Aortic dissection related biomarker, kit, system and application thereof

Abstract

The invention discloses a biomarker related to aortic dissection, a kit, a system and application thereof, and relates to the technical field of biological medicine. The biomarkers related to the aortic dissection comprise a first biomarker and / or a second biomarker; the first biomarker is used for diagnosing the aortic dissection, and the first biomarker comprises RCN1, SERPINA3 and CPN1; the second biomarker is used for predicting the clinical death risk of aortic dissection patients, and the second biomarker comprises SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5. According to the invention, the plasma-based protein biomarker capable of accurately screening the aortic dissection patient and predicting the death risk of the aortic dissection patient is successfully screened out, which is also beneficial to exploring the pathogenesis of the disease and has important significance on efficient, accurate and low-cost diagnosis of the aortic dissection patient.

Description

Technical Field

[0001] The present invention relates to the field of biomedicine technology, and specifically refers to a biomarker, a kit, a system and applications thereof related to aortic dissection. Background Art

[0002] Acute aortic dissection is a dangerous cardiovascular disease. It is a catastrophic disease caused by damage to the innermost layer of the aorta, which causes blood to flow through the aortic wall and separate. The treatment of aortic dissection is relatively mature. Stanford type A aortic dissection is mainly treated by surgery, and type B is mainly treated by endovascular treatment. Although the treatment guidelines have been well established, the mortality rate of patients with aortic dissection is very high when not treated, increasing by 1-3% per hour, up to 21% within 24 hours, and up to 74% within 1 week. The mortality rate of patients with Stanford type A aortic dissection after surgery due to various complications is as high as 25%, and Stanford type B also has the risk of death due to postoperative vascular tearing or distal dissection rupture. Early identification of aortic dissection and high-risk patients, and maximizing the time from symptom onset to receiving appropriate treatment are the key to improving patient survival.

[0003] The gold standard for diagnosing aortic dissection is CTA. Combined with the use of imaging techniques, the use of highly sensitive and specific biomarkers to quickly diagnose aortic dissection before the patient arrives at a hospital with surgical capabilities will greatly improve survival rates. And pre-hospital assessment of the patient's prognosis, early identification of patients with serious consequences, and attention during the diagnosis and treatment process will also improve the patient's outcome. However, most grassroots hospitals do not have equipment such as CTA, magnetic resonance angiography (MRA), and transthoracic ultrasound, which increases the difficulty and complexity of diagnosing aortic dissection.

[0004] There are several potential biomarkers in the prior art for acute aortic dissection, including CTRP-9, C-reactive protein, D-dimer, etc. For example, the Chinese invention patent with application number CN201910839515.3 screens acute aortic dissection by measuring CTRP-9 and D-dimer concentrations in plasma. However, the repeatability, specificity, and sensitivity of these potential biomarkers in an expanded population remain unclear, and there are no specific biomarkers associated with the prognosis of patients with aortic dissection.

[0005] Therefore, there is an urgent need to identify reliable early biomarkers of aortic dissection to assist in the diagnosis of whether patients with chest pain have aortic dissection, as well as predictive technology for the clinical prognosis of confirmed patients, to guide the clinical diagnosis and treatment process. Summary of the invention

[0006] In order to overcome the shortcomings of the above-mentioned technologies, the purpose of the present invention is to provide a biomarker, a kit, a system and its application related to aortic dissection, to explore the biomarkers related to aortic dissection and to construct a model and system for the diagnosis of aortic dissection, so as to achieve accurate and rapid diagnosis of aortic dissection.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0008] A biomarker associated with aortic dissection, the biomarker comprising a first biomarker and / or a second biomarker; the first biomarker is used for diagnosing aortic dissection, and the first biomarker comprises RCN1, SERPINA3 and CPN1; the second biomarker is used for predicting the clinical mortality risk of patients with aortic dissection, and the second biomarker comprises SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

[0009] The information of the above biomarkers can be found in the Uniprot database (https: / / www.uniprot.org).

[0010] The present invention also provides use of a first biomarker or a detection reagent thereof in preparing a product for diagnosing aortic dissection, wherein the first biomarker includes RCN1, SERPINA3 and CPN1.

[0011] The present invention also provides a kit for diagnosing aortic dissection, the kit comprising a reagent for detecting the expression level of a first biomarker in a sample; the first biomarker includes RCN1, SERPINA3 and CPN1; the sample is blood or plasma of a subject.

[0012] The present invention also provides the use of a second biomarker or a detection reagent thereof in the preparation of a product for predicting the clinical mortality risk of patients with aortic dissection, wherein the second biomarker includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

[0013] The present invention also provides a kit for predicting the clinical mortality risk of patients with aortic dissection, the kit comprising a reagent for detecting the expression level of a second biomarker in a sample; the second biomarker includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5; and the sample is the blood or plasma of the subject.

