Application of ACSL4 in predicting prognosis of poor drug-induced liver injury

Abstract

The invention relates to the technical field of biological medicine, and discloses application of ACSL4 in predicting prognosis of drug-induced liver injury. The screened biomarker ACSL4 is remarkably and highly expressed in prognosis patients with poor drug-induced liver injury, the prognosis condition of the patients can be effectively evaluated by detecting the expression level of the biomarker, and the biomarker ACSL4 has high diagnostic value.

Description

Technical Field

[0001] The present invention relates to the field of biomedicine technology, and particularly relates to the application of ACSL4 in predicting the poor prognosis of drug-induced liver injury. Background Art

[0002] Clinically, about 5-10% of patients with drug-induced liver injury (DILI) will have poor prognosis, including liver failure, liver transplantation or death. Therefore, the prediction of poor prognosis in DILI patients is of great significance for clinical decision-making. At present, the methods used clinically to predict the prognosis of DILI mainly rely on traditional clinical indicators and routine biochemical tests, but these methods have the following limitations:

[0003] (1) Low accuracy: Traditional indicators are difficult to accurately predict the poor prognosis of patients at an early stage;

[0004] (2) Lack of individuation and precision in detection methods: Scoring systems such as MELD score and Roble-Diaz are applicable to all liver diseases, but lack specificity in predicting the poor prognosis of DILI;

[0005] (3) Relatively complex usage methods: The above models all require two or more biochemical test indicators or need to convert the test indicators between units, and the usage experience is relatively poor. Summary of the Invention

[0006] The object of the present invention is to overcome the problems of the existing models, such as low diagnostic accuracy, lack of specificity, and relatively complex usage methods, and to provide a biomarker ACSL4 for predicting the poor prognosis of drug-induced liver injury.

[0007] To achieve the above object, the present invention provides the application of ACSL4 in the preparation of a kit for predicting the poor prognosis of drug-induced liver injury.

[0008] The present invention provides the application of a reagent for detecting ACSL4 in the preparation of a kit for predicting the poor prognosis of drug-induced liver injury.

[0009] The present invention provides a system for predicting the poor prognosis of drug-induced liver injury in a subject, and the system includes:

[0010] (a) A module for inputting features related to the poor prognosis of drug-induced liver injury, which is used to input features related to the poor prognosis of drug-induced liver injury in a subject, and the features related to the poor prognosis of drug-induced liver injury include: ACSL4;

[0011] (b) The adverse prognosis discrimination and processing module for drug-induced liver injury is used to input the features related to the adverse prognosis of drug-induced liver injury into the judgment model for the adverse prognosis of drug-induced liver injury, so as to obtain the risk possibility; and compare the risk possibility with the risk degree threshold of the adverse prognosis of drug-induced liver injury, so as to obtain the auxiliary diagnosis result;

[0012] (c) The auxiliary diagnosis result output module is used to output the above-mentioned auxiliary diagnosis result.

[0013] The novel biomarker ACSL4 discovered by the present invention is used to predict the adverse prognosis of DILI patients, and has high specificity and accuracy. The novel biomarker of the present invention only contains one index, and the detection process is convenient and fast, which can assist the diagnosis of pathologists, provide a reliable basis for the treatment plan, and avoid the occurrence of adverse prognosis. Description of the Drawings

[0014] Figure 1A-1F Shows the clinical cohort sample collection process in the embodiment of the present invention;

[0015] Figure 2A Shows the ACSL4 detection levels of patients in the good prognosis group and the poor prognosis group in the embodiment of the present invention;

[0016] Figure 2B Shows the comparison results of using ACSL4 as a biomarker for prediction and the previous prediction methods in the embodiment of the present invention. Detailed Embodiments

[0017] In the ranges disclosed herein, the endpoints and any values are not limited to the exact ranges or values, and these ranges or values should be understood to include values close to these ranges or values. For numerical ranges, the endpoints of each range, between the endpoints of each range and a single point value, and between single point values can be combined with each other to obtain one or more new numerical ranges, and these numerical ranges should be regarded as specifically disclosed herein.

