from multiple individuals who do not have liver cancer, or control or baseline expression levels of the average expression level of individuals known to be free of liver cancer.

[0036] In some implementations, a value is calculated based on the following logistic regression formula, compared with a threshold, and a value greater than the threshold indicates that the individual has or may have liver cancer:

[0037] logit(P)=log(P / (1-P))=1.986+2.508×taurine chenodeoxycholic acid concentration–4.499×1-oleoyl-L-α-lysophosphatidic acid concentration-2.088×NG,NG-dimethyl-L-arginine concentration-1.762×1-stearoyl-2-arachidonicoyl-sn-glycerol concentration+0.191×prolyl-valine concentration.

[0038] In a specific implementation, the cutoff is 0.68. If the value calculated based on the above formula is greater than 0.68, the individual is considered to have liver cancer; if it is less than 0.68, the individual is considered to be healthy.

[0039] The “threshold” mentioned in this article is determined based on the definition of logistic regression.

[0040] In some implementations, through non-targeted metabolomics analysis, the inventors discovered differences in the abundance of serum and fecal metabolites between liver cancer patients and healthy individuals, including 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidic acid.

[0041] In some embodiments, this application also provides a method for detecting metabolites in a sample, comprising contacting the sample to be tested with a detection reagent and then performing LC-MS / MS detection. LC-MS / MS assay refers to an analytical technique that combines the separating power of high-performance liquid chromatography (HPLC) with the mass analysis of mass spectrometry. This technique allows for targeted detection of molecules of interest. For example, quantitative analysis of one or more metabolites. Those skilled in the art will recognize that biomarkers in a sample can also be detected in other suitable ways.

[0042] The methods and kits of this application can also be used to monitor the effectiveness of liver cancer treatments. In an optional implementation, if the treatment regimen is effective, the levels of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, and taurine chenodeoxycholic acid in samples from liver cancer patients may decrease over time, while the levels of prolyl-valine and 1-oleoyl-L-α-lysophosphatidylcholine may increase over time; if the treatment regimen is ineffective, the levels of the biomarkers will not change or may increase or decrease over time, depending on which biomarker is being detected in the individual.

[0043] In some implementations, the non-invasive diagnostic methods for liver cancer based on serum metabolic markers described herein offer several advantages over existing methods. They provide a blood-based alternative to invasive procedures such as liver biopsy, reducing patient discomfort and associated risks.

[0044] In the specific implementation plan, the novel combination of the above five liver cancer metabolic markers was developed by the inventors based on a unique internal cohort and has undergone rigorous verification in an independent cohort to ensure its stability and reliability.

[0045] In some implementations, the inventors employed advanced machine learning algorithms during development to improve the predictive performance of the combination.

[0046] In some implementation schemes, the combination of five liver cancer metabolic biomarkers provided in this application exhibits significant advantages in sensitivity and specificity compared to existing biomarker combinations. This superior performance makes it a promising candidate for clinical diagnosis of liver cancer, providing physicians with a more accurate and reliable diagnostic tool.

[0047] In some implementation schemes, the effectiveness of the combination of five liver cancer metabolic biomarkers provided in this application has been clearly demonstrated through targeted metabolomics, further consolidating its reliability and accuracy in liver cancer screening. This validation method ensures that the above combination not only performs well under laboratory conditions, but also plays a unique diagnostic role in practical applications.

[0048] In some implementation schemes, the compatibility of the biomarker combinations described herein with metabolomics analysis makes them more convenient and efficient in practical applications. This compatibility allows physicians to use existing metabolomics analysis equipment and technologies to perform simple and effective measurements of biomarkers.

[0049] In some implementations, the biomarker combinations described herein outperform existing or prior products in terms of performance, reliability, compatibility, and / or usability, offering new hope and possibilities for the early diagnosis and precision treatment of liver cancer.

[0050] In this specification and claims, the words “comprising,” “including,” and “containing” mean “including but not limited to” and are not intended to exclude other parts, additives, components, or steps.

[0051] It should be understood that the features, characteristics, components or steps described in a particular aspect, embodiment or example of this application may be applied to any other aspect, embodiment or example described herein, unless there is any contradiction.

[0052] The foregoing disclosure generally describes the present invention, and the following embodiments further illustrate the invention. These embodiments are described merely to illustrate the invention and not to limit its scope. Although specific terms and values ​​are used herein, they are also understood to be exemplary and do not limit the scope of the invention. Unless specifically indicated, the experimental methods and techniques described herein are methods and techniques well known to those skilled in the art.