[0014] The present invention also provides a method for constructing an aortic dissection diagnosis model, comprising the following steps:

[0015] The plasma protein data of healthy people are obtained to construct a healthy sample group, and the plasma proteome data of patients with aortic dissection are obtained to construct a disease sample group; the disease sample group includes the plasma proteome data of patients with aortic dissection whose clinical outcome is death and the sample group data with good prognosis; each sample in the healthy sample group and the disease sample group includes the expression level of each protein feature in the plasma proteome and the annotation information of the auxiliary examination of whether the corresponding subject is diagnosed;

[0016] The healthy sample group and the disease sample group are randomly divided into a training set and a test set;

[0017] The training set is subjected to differential protein analysis, and differential proteins are screened using t-test and P value, so as to obtain protein features whose expression level difference between the plasma proteome of healthy people and patients with aortic dissection satisfies P<0.05 as the optimal marker combination for diagnosing aortic dissection; the optimal marker combination for diagnosing aortic dissection includes RCN1, SERPINA3 and CPN1;

[0018] Taking the expression level of the optimal marker for diagnosing aortic dissection of each sample as feature information, using training set and test set data, and constructing an aortic dissection diagnosis model based on a machine learning algorithm;

[0019] The aortic dissection diagnostic model is used to obtain risk parameters for aortic dissection in a subject based on the expression level of the optimal marker combination for diagnosing aortic dissection.

[0020] The present invention also provides an aortic dissection diagnostic model, which is constructed by the above method for constructing an aortic dissection diagnostic model. The risk parameter of aortic dissection in a subject is obtained by inputting the optimal marker combination into the aortic dissection diagnostic model for analysis.

[0021] The present invention also provides a method for constructing a clinical mortality risk prediction model for patients with aortic dissection, comprising the following steps:

[0022] The plasma proteome data of patients with aortic dissection whose clinical outcome was good were obtained as the sample group with good prognosis, and the plasma proteome data of patients with aortic dissection whose clinical outcome was death were obtained as the sample group with poor prognosis; wherein each sample includes the clinical characteristic information of the corresponding patient, the expression level of each protein feature in the plasma proteome, and the annotation information of whether the corresponding patient died;

[0023] The good prognosis sample group and the poor prognosis sample group are randomly divided into a training set and a test set;

[0024] The training set is subjected to differential protein analysis, and differential proteins are screened using t-test and P value, so as to obtain protein features in the plasma proteome whose expression level difference satisfies P<0.05 as the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection; the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5;

[0025] The expression level and clinical characteristic information of the optimal marker for predicting the clinical mortality risk of patients with aortic dissection of each sample are used as characteristic information, a clinical mortality risk prediction model for patients with aortic dissection is constructed based on the training set using a machine learning algorithm, and the optimal clinical feature combination is determined based on the feature importance of each clinical feature in the predictive performance of the clinical mortality risk prediction model for patients with aortic dissection;

[0026] Using the expression level of the optimal marker for predicting the clinical mortality risk of patients with aortic dissection and the information of the optimal clinical characteristics of each sample in the test set data to evaluate the predictive performance of the clinical mortality risk prediction model for patients with aortic dissection;

[0027] The clinical mortality risk prediction model for patients with aortic dissection is used to obtain the risk parameters of clinical mortality in patients with aortic dissection based on the expression level of the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection and the optimal clinical feature combination.

[0028] Furthermore, the clinical characteristic information includes gender, age, CTA, BMI, BaPWV, and clinical outcome.

[0029] As a preferred solution, the optimal clinical feature combination includes gender, age, CTA, BMI, and BaPWV. Combining the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection with the optimal clinical feature combination for predicting the clinical mortality risk of patients with aortic dissection is more accurate than using the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection alone.

[0030] Furthermore, the machine learning algorithm is selected from any one of the following algorithms: logistic regression algorithm, linear regression algorithm, random forest algorithm, neural network algorithm, support vector machine algorithm, Bayesian classification algorithm, gradient boosting algorithm, K nearest neighbor algorithm and decision tree algorithm.

[0031] The present invention also provides a clinical death risk prediction model for patients with aortic dissection, which is constructed by the above method for constructing a clinical death risk prediction model for patients with aortic dissection. The clinical death risk parameters of patients with aortic dissection are obtained by inputting the optimal marker for clinical death risk prediction of patients with aortic dissection into the clinical death risk prediction model for patients with aortic dissection for analysis.

[0032] The present invention also provides a system for diagnosing aortic dissection or predicting the clinical mortality risk of a patient with aortic dissection, comprising a processor and a display;

[0033] The processor is configured to predict the risk of aortic dissection in a subject or the clinical risk of clinical death in aortic dissection patients using an aortic dissection diagnostic model or a clinical risk of death prediction model for aortic dissection patients based on the expression level of the biomarker associated with aortic dissection; the display is used to present the risk parameter of aortic dissection in the subject or the clinical risk of death predicted by the processor;

[0034] The biomarker associated with aortic dissection is the biomarker associated with aortic dissection described above.