[0018] In the present invention, the term "prognosis" refers to the prediction of the development process and final outcome of a certain disease. According to whether the disease is treated during its occurrence or progression, the prognosis can be divided into natural prognosis (i.e., the natural progression of the disease without intervention) and treatment prognosis (i.e., the outcome of the disease after treatment intervention). In the present invention, the terms "poor prognosis" and "adverse prognosis" can be used interchangeably. "Prediction" refers to the estimation of the patient's future possible health status, disease progression, complication risk, and response to treatment. Prediction helps doctors formulate personalized treatment plans, evaluate treatment effects, and provide prognosis information for patients, so as to optimize disease management. The significance of predicting the adverse prognosis of drug-induced liver injury mainly includes the following aspects:

[0019] (1) Early identification of high-risk patients: By predicting poor prognosis, patients who may develop liver failure, require liver transplantation, or are at risk of death can be identified in advance, enabling more proactive intervention measures to be taken;

[0020] (2) Optimization of treatment plans: For patients predicted to have a poor prognosis, doctors can promptly adjust the medication plan, avoid using drugs that may exacerbate liver injury, and select safer alternative treatment plans;

[0021] (3) Improvement of patient management: Predicting a poor prognosis helps in more refined stratified management of patients. High-risk patients can receive closer monitoring and more frequent follow-ups to detect changes in the condition at an early stage;

[0022] (4) Reduction of medical costs: Through early identification and intervention, the high medical costs associated with liver failure and liver transplantation can be reduced, improving the utilization efficiency of medical resources;

[0023] (5) Improvement of patient survival rate and quality of life: Timely adjustment of treatment plans and enhanced management can significantly reduce the incidence of poor prognosis, extend the patient's survival period, and improve the quality of life.

[0024] In the present invention, poor prognosis of drug-induced liver injury includes liver failure, the need for liver transplantation, or the risk of death. The present invention is particularly suitable for analyzing the expression level of ACSL4 by taking blood or tissue samples when the subject is ill, so as to identify in advance patients who may have a poor prognosis based on the test results and promptly adjust the medication plan.

[0025] In the present invention, ACSL4 (Acyl-CoA synthetase long-chain family member 4) is long-chain acyl-CoA synthetase 4, and its amino acid sequence can be referred to NCBI Reference Sequence: NM_001318509.2. The aliases of ACSL4 are ACS4, FACL4, LACS4, MRX63, MRX68, XLID63. It is a protein encoded by the ACSL4 gene in the human genome, and this gene is located in the region of Xq22.3 - 23 on the X chromosome. ACSL4 is expressed in a variety of tissues, especially at relatively high levels in tissues such as the adrenal gland and appendix. Its mechanism of action is to catalyze the conversion of free long-chain fatty acids into acyl-CoA esters, participate in lipid biosynthesis and fatty acid degradation, and preferentially utilize arachidonic acid as a substrate. ACSL4 plays an important role in regulating lipid metabolism, inflammatory responses, insulin secretion, skeletal muscle protein metabolism, and cell death modes such as ferroptosis. The abnormal expression of ACSL4 is associated with a variety of diseases, including non-alcoholic fatty liver disease, cancer, ischemic stroke, and acute kidney injury.

[0026] The first aspect of the present invention provides an application of ACSL4 in the preparation of a kit for predicting poor prognosis of drug-induced liver injury.

[0027] The second aspect of the present invention provides an application of a reagent for detecting ACSL4 in the preparation of a kit for predicting poor prognosis of drug-induced liver injury.

[0028] In the present invention, the reagent may be a commonly used reagent in the art capable of detecting ACSL4, such as an ELISA (enzyme-linked immunosorbent assay) reagent. Preferably, the reagent includes an ELISA reagent for quantitative detection of ACSL4 (such as an enzyme-labeled antibody, a substrate solution, a washing solution, a termination solution, a standard, a control, a diluent, a secondary antibody, etc.) and / or tools (such as a solid-phase carrier, a sealing film, an instruction manual, etc.). The enzyme-labeled antibody (conjugate) is an antibody that specifically binds to the target molecule and is usually conjugated to an enzyme (such as horseradish peroxidase HRP) for amplifying the detection signal. The substrate solution reacts with the enzyme to generate a colored or fluorescent product for quantitative or qualitative analysis of the concentration of the target molecule. The washing solution is used to wash the microplate to remove unbound substances and reduce background interference, and commonly used is phosphate buffered saline containing Tween 20 (polysorbate-20). The termination solution is used to stop the enzymatic reaction and fix the color or fluorescent signal, and commonly used is a sulfuric acid solution, and the concentration can be adjusted according to experimental requirements. The standard is used to establish a standard curve, and the controls (negative control and positive control) are used to verify the accuracy and reliability of the experiment. The diluent is used to dilute the sample and the conjugate to ensure that their concentrations are suitable for the experimental conditions. The secondary antibody (optional) is used to enhance the binding of the detection antibody to the target molecule and further improve the detection sensitivity. The solid-phase carrier is coated with an antigen or an antibody on its surface for capturing the target molecule in the sample. The sealing film is used to cover the microplate during incubation to prevent evaporation and contamination. The instruction manual provides detailed operation steps, reaction conditions, and result interpretations to ensure that the experimenter uses it correctly.