[0053] Example

[0054] Experimental related methods

[0055] Metabolite extraction

[0056] 100 μL of plasma sample or 25 mg of fecal sample was mixed with 500 μL or 300 μL of extraction buffer (methanol to water = 3:1 (V / V), containing an isotope-labeled internal standard), respectively. The sample mixture was ground at 35 Hz for 4 minutes and sonicated in an ice-water bath for 5 minutes each time, for a total of 3 times. Then, it was incubated at -40 °C for 1 hour to precipitate proteins. The samples were then centrifuged at 12,000 rpm for 15 minutes at 4 °C, and the supernatant was collected. Quality control samples were prepared by mixing equal volumes of the supernatant from all samples.

[0057] Metabolomics analysis

[0058] Metabolites in plasma and fecal samples were analyzed using liquid chromatography-mass spectrometry (LC-MS). Chromatographic separation was performed using a UHPLC system (Vanquish, Thermo Fisher Scientific, Waltham, MA) coupled with a UPLC HSS T3 column (2.1 mm × 100 mm, 1.8 μm) to the mass spectrometer. The mobile phase consisted of a solution of 5 mmol / L ammonium acetate and 5 mmol / L acetic acid in water (A) and acetonitrile (B). The autosampler temperature was 4 °C, and the injection volume was 2 μL.

[0059] For targeted metabolomics analysis, an Orbitrap Exploris 120 mass spectrometer (Orbitrap MS, Thermo Fisher Scientific) was used, and MS / MS spectra were acquired in information-correlated acquisition mode using the acquisition software (Xcalibur, Thermo Fisher Scientific). ESI source conditions were set as follows: sheath flow rate = 50 Arb; auxiliary gas flow rate = 15 Arb; capillary temperature = 320℃; full MS resolution = 60000; MS / MS resolution = 15000; collision energy in NCE mode = 10 / 30 / 60; spray voltage = 3.8 kV (positive) or -3.4 kV (negative).

[0060] Metabolomics data analysis and metabolite annotation

[0061] The raw data files were converted to mzXML format by ProteoWizard and processed by the R package XCMS (version 3.2) to generate a data matrix consisting of retention time (RT), mass-to-charge ratio (m / z) values, and peak abundance. Peak annotation was then performed using the R package CAMERA. Peaks with RSD > 30% in quality control samples or missing values ​​(intensity = 0) in > 50% of samples were discarded. The area of ​​each peak was normalized using total ion current in polar metabolomics. To eliminate technical variations and avoid batch effects, each peak was normalized using the LOESS method, which is used to predict and correct the intensity of the same metabolite in real samples. MS and MS / MS spectra were examined based on accurate quality (±10 ppm, ±25 ppm) and RT values ​​(±30 sec). Only matching MS and MS / MS spectra were included for metabolite annotation. Filtered MS / MS spectra were evaluated using an internal database (BiotreeDB) for polar metabolite annotation based on accurate quality (m / z, ±10 ppm), RT, and spectral mode. The MS / MS spectral matching score is calculated using a dot product algorithm based on peak fragments and intensities.

[0062] Selection of metabolic biomarkers

[0063] Endogenous metabolites were selected using the Human Metabolome Database (HMDB 4.0). A multivariate ANOVA with 999 permutations of the Bray-Curtis distance was performed to calculate statistical differences between metabolic characteristics and clinical parameters (including age, sex, BMI, and tumor location). To assess the diagnostic potential of the selected metabolites, three machine learning models (random forest, least absolute shrinkage and selection operator (LASSO), and support vector machine) and a stepwise logistic regression model were built using R packages, and 1000 cross-validations were performed. ROC analysis was conducted using the pROC package in R.

[0064] Statistical analysis

[0065] Log-transformation was performed on the pooled UHPLC-MS data (positive and negative modes) for subsequent analysis. A two-tailed Wilcoxon rank-sum test was performed to measure the significance of each metabolite between liver cancer and normal control samples, and the results were adjusted for multiple tests using false discovery rate correction (Benjamini-Hochberg method). Partial least-squares discriminant analysis (PLS-DA) and orthogonal PLS-DA (OPLS-DA) were performed using permutation tests (R package ropls, version 1.26.4). Metabolite difference analysis was performed using the R package MetaboAnalystR (version 3.0.3). Metabolites with a p-value <0.05 were considered significantly different.