[0035] Furthermore, the display also presents diagnosis and treatment recommendations corresponding to the risk parameters.

[0036] Furthermore, when the system uses the aortic dissection diagnostic model to predict the risk of aortic dissection in a subject, the prediction is made based on the expression level of the first biomarker.

[0037] Furthermore, when the system uses the clinical mortality risk prediction model for patients with aortic dissection to predict the clinical mortality risk of patients with clinical aortic dissection, it is based on the expression level of the second biomarker and clinical characteristic information; the clinical characteristic information includes computed tomography angiography (CTA), body mass index (BMI), brachial-ankle pulse wave velocity (BaPWV), gender, and age.

[0038] Furthermore, the system comprises a liquid chromatography device and a mass spectrometry device, wherein the liquid chromatography device is used to extract proteins from the collected plasma sample of the subject, and the mass spectrometry device is used to quantitatively analyze the extracted proteins.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention provides biomarkers, kits, systems and applications related to aortic dissection, mines biomarkers related to aortic dissection and constructs a model and system for diagnosing aortic dissection, and achieves accurate and rapid diagnosis of aortic dissection and accurate prediction of the clinical mortality risk of patients with aortic dissection by screening out 10 protein markers, with the advantages of low cost, simple method, high accuracy and high efficiency.

[0041] The present invention establishes an association of 10 protein features with the diagnosis and prediction of clinical outcomes of subjects with aortic dissection and a corresponding prediction model. The acquisition of these protein feature information and clinical feature information is low-invasive for the subject, wherein the protein feature information is obtained by using the subject's plasma sample (usually collected and prepared in other medical procedures for aortic dissection) by means of proteomics analysis, mass spectrometry analysis, etc., and the clinical features should have been obtained during the subject's standardized diagnosis and treatment and follow-up process. Therefore, the feature information used in the present invention has very good compatibility with existing aortic dissection-related medical procedures. In addition, the negative predictive value of the aortic dissection diagnosis model and the aortic dissection patient clinical death risk prediction model of the present invention are both close to 95%, the area under the AUC curve is close to or exceeds 0.9, and the specificity, sensitivity, and accuracy are all high, which can efficiently and accurately screen patients with aortic dissection and patients with a high risk of clinical death, especially the situation of non-dissection patients can be accurately excluded, unnecessary treatment and waste of medical resources can be reduced, and the psychological pressure of these patients on the disease can be alleviated. In addition, the present invention can perform efficient and accurate clinical outcome predictions for patients with aortic dissection as early as possible without the need for additional invasive examinations, and monitor various indicators more closely, regularly, and frequently during treatment to prevent serious cardiovascular events or even death, and guide clinical decision-making, form individualized and accurate diagnosis and treatment plans, and improve prognosis.

[0042] The present invention screens out patients with aortic dissection and those with poor clinical outcomes in a relatively non-invasive manner only through the blood markers of the subjects, or through the blood markers combined with clinical indicators, so as to implement more precise treatment policies, including accurately excluding patients with good prognosis in the healthy population or the diseased group. For patients with aortic dissection who are at high risk of adverse outcomes such as postoperative death, less invasive surgical methods can be carefully selected during clinical decision-making, and the predictive indicators and other clinical review items provided by the present invention can be regularly monitored at a higher frequency to prevent these patients from suffering unnecessary cardiovascular risk events or even death.

[0043] The present invention successfully screened out plasma-based protein biomarkers that can accurately screen patients with aortic dissection and predict the risk of death in patients with aortic dissection. This will also help to explore the pathogenesis of this disease and is of great significance for the efficient, accurate and low-cost diagnosis of patients with aortic dissection. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of some components of an aortic dissection diagnosis system and a clinical mortality risk prediction system for aortic dissection patients according to an embodiment of the present invention;

[0045] Figure 2 A schematic diagram of the importance of three protein features as the optimal marker combination for diagnosing aortic dissection screened by the embodiment of the present invention (part A in the figure) and a schematic diagram of the performance results of the aortic dissection diagnosis model on the test set (parts B and C in the figure);

[0046] Figure 3 A schematic diagram of the importance of seven protein features for predicting the clinical mortality risk of patients with aortic dissection screened in an embodiment of the present invention (part D in the figure) and a schematic diagram of the performance results of the clinical mortality risk prediction model for patients with aortic dissection on a test set (parts E and F in the figure). DETAILED DESCRIPTION

[0047] In order to better explain the present invention, the main contents of the present invention are further explained below in conjunction with specific embodiments, but the contents of the present invention are not limited to the following embodiments.