[0029] In the present invention, the ACSL4 mainly refers to ACSL4 in serum and / or plasma (blood). Preferably, the reagent is used to detect ACSL4 in serum and / or plasma (blood).

[0030] In the present invention, preferably, the object to be predicted is a mammal, particularly a primate, and more preferably a human.

[0031] The third aspect of the present invention provides a system for predicting poor prognosis of drug-induced liver injury in an object, and the system includes:

[0032] (a) A module for inputting features related to poor prognosis of drug-induced liver injury, which is used to input features related to poor prognosis of drug-induced liver injury of the object, and the features related to poor prognosis of drug-induced liver injury include: ACSL4;

[0033] (b) The adverse prognosis discrimination and treatment module for drug-induced liver injury is used to input the features related to the adverse prognosis of drug-induced liver injury into the adverse prognosis judgment model of drug-induced liver injury, so as to obtain the risk probability; and compare the risk probability with the risk degree threshold of the adverse prognosis of drug-induced liver injury, so as to obtain an auxiliary diagnosis result;

[0034] (c) The auxiliary diagnosis result output module is used to output the auxiliary diagnosis result.

[0035] Preferably, in step (a), the related features include the expression level of the biomarker ACSL4.

[0036] Preferably, when the risk probability is greater than 0.655, it is predicted as an adverse prognosis of drug-induced liver injury.

[0037] Preferably, the features related to the adverse prognosis of drug-induced liver injury are derived from serum and / or plasma (blood).

[0038] In the present invention, as described above, preferably, the subject is a mammal, particularly a primate, and more preferably a human.

[0039] The present invention screens out the significantly abnormal candidate biomarker ACSL4 by analyzing the protein expression profile in the liver tissue of patients with adverse prognosis of DILI. This biomarker ACSL4 is used to predict the adverse prognosis of DILI, with high accuracy and strong specificity.

[0040] The present invention will be described in detail below through examples. In the following examples, unless otherwise specified, the methods used are conventional methods in the art, and the reagents used are conventional reagents and can be obtained through commercial purchase. For the experimental methods without specific conditions noted in the examples, they are usually carried out under conventional conditions or according to the conditions recommended by the manufacturer.

[0041] The definition of patients with drug-induced liver injury refers to the "Chinese Guidelines for Drug-induced Liver Injury (2023 Edition)": (1) There is a temporal causal relationship between the patient's liver injury and the drug use situation; (2) The Roussel Uclaf Causality Assessment Score ≥ 6 points, and the biochemical indicators at the time of onset meet one of the following (the RUCAM scoring scale is shown in the appendix): ① The level of Alanine aminotransferase (ALT) ≥ 5 × the Upper limit of normal (ULN); ② Or the level of Alkaline phosphatase (ALP) ≥ 2 × ULN and the ALP elevation related to bone disease is excluded; ③ Or the level of ALT ≥ 3 × ULN and the total bilirubin (TB) > 2 × ULN.

[0042] Good prognosis is defined as the normalization of biochemical indicators in DILI patients within 1 year, and poor prognosis is defined as the occurrence of liver failure, liver transplantation, or death in patients, as follows:

[0043] The definition of liver failure refers to the "Diagnosis and Treatment Guidelines for Liver Failure (2024 Edition)": For patients without a history of underlying liver disease, with an acute onset, and presenting with grade II or higher hepatic encephalopathy (classified according to the four-grade classification method) within 4 weeks and having the following manifestations: (1) Severe digestive tract symptoms such as fatigue, anorexia, abdominal distension, nausea, and vomiting; (2) Coagulation dysfunction, with an international normalized ratio ≥ 1.5 or a prothrombin activity ≤ 40%, and excluding other causes; (3) Progressive elevation of total bilirubin.