[0066] Example 1: Construction of a combination of five metabolic biomarkers

[0067] We collected serum and stool samples from two independent cohorts comprising 107 hepatocellular carcinoma patients and 103 healthy individuals. Figure 1 The study included a cohort of 57 patients with liver cancer and 53 healthy individuals, and a validation cohort of 50 patients with liver cancer and 50 healthy individuals. The majority of liver cancer cases in both cohorts were caused by hepatitis B virus infection.

[0068] Through non-targeted metabolomics analysis, we found differences in serum metabolite abundance between liver cancer patients and healthy individuals. Figure 2 A). 33 metabolites were elevated in liver cancer, and 16 metabolites were decreased in liver cancer. Figure 2 B). Similarly, there were significant differences in fecal metabolites between liver cancer patients and healthy individuals. Figure 3 A). 20 metabolites were elevated in liver cancer, and 35 metabolites were decreased in liver cancer. Figure 3 B). Further pathway analysis revealed significant changes in the arginine and proline metabolic pathways and long-chain fatty acid oxidation in the serum of liver cancer patients. Figure 4 A). Furthermore, significant alterations were observed in the feces of liver cancer patients in pathways such as pyrimidine metabolism and amino sugar metabolism. Figure 4 B).

[0069] Through combined analysis of metabolites in serum and feces, we found increased levels of γ-L-glutamyl-L-valine and NG,NG-dimethyl-L-arginine, and decreased levels of butyl lactate in the serum and feces of liver cancer patients. To validate these findings, we further investigated the functions of these characteristic metabolites. Through a series of in vitro functional experiments, we confirmed that γ-L-glutamyl-L-valine has a significant pro-tumor effect. Specifically, it significantly promotes the growth of liver cancer cells, inhibits apoptosis in liver cancer cells, and pushes cells into the S phase, thereby accelerating cell proliferation. Furthermore, we also found that γ-L-glutamyl-L-valine can promote the proliferation of liver cancer organoids (…). Figure 5 In contrast, butyl lactate exhibits significant antitumor effects. Through in vitro experiments, we observed that butyl lactate can inhibit the growth of liver cancer cells and promote their apoptosis, thereby exerting an antitumor effect. Figure 5 EH).

[0070] Building upon this foundation, in a cohort study comprising 57 hepatocellular carcinoma (HCC) patients and 53 healthy controls, we successfully constructed a combination of five serum metabolic biomarkers using the random forest method. These five biomarkers are: 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine. The inventors found that this combination of metabolites can accurately distinguish HCC patients from healthy controls. It exhibits extremely high accuracy (AUC = 0.987) in differentiating HCC patients from healthy individuals, with a sensitivity of 97.6% and a specificity of 92.7%. Figure 6 A).

[0071] The logistic regression formula is logit(P) = log(P / (1-P)) = 1.986 + 2.508 × taurine chenodeoxycholic acid concentration – 4.499 × 1-oleoyl-L-α-lysophosphatidic acid concentration – 2.088 × NG,NG-dimethyl-L-arginine concentration – 1.762 × 1-stearoyl-2-arachidonicoyl-sn-glycerol concentration + 0.191 × prolyl-valine concentration. The cutoff value is 0.68. Values ​​calculated based on the above logistic regression formula greater than 0.68 are considered liver cancer patients, and values ​​lower than 0.68 are considered healthy individuals.

[0072] Example 2: Validating the effectiveness of serum biomarker combinations using targeted metabolomics.

[0073] We also used targeted metabolomics to validate the efficacy of a biomarker combination composed of five metabolites: 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidic acid. The results showed an AUC of 0.950, a sensitivity of 95.8%, and a specificity of 87.8%. Figure 6 B) further confirms its accuracy and reliability.

[0074] Example 3 validates the effectiveness of serum biomarker combinations using other methods.

[0075] We further validated the effectiveness of the biomarker combination consisting of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidyl acid using multiple methods including Lasso (AUC 0.961), SVM (AUC 0.955), and logistic regression (AUC 0.948). Figure 7 This ensures the stability and reliability of the diagnostic results.

[0076] Example 4: Diagnostic efficacy of biomarker combinations in an independent validation cohort.

[0077] We further validated the diagnostic efficacy of a serum biomarker combination consisting of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine (ADMA), taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine in an independent validation cohort of 50 hepatocellular carcinoma patients and 50 healthy controls. The results showed that the biomarker combination had an AUC of 0.929, a sensitivity of 85.3%, and a specificity of 94.0%. Figure 6 A) confirms its diagnostic value in practical applications.