[0048] Example 1: Biomarkers associated with aortic dissection and screening process thereof

[0049] Step 1: Experimental batch design and quality control sample preparation

[0050] Experimental design is performed for clinical cohort samples, taking into account the different clinical characteristics of the population, and using any applicable proteomic big data design and quality control tools to make the samples as evenly dispersed as possible to minimize batch effects. In addition, QC (quality control) samples can be prepared while processing samples, that is, mixed samples of equal volume of experimental samples, and the system stability can be evaluated throughout the experiment.

[0051] Step 2: Protein extraction and peptide digestion

[0052] The plasma sample to be tested was thawed at 4°C. An appropriate amount of sample was taken and added to a new 1.5mL EP tube, and 50mM ammonium bicarbonate solution was added to a total volume of 100μL. The sample was heated at 95°C for 3min to denature the protein, and then cooled to room temperature and then pancreatic protease was added for enzymatic hydrolysis. The enzymatic hydrolysis was carried out at 37°C for 16h, and the enzymatic hydrolysis product was extracted and freeze-dried. Desalting was then performed. After the peptide was freeze-dried, 100μL of 0.1% formic acid solution was added to re-dissolve it, and the sample was injected at a ratio of 2% for detection.

[0053] The samples were separated using the nanoliter flow rate high performance liquid system EASY-nLC 1200. Mobile phase A was 0.1% formic acid in water, and mobile phase B was 0.1% formic acid in acetonitrile (the volume ratio of acetonitrile to formic acid was 4:1). First, the loading column and the analytical column were equilibrated with 100% mobile phase A, and then the enzymatic peptides of the sample were transported to the loading column (2cm, ID100μm, 3μm, C18) by the autosampler, and then separated by the analytical column (15cm, ID150μm, 1.9μm, C18) at a flow rate of 600nL / min.

[0054] Step 3: Mass spectrometry data acquisition and analysis

[0055] 1. Mass spectrometry analysis: After the sample is separated by liquid chromatography in step 2, it is analyzed by mass spectrometry using an HFX mass spectrometer. The detection method is positive ion, the parent ion scanning range is 300-1400 m / z, the primary mass spectrometry resolution is 60,000 at 200 m / z, the AGC (Automatic gain control) target is 3e6, and the Maximum IT is 20 ms. The mass-to-charge ratio of the peptide and the peptide fragments is collected according to the following method: 30 DIA Scans are collected after each full scan, using the HCD fragmentation mode, the Normalized Collision Energy is 27%, the Isolation window varies according to the isolation window, and the secondary mass spectrometry resolution is 1,5000 at 200 m / z.

[0056] 2. Quality control analysis: From the correlation data of plasma proteome QC samples, it can be seen that the median correlation of plasma proteome QC samples is 98%, indicating that the experimental data has high consistency and repeatability.

[0057] 3. Protein difference analysis: The condition P value was set to be less than 0.05, and the differential proteins between healthy subjects, patients with good prognosis of aortic dissection, and patients with poor prognosis of aortic dissection were screened, and the screened proteins were subjected to the next step of targeted verification.

[0058] Step 4: Targeted proteome analysis

[0059] First, the quantitative information of the target protein set was normalized (z-score). Then, the pheatmap R package was used to classify the samples and protein expression levels at the same time (distance algorithm: Euclidean, connection method: Average linkage).

[0060] The raw data of mass spectrometry analysis is a RAW file, and the iProteome one-stop data analysis cloud platform is used for library search qualitative and quantitative analysis. The library search parameters are set as follows: Enzyme: Trypsin; Fixed modifications: Carbamidomethylation (C); Variable modifications: Methionine oxidation, acetyl (protein N-terminus); Missed cleavages: 2Peptide; Mass Tolerance: 20ppm; Mass Tolerance: 0.05Da.

[0061] In addition, the use of a liquid chromatography-mass spectrometer with high mass accuracy and high resolution can maintain good mass deviation during data acquisition, and ultimately obtain high-quality MS1 and MS2 spectra. The mass deviation of all identified peptides is mainly distributed within 10ppm, indicating that the identification results are accurate and reliable. The spectrum data is then analyzed in combination with the Mascot search engine to obtain the score of each MS2 spectrum. Combined with the distribution of the obtained excellent peptide scores, it is further demonstrated that mass spectrometry instruments can produce high-quality experimental data. At the same time, DIA data uses Peptide FDR≤0.05 as the screening criterion in qualitative analysis.

[0062] DIA (data-independent acquisition) refers to a data-independent scanning mode, which is a holographic mass spectrometry data acquisition mode based on the electrostatic field orbital trap Orbitrap. After the primary mass spectrometry detection, DIA will fragment the parent ions within a specific mass-to-charge ratio range, collect the corresponding fragment ions, and quickly scan all the fragment ions in the adjacent parent ion windows in sequence, so as to perform qualitative and quantitative analysis of proteins.