[0044] Liver transplantation is defined as the treatment of DILI patients who must undergo liver transplantation due to end-stage liver disease or liver failure.

[0045] Example 1

[0046] Screening of Biomarker ACSL4

[0047] (I) Preparation of Liver Tissue Protein Samples

[0048] 1. Sample Collection

[0049] Collect liver tissue samples from the healthy control group (N = 6) and patients with poor prognosis of drug-induced liver injury (DILI) (N = 6). Before obtaining the samples, all samples have been approved by the Ethics Committee of Beijing Friendship Hospital, Capital Medical University, and the patient informed consent form has been signed (Ethics number: 2022-P2-063).

[0050] 2. Sample Processing and Protein Extraction

[0051] Take out the frozen liver tissue samples, add liquid nitrogen, grind them thoroughly, and take an appropriate amount of the sample into a 1.5 mL centrifuge tube. Add sample lysis buffer, phosphatase inhibitor, and protease inhibitor PMSF to the centrifuge tube to make their final concentration 1 mM. Grind the sample with a cold grinder at -35 °C, 60 Hz, for 120 s, and then repeat this operation once. After grinding, centrifuge the solution at 12,000 rpm for 10 min at 4 °C and take the supernatant. The supernatant is the total protein solution of the sample. Measure the protein concentration using the BCA method, and store the protein solution at -80 °C for later use.

[0052] 3. Protease Digestion and Peptide Desalting

[0053] The protein concentration was determined using the BCA method (Thermofisher). According to the measured concentration, an appropriate amount of protein was taken from each sample, and the samples in different groups were diluted and adjusted to the same concentration and volume with the lysis buffer. Dithiothreitol was added to the above protein solution to a final concentration of 5 mM. After mixing, it was incubated at 55 °C for 30 min, and then cooled on ice to room temperature. Subsequently, the corresponding volume of iodoacetamide was added to a final concentration of 10 mM, mixed well, and placed in the dark at room temperature for 15 min. Then, 6 volumes of acetone were added, and it was placed at -20 °C for more than four hours. The precipitate was collected by centrifugation at 4 °C, 8000×g for 10 minutes. After the acetone had evaporated for 2 - 3 min, the precipitate was redissolved in 100 μL of 50 mM NH4HCO3. Trypsin (1 mg / mL) at 1 / 50 of the sample mass was added, and digestion was carried out overnight at 37 °C. The enzymatically digested samples were lyophilized and stored at -80 °C. The peptides after enzymatic digestion were desalted using a SOLA TM SPE 96-well plate. The column was activated with 200 μL of methanol, and this was repeated twice. Then, the column was activated with 200 μL of pure water (containing 0.1% formic acid), and this was repeated twice. A sample volume of 50 - 500 μL was added, the vacuum was adjusted, and the droplet rate was maintained at 1 mL / min (about 1 drop / second). The sample addition was repeated once. The column was washed with 200 μL of 0.1% formic acid - water, and this was repeated twice. Finally, the peptides were eluted with 150 μL of 50% acetonitrile - water (containing 0.1% formic acid), and this was repeated twice, for a total of three times, to obtain 450 μL of eluate, which was dried under vacuum.

[0054] (II) High-throughput sequencing of proteins

[0055] 1. Before mass spectrometry injection, each sample was mixed with an internal standard in a volume ratio of iRT: sample to be measured = 1:20. Equal amounts of peptides were taken from all enzymatically digested samples and separated using an EASY-nLC 1200 liquid phase. Mobile phase A was 0.1% FA aqueous solution, and mobile phase B was ACN containing 0.1% FA. Gradient elution conditions: 0 - 20 min, 5 - 22% B; 20 - 24 min, 22 - 37% B; 24 - 27 min, 37 - 80% B; 27 - 30 min, 80% B. The peptides were separated by the ultra-high performance liquid system and then injected into a timsTOFPro mass spectrometer (Bruker) for analysis. The mass spectrometry conditions were as follows: capillary voltage was 1.4 KV, drying gas temperature was 180 °C, drying gas flow rate was 3.0 L / min, mass spectrometry scanning range was 100 - 1700 m / z, ion mobility range was 0.7 - 1.3 Vs / cm2, and collision energy range was 20 - 59 eV.