[0078] Example 5: Diagnostic efficacy of biomarker combinations for early and late-stage liver cancer

[0079] In the aforementioned cohort of 57 patients with hepatocellular carcinoma and 53 healthy individuals, we also examined the diagnostic efficacy of a combination of serum biomarkers consisting of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidic acid in early and late-stage hepatocellular carcinoma.

[0080] The results showed that the serum biomarker combination composed of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine had an AUC of 0.926, a sensitivity of 80.0%, and a specificity of 96.0% in the diagnosis of early-stage liver cancer (stage I + II). Figure 8 A); In the diagnosis of advanced liver cancer (stage III + IV), the AUC was 0.943, the sensitivity was 86.4%, and the specificity was 96.0%. Figure 8 B).

[0081] Example 6: Predictive ability of biomarker combinations for survival in liver cancer patients

[0082] We also examined the predictive ability of a serum metabolite combination consisting of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine for survival in the aforementioned cohort of 57 hepatocellular carcinoma patients and 53 healthy individuals. The results showed that the AUC of this metabolic biomarker combination was 0.736, with a sensitivity of 85.3% and a specificity of 63.3%. Figure 6 C).

[0083] The above results indicate that the combination of biomarkers containing five metabolites—1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine (ADMA), taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidic acid—has a good distinguishing effect between normal individuals and liver cancer patients, and can be effectively used for the diagnosis or prognosis of liver cancer.

[0084] Various changes and equivalent substitutions may be made to the embodiments disclosed in this application without departing from the spirit and scope of this disclosure. Unless the context otherwise requires, any feature, step, or embodiment of the embodiments disclosed herein may be used in combination with any other feature or embodiment.

Claims

1. A kit for diagnosing or identifying liver cancer in an individual, comprising reagents for detecting metabolite levels in a sample from said individual, said metabolites comprising one or more of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophosphatidic acid.

2. The kit according to claim 1, wherein the liver cancer is hepatocellular carcinoma.

3. The kit of claim 1, wherein the sample is selected from serum, feces, plasma or whole blood, preferably serum or feces, more preferably serum.

4. The kit according to any one of claims 1-3, wherein the metabolite is selected from two or more of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine and 1-oleoyl-L-α-lysophosphatidic acid, preferably five.

5. The kit according to any one of claims 1-3, wherein the kit is a kit for performing LC-MS / MS analysis and contains mass spectrometry standards for detecting the metabolites.

6. Use of a reagent for detecting metabolite levels in a sample from an individual in the preparation of a kit or drug for diagnosing or identifying liver cancer in said individual, said metabolite comprising one or more of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, prolyl-valine, and 1-oleoyl-L-α-lysophosphatidylcholine.

7. The use as claimed in claim 6, wherein the sample is selected from serum, feces, plasma or whole blood, preferably serum or feces, more preferably serum.

8. The use as described in claim 6 or 7, wherein the metabolite is 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophosphatidylcholine.

9. The use as claimed in claim 8, wherein if the levels of 1-stearoyl-2-arachidonicoyl-sn-glycerol, NG,NG-dimethyl-L-arginine, and taurine chenodeoxycholic acid in a sample from said individual are higher than those in a control sample, and the levels of prolyl-valine and 1-oleoyl-L-α-lysophosphatidylcholine in a sample from said individual are lower than those in a control sample, then said individual has or may have liver cancer; or A value is calculated based on the following logistic regression formula and compared with a threshold. A value greater than the threshold indicates that the individual has or may have liver cancer: logit(P)=log(P / (1-P))=1.986+2.508×taurine chenodeoxycholic acid concentration–4.499×1-oleoyl-L-α-lysophosphatidic acid concentration-2.088×NG,NG-dimethyl-L-arginine concentration-1.762×1-stearoyl-2-arachidonicoyl-sn-glycerol concentration+0.191×prolyl-valine concentration.

10. A method for detecting metabolite levels in a sample, comprising treating the sample with a reagent from the kit of any one of claims 1-5, wherein the metabolite comprises one or more of 1-stearoyl-2-arachidonic-sn-glycerol, NG,NG-dimethyl-L-arginine, taurine chenodeoxycholic acid, proline-valine, and 1-oleoyl-L-α-lysophosphatidic acid, preferably, the sample being selected from serum, feces, plasma, or whole blood, more preferably serum.

Images