[0063] Based on the above steps one to four, the expression levels of various proteins serving as biomarkers can be detected from the plasma samples of the subjects, and the differential proteomes between healthy people and patients with aortic dissection, as well as the differential proteomes between patients with aortic dissection with good and poor prognosis can be screened out, and the expression levels of each protein feature in the differential proteome can also be obtained.

[0064] Among them, the differential proteome between healthy people and patients with aortic dissection is the first biomarker, including RCN1, SERPINA3 and CPN1; the differential proteome between patients with good prognosis and poor prognosis is the second biomarker, including SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

[0065] The above ten biomarkers are all proteins. Currently, the main functions of various proteins are considered to be as follows in clinical practice.

[0066] RCN1: Human reticulin 1, as a potential regulator, plays a role in regulating calcium-dependent activities within the lumen of the endoplasmic reticulum (ER) or post-ER compartments. RCN1 Protein, Human (P. pastoris, His) is a recombinant RCN1 protein expressed from P. pastoris with an N-His tag, and the specific mechanism by which it exerts regulatory effects on calcium-dependent processes remains to be fully elucidated, which has prompted further investigation of its functional significance in these cellular compartments.

[0067] SERPINA3: α-1 antichymotrypsin, cytochrome P4502E1 (CYP2E1) can be induced by compounds such as ethanol, and participates in the metabolic activation of various pro-carcinogens (N-nitrosamines, aniline, vinyl chloride, ethyl carbamate), inducing the occurrence of tumors. Receptor tyrosine kinase participates in the development and maturation of the central and peripheral nervous systems by regulating the proliferation, differentiation and survival of sympathetic nerves and neurons.

[0068] CPN1: Calcium-dependent membrane-bound protein Copine protein 1, is a calcium-dependent phospholipid-binding protein that actively participates in calcium-mediated intracellular processes. Copine-1 participates in the TNF-α receptor signaling pathway and exhibits calcium-dependent phospholipid-binding properties. In addition, it acts in a calcium-independent manner through the AKT-dependent signaling cascade to induce neurite outgrowth, thereby playing a key role in the differentiation of neural progenitor cells. Copine-1 may also promote the recruitment of target proteins to the cell membrane in a calcium-dependent manner and participate in membrane trafficking.

[0069] SYTL1: Synaptotagmin 1, Synaptotagmin 1 (SYT1) is an integral membrane protein of synaptic vesicles and is thought to act as a Ca(2+) sensor during vesicle trafficking and exocytosis. The calcium it binds to synaptotagmin-1 is involved in triggering neurotransmitter release at synapses. It plays a vital role in the nervous system, especially in the docking and fusion of synaptic vesicles with the presynaptic membrane. This role is important for the maintenance of mature neurons and neurotransmitter release.

[0070] B2M: β2-microglobulin (β2-MG), is the light chain of human leukocyte antigen (HLA) class I antigen, a low molecular weight serum globulin produced by lymphocytes, platelets, polymorphonuclear leukocytes, and exists on the surface of various cells outside mature red blood cells and placental trophoblast cells. β2-MG is widely present in plasma, urine, cerebrospinal fluid, saliva and colostrum. Clinically, blood and urine β2-MG determination can evaluate renal tubular function.

[0071] LYVE1: Lymphatic endothelial cell hyaluronic acid receptor 1, Lymphatic endothelial receptor-1 is a close relative of the leukocyte receptor CD44, is the main receptor of lymphatic endothelial cell hyaluronic acid (HA), and is a common marker for distinguishing blood vessels and lymphatic vessels. It is a lymphatic docking receptor of dendritic cells, selectively binds to their surface HA glycocalyx, regulates the entry of peripheral lymphatic vessels and migration to downstream lymph nodes, thereby activating immunity.

[0072] TTR: Transthyretin, a homotetrameric protein mainly synthesized by the liver and choroid plexus, is a protein that transports thyroxine and retinol-binding retinol in plasma and cerebrospinal fluid. It exists in plasma, serum and cerebrospinal fluid and is mainly synthesized by the liver and choroid plexus.

[0073] PRSS2: Serine protease 2, is a serine protease, a member of the protease family. This enzyme plays an important role in mammals, especially in digestion, coagulation and complement systems. The activation of serine protease is achieved through the change of a group of amino acid residues in its active center, one of which must be serine, which is the origin of its name.

[0074] CSTF1: Cleavage stimulatory factor subunit 1, this gene encodes one of the three subunits that combine to form cleavage stimulatory factor (CSTF). CSTF is involved in polyadenylation and 3' end cleavage of pre-mRNA. Similar to mammalian G protein β subunits, this protein contains transducin-like repeat sequences. In addition, transcript variants with different 5'UTRs but encoding the same protein have been found for this gene. The CSTF1 gene is widely expressed in multiple tissues, including testis, lymph nodes, etc., and is also expressed in 25 other tissues.