[0056] 2. Use Spectronaut Pulsar TM18.4 (Biognosys, Swiss) software processes the DIA raw data. The mass spectrometry retrieval parameters are as follows: precursor mass value threshold of 0.01, protein mass value threshold of 0.01, fixed modification of Carbamidomethyl (C), variable modifications of Oxidation (M) and Acetyl (N-term), and a maximum of 2 missed cleavage sites.

[0057] 3. Data normalization and statistical analysis: The protein expression data is normalized (total protein mass normalization method) to ensure that the expression levels between different samples can be compared.

[0058] (III) Screening of candidate biomarkers

[0059] The high-throughput sequencing data is analyzed by bioinformatics methods. Through analysis means such as clustering, differential expression analysis, GO / KEGG enrichment analysis, and protein-protein interaction analysis, candidate biomarkers are screened.

[0060] 1. Use R packages such as "DeSeq2", "Limma", "EdgeR", etc. to perform differential expression analysis on the expression matrix;

[0061] 2. Use R packages such as "EnhancedVolcano", "ggplot2", etc. to draw volcano plots for differentially expressed genes ( Figure 1A );

[0062] 3. Use R packages such as "pheatmap", "ComplexHeatmap", etc. to draw heatmaps for differentially expressed genes ( Figure 1B );

[0063] 4. Use "clusterProfiler" to perform GO and KEGG enrichment analysis on differentially expressed genes, and select a relevant gene list for the next step of analysis in combination with the disease pathogenesis ( Figure 1C -D);

[0064] 5. Use R packages such as "VennDiagram" to draw Venn diagrams for differentially expressed genes and the gene list selected in the previous step to find differentially expressed genes related to the disease onset ( Figure 1E );

[0065] 6. Use the String website and Cytoscape software to perform protein-protein interaction analysis to find the key gene ACSL4 related to the disease pathogenesis ( Figure 1F ).

[0066] Example 2

[0067] Verification of biomarker ACSL4 and comparison with previous models

[0068] (I) Sample size calculation

[0069] Before collecting cases, sample size calculation is required. The present invention uses the Events per Variable (EPV) rule for sample size calculation:

[0070] 1. Sample size calculation formula

[0071]

[0072] 2. The sample size calculation is based on the following assumptions and parameters

[0073] (1) Number of events: For the Logistic regression model, it is usually required that each independent variable corresponds to at least n events. Considering that the incidence rate of DILI is relatively low compared with common diseases, it is required that each independent variable corresponds to at least 5 - 8 events;

[0074] (2) Event incidence rate: Combining the reports in previous literature, the estimated event incidence rate is 5 - 10%.

[0075] 3. Sample size calculation results

[0076] According to the previous model and the required number of events, the number of independent variables = 3, and the corresponding number of events (poor prognosis group) is 15 - 24. According to the event incidence rate, the minimum total sample size calculated is N = 150 cases. Considering possible loss to follow-up or data missing, the sample size is increased by 10%. Finally, the total expected sample size is N = 165 cases, including 15 poor prognosis events and 150 good prognosis events.

[0077] (II) Detection of candidate markers

[0078] 1. After obtaining the approval of the Ethics Committee of Beijing Friendship Hospital, Capital Medical University, a sufficient number of patients with poor prognosis (N = 19) and patients with good prognosis (N = 163) are recruited. In addition, samples of healthy people are collected as controls (N = 40).

[0079] 2. Collect blood samples from patients. Enzyme - linked immunosorbent assay (ELISA) is selected for quantitative detection. Ensure the use of a validated ELISA kit (FineTest), and the detection process is strictly carried out according to the kit instructions. At the same time, collect the clinical data and laboratory indexes of patients to calculate the scores of previous different prognosis evaluation methods.

[0080] (III) Data statistics and analysis

[0081] 1. Group according to the prognosis of DILI patients, and compare the ACSL4 detection levels of patients in the good prognosis group and the poor prognosis group ( Figure 2A) The results showed that the expression level of ACSL4 was significantly increased in the poor prognosis group, and the difference was statistically significant (**** indicates P < 0.0001);

[0082] 2. Based on the clinical indicators of the patients, calculations were performed according to the definitions of Hy’s Law, new Hy’s Law, MELD, Roble Diaz, and DrIL ToxALF Score.

[0083] (IV) Model construction

[0084] 1. Based on the ACSL4 detection level and the different scores obtained in the previous step, combined with the prognosis of the patients, using R packages such as "pROC", "caret", and "glmnet", Receiver Operating Characteristic (ROC) curves were plotted and the accuracy, sensitivity, specificity, and AUROC of the model were calculated. The specific matrix (Table 1) and formulas are as follows.

[0085] Table 1

[0086]

[0087]

[0088] Accuracy: The proportion of correct predictions among all predictions

[0089]

[0090] Sensitivity: The proportion of correctly identified positive samples among all positive samples

[0091]

[0092] Specificity: The proportion of all correctly predicted negative samples among all actual negative samples

[0093]

[0094] AUROC is a curve plotted with sensitivity as the vertical axis and specificity as the horizontal axis. The area under the curve is calculated by integration and there is no single formula.

[0095] 2. Analyze the diagnostic efficacy of candidate markers in DILI patients.

[0096] 3. Compare the predictive efficacy of ACSL4 with that of previous tools for predicting poor prognosis of DILI. The models for comparison include the MELD score, the Roble-Diaz score, the DrILTox ALF score, Hy's Law, and New Hy's Law. Each model involves 2 - 3 parameters. Compared with previous models, the area under the curve of the ACSL4 model is higher than that of other models, and the sensitivity and specificity of the model are more stable than those of other models (Table 2 and Figure 2B ).

[0097] Table 2

[0098] Model AUROC Accuracy Sensitivity Specificity Hy’s Law 0.701 0.685 0.674 0.727 nHy’s Law 0.599 0.685 0.744 0.455 MELD Score 0.825 0.815 0.791 0.909 Roble-Diaz Model 0.658 0.778 0.86 0.455 DrILTox ALF Score 0.810 0.704 0.651 0.909 ACSL4 0.875 0.833 0.837 0.818

[0099] 4. To facilitate the use of ACSL4 by clinicians to predict the prognosis of DILI patients, a logistic regression model was used to establish a mathematical relationship between the expression level of ACSL4 and the risk of poor prognosis. The formula for the univariate logistic regression model is as follows:

[0100]

[0101] Where: P represents the probability of a patient experiencing a poor prognosis event; X represents the quantitative test value of ACSL4; β0 is the model intercept; β1 is the regression coefficient corresponding to the biomarker.

[0102] 5. After fitting, the final formula is:

[0103]

[0104] When P is greater than 0.655, the probability of a DILI patient experiencing a poor prognosis event is relatively high.

[0105] From the above results, it can be seen that by using the biomarker ACSL4 provided by the present invention, the poor prognosis of DILI patients can be accurately predicted.

[0106] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited thereto. Within the scope of the technical concept of the present invention, various simple modifications can be made to the technical solutions of the present invention, including any other suitable combination of each technical feature. These simple modifications and combinations should also be regarded as the content disclosed by the present invention and fall within the protection scope of the present invention.

Claims

1. Use of ACSL4 in the preparation of a kit for predicting poor prognosis of drug-induced liver injury.

2. Use of a reagent for detecting ACSL4 in the preparation of a kit for predicting poor prognosis of drug-induced liver injury.

3. The application according to claim 2, wherein, The reagent includes an ELISA reagent for quantitative detection of ACSL4.

4. The application according to claim 2, wherein, The reagent is used to detect ACSL4 in serum and / or plasma.

5. The application according to any one of claims 1-4, wherein, The object to be predicted is a mammal.

6. The application according to claim 5, wherein The object to be predicted is a human.

7. A system for predicting poor prognosis of drug-induced liver injury in an object, characterized in that, The system includes: (a) A module for inputting features related to poor prognosis of drug-induced liver injury, which is used to input features related to poor prognosis of drug-induced liver injury of an object, and the features related to poor prognosis of drug-induced liver injury include: ACSL4; (b) A discrimination and processing module for poor prognosis of drug-induced liver injury, which is used to input the features related to poor prognosis of drug-induced liver injury into a judgment model for poor prognosis of drug-induced liver injury to obtain a risk probability; and compare the risk probability with a risk threshold for poor prognosis of drug-induced liver injury to obtain an auxiliary diagnosis result; (c) An output module for the auxiliary diagnosis result, which is used to output the auxiliary diagnosis result.

8. The system according to claim 7, wherein, When the risk probability is greater than 0.655, it is predicted as poor prognosis of drug-induced liver injury.

9. The system according to claim 7 or 8, wherein, The object is a mammal.

10. The system according to claim 7 or 8, wherein, The features related to poor prognosis of drug-induced liver injury are derived from serum and / or plasma.

Images