[0075] H2BC5: H2B clustering histone 5, histone is the basic nuclear protein that constitutes the nucleosome structure of the chromosome fiber of eukaryotic organisms. The nucleosome consists of about 146bp of DNA wrapped around the histone octamer, which is composed of each of the four core histones (H2A, H2B, H3 and H4). This gene has no introns and encodes a replication-dependent histone, which is a member of the histone H2B family.

[0076] Example 2: Aortic dissection diagnosis model and aortic dissection patient clinical mortality risk prediction model and construction method thereof

[0077] 1. Aortic dissection diagnosis model

[0078] The aortic dissection diagnostic model of this embodiment has an input parameter of the expression level of the first biomarker in Example 1, and an output parameter of the risk parameter of aortic dissection in the subject. The method for constructing the model includes the following steps:

[0079] 1) Obtain plasma protein data of a group of healthy people as a healthy sample group, and obtain plasma proteome data of a group of patients with aortic dissection as a disease sample group, including sample groups with good prognosis and poor prognosis in clinical outcomes, and each sample in the healthy sample group and the disease sample group includes the expression level of each protein feature in the plasma proteome and the corresponding annotation information of the patient's auxiliary examination for diagnosis.

[0080] 2) The healthy group and the aortic dissection (good prognosis + poor prognosis) patient sample group are divided into a training set and a test set.

[0081] 3) Using t-test and P value to screen features, and determine the protein features in the plasma proteome that satisfy the expression level difference satisfying P<0.05, as the optimal marker combination for predicting whether the subject suffers from aortic dissection; this step is the screening process in Example 1, and the optimal marker combination for predicting the risk of aortic dissection in the subject is the first biomarker in Example 1.

[0082] 4) Based on the expression level of the optimal marker combination of each sample in the training set, a random forest algorithm is used to construct and train an application prediction model, and the application prediction model is tested based on the test set to evaluate the prediction performance, and the application prediction model whose prediction performance exceeds the performance index threshold is used as an aortic dissection diagnosis model to provide a risk prediction result for aortic dissection based on the expression level of the optimal marker combination in the subject's plasma.

[0083] 2. Clinical mortality risk prediction model for patients with aortic dissection

[0084] The clinical mortality risk prediction model for patients with aortic dissection in this embodiment has an input parameter of the expression level of the second biomarker in Example 1, and an output parameter of the clinical mortality risk parameter for patients with aortic dissection, and the construction method thereof comprises the following steps:

[0085] 1) Obtain plasma proteome data of a group of patients with aortic dissection with good clinical outcomes as a good prognosis sample group, and obtain plasma proteome data of a group of patients with aortic dissection with poor clinical outcomes (death) as a poor prognosis sample group, wherein each sample in the good prognosis sample group and the poor prognosis sample group includes the expression level of each protein feature in the plasma proteome and the annotation information of whether the corresponding patient has died; obtain clinical characteristic information of the patient corresponding to each sample, wherein the clinical characteristic information at least includes gender, age, CTA, BMI, BaPWV, and clinical outcome; wherein, a good clinical outcome refers to a patient who is discharged from the hospital after the condition improves during clinical treatment.

[0086] 2) Dividing the aortic dissection sample group with good prognosis and the aortic dissection sample group with poor prognosis into a training set and a test set.

[0087] 3) Based on the training set, t-test and P-value are used to screen features, and each protein feature in the plasma proteome that satisfies the expression level difference satisfying P<0.05 is determined as the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection; this step is the screening process in Example 1, and the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection is the second biomarker in Example 1.

[0088] 4) Based on the expression level and clinical feature information of the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection in each sample in the training set, a random forest algorithm is used to construct and train an application prediction model, and the optimal clinical feature combination is determined based on the feature importance of each clinical feature in the predictive performance of the application prediction model; the application prediction model is tested to evaluate the prediction performance, and the application prediction model whose prediction performance exceeds the performance index threshold is used as a clinical mortality risk prediction model for patients with aortic dissection to give a prediction result about the mortality risk of aortic dissection. The prediction result is expressed by a risk parameter, including but not limited to a percentage, such as "the probability of suffering from aortic dissection: 90%", etc. The actual risk parameter can also be other numerical ranges, and the content and method of presentation can also be set as needed, which will not be repeated here.

[0089] Among them, the optimal clinical feature combination is to select the top five features as the optimal clinical feature combination based on clinical experience after ranking the features by importance.

[0090] In the above modeling method, the preset condition of the difference in expression level is that the P value is less than 0.05, that is, P-value<0.05, where P-value is the P value. The specific calculation method is known to those skilled in the art and will not be repeated here.

[0091] The above method of constructing and training an application prediction model using a random forest algorithm includes the following steps:

[0092] 1) Data cleaning and filling;

[0093] 2) The data is randomly divided into training set and test set.

[0094] 3) Perform model training and parameter optimization in the training set. First, perform random search to explore multiple parameter combinations, and then select the model with the best performance, ntree=500, and the minimum size of the termination node is 1. Three-fold cross validation is used to ensure that the model has good generalization ability.

[0095] The number of trees ntree in the optimal parameters is set to 500, and each tree randomly extracts samples and features for learning and construction. Due to the randomness of samples and features, random forests are not prone to overfitting. The introduction of this randomness helps to improve the generalization ability of the model. Each internal node represents a "test" of an attribute (for example, aortic dissection or not), each branch represents the result of the test, and each leaf node represents a class label (decided after calculating all attributes). Nodes without child nodes are leaves. The minimum size of the terminal node is 1, that is, each terminal node has at least one sample, and the model can explore the patterns in the data more comprehensively.

[0096] After the above-mentioned application prediction model based on the random forest algorithm is trained, it is necessary to further test the application prediction model based on the test set to evaluate the prediction performance, and use the application prediction model whose prediction performance exceeds the performance index threshold to give a prediction result on the risk of aortic dissection or a prediction result on the clinical mortality risk of patients with aortic dissection based on the expression level of the above-mentioned biomarkers.

[0097] The performance indicator may be AUC or other performance indicators, or other performance indicators and their corresponding thresholds may be selected according to actual conditions.

[0098] The performance indicators of the above-mentioned aortic dissection diagnosis model (model 1) and the clinical mortality risk prediction model for aortic dissection patients (model 2) on the test set are shown in Table 1. Figure 2 (B, C) and Figure 3 As shown in (E, F), the aortic dissection diagnostic model only uses the three protein features of the first biomarker, while the aortic dissection patient clinical mortality risk prediction model uses the second biomarker and the optimal clinical feature information to predict the risk of aortic dissection in the subject and whether the aortic dissection-affected limb will die.

[0099] Table 1: Performance indicators of Model 1 and Model 2 on the test set

[0100]

[0101] The area under the curve (AUC) of the AUC curve is the area covered under the receiver operating characteristic curve. The AUC value is not greater than 1. The larger the AUC value, the better the model performance. When 0.5 < AUC < 1, it indicates that the model prediction performance is better than random prediction and the performance is good.

[0102] As can be seen from Table 1, whether it is the aortic dissection diagnosis model or the clinical death risk prediction model for aortic dissection patients, the area under the AUC curve is close to or exceeds 0.9, and the specificity, sensitivity, and accuracy are all relatively high, indicating that the prediction performances of both models are very good and the prediction results are highly credible. Therefore, unnecessary waste of medical resources and side effects of patients and anxiety about the disease caused by overtreatment can be avoided.

[0103] In addition, through the random forest algorithm, the feature importance of each protein in the plasma protein data is analyzed, and the analysis results are respectively as Figure 2 (A) and Figure 3 (D) shown. In the figure, the importance of protein features decreases sequentially from top to bottom. It can be Figure 2 seen that the importance rankings of the three protein features as the optimal biomarker combination for diagnosing aortic dissection are the most prominent. It can be Figure 3 seen that the importance rankings of the seven protein features as the optimal biomarker combination for predicting the clinical death risk of aortic dissection patients are the most prominent. The feature importance analysis results of the random forest algorithm further verify the screening results of the biomarkers in Example 1.

[0104] Example 3: Aortic Dissection Diagnosis System and Clinical Death Risk Prediction System for Aortic Dissection Patients I. Aortic Dissection Diagnosis System

[0105] As Figure 1 shown, the aortic dissection diagnosis system 001 includes a first processor 002 and a first display 003. Among them, the first processor 002 predicts the risk parameters of the subject suffering from aortic dissection by obtaining the expression levels of RCN1, SERPINA3, and CPN1 of the subject, and uses the aortic dissection diagnosis model in Example 2, and enables the first display 003 to present the predicted risk parameters of the subject suffering from aortic dissection and the corresponding diagnosis and treatment suggestions.

[0106] The risk parameters can be expressed as a percentage, such as "probability of suffering from aortic dissection: 90%". Specifically, according to the clinical experience of medical workers, two risk parameter thresholds, low and high, can be set for the risk parameters of the subject suffering from aortic dissection. The low threshold is 20%, and the high threshold is 50%. The diagnosis and treatment suggestions are as follows:

[0107] If the predicted risk parameter is less than the low threshold, the risk of aortic dissection is considered to be very low and the patient can be treated according to other clinical chest pain diseases. Some patients (e.g., young patients with no smoking history, normal BaPWV, normal BMI and no family genetic diseases) can be exempted from invasive imaging examinations.

[0108] When the low threshold < the predicted risk parameter ≤ the high threshold, further exploration is necessary in combination with other medical methods;

[0109] When the predicted risk parameter is greater than the high threshold, it means that the subject is at a very high risk of aortic dissection, and it is recommended to conduct detailed and other risky price difference items in advance, such as CTA, cardiac ultrasound, DSA or MRI, etc.; if the imaging examination reveals abnormal phenomena such as double-chamber changes in the aorta, aortic dilatation or intramural hematoma, a value higher than the high threshold can also help clinicians make a tendency diagnosis of aortic dissection. The present invention does not specifically limit the specific setting of the threshold and the specific application of the prediction results by medical workers.

[0110] 2. Clinical mortality risk prediction system for patients with aortic dissection

[0111] The clinical death risk prediction system 004 for patients with aortic dissection includes a second processor 005 and a second display 006. Before imaging finds that the subject has a ruptured aortic dissection or has a poor clinical outcome, the clinical death risk prediction model for patients with aortic dissection in Example 2 is used to predict the risk parameters of clinical death of patients with aortic dissection, and based on the predicted risk parameters of the subject's death, the second display 006 presents the predicted risk parameters and corresponding diagnosis and treatment recommendations.

[0112] For subjects diagnosed with aortic dissection, two risk parameter thresholds, low and high, were set.

[0113] When the predicted risk parameter is lower than the low threshold (the specific value can be set to 20%, for example), the risk of death of the subject can be considered very low, and the subject can be treated according to the conventional treatment process, and some patients (for example, young patients with no smoking history, normal BaPWV, normal BMI and no family genetic diseases) can be exempted from repeated imaging examinations, etc.;

[0114] In cases where the predicted risk parameter is between the low and high thresholds, further exploration is necessary in combination with other medical methods;

[0115] When the predicted risk parameter is higher than the high threshold (the specific value can be set to 50%, for example), it means that the risk of death of the subject is extremely high. The prediction result is for the reference of the doctor, so that he can judge whether it is necessary to carry out detailed and other risky price difference items in advance (for example, CTA, cardiac ultrasound, DSA or MRI, etc.), and recommend close monitoring of serious cardiovascular events such as death; for imaging findings of severe lung infection, neurological complications, double-chamber changes in other parts of the aorta, or aortic dilatation or surgical anastomotic hematoma during the examination, it is higher than the high threshold. It can also help clinicians make a tendency prediction that the prognosis of patients with aortic dissection is extremely poor or even death. The present invention does not specifically limit the specific setting of the threshold and the specific application of the prediction results by medical workers.

[0116] Among them, close monitoring can further include: 1) shortening the time interval for postoperative review of patients with aortic dissection who have not yet experienced severe pain symptoms in other parts of the body, so as to detect possible aortic dissection rupture early; 2) completely eliminating the entire aortic rupture and false lumen during aortic dissection surgery to avoid the recurrence of serious cardiovascular events or even death; 3) for patients with unsatisfactory clinical prognosis, if the predicted risk parameters are higher than the first threshold, it is recommended that the clinic consider the necessity of more stringent monitoring of various biochemical indicators and cardiovascular function after surgery.

[0117] The above-mentioned second display 006 can be the first display 003. When the first processor 002 processes the risk of the subject's illness and gives that the subject's illness risk parameter is higher than a high threshold, the second processor 004 is used to predict the subject's death risk, so that the first display 003 simultaneously displays the illness risk parameters and the patient's clinical death risk parameters and corresponding diagnosis and treatment opinions.

[0118] The aortic dissection diagnosis system and the clinical mortality risk prediction system for patients with aortic dissection may also include liquid chromatography equipment and mass spectrometry analysis equipment to obtain samples of plasma collected from subjects, analyze and quantify them through targeted proteomics (MRMHR), and detect the expression levels of biomarkers therefrom.

[0119] In some embodiments, the plasma used to obtain the expression level of the subject's biomarker can be obtained simultaneously when the subject undergoes a routine blood test, and the amount of plasma required is small, and the subject's blood loss or blood collection trauma is not increased. No additional blood collection is required, which to a certain extent reduces the number of unnecessary CTA and MRI examinations for low-risk patients and the number of invasive examinations for some low-risk patients, and reduces the psychological pressure, contrast agent risks, radiation exposure and body invasion of the subjects.

[0120] Example 4

[0121] Use of a first biomarker or a detection reagent thereof in the preparation of a product for diagnosing aortic dissection, wherein the first biomarker includes RCN1, SERPINA3 and CPN1.

[0122] Example 5

[0123] A kit for diagnosing aortic dissection, the kit comprising a reagent for detecting the expression level of a first biomarker in a plasma sample from a subject; the first biomarker comprises RCN1, SERPINA3 and CPN1.

[0124] Example 6

[0125] The use of a second biomarker or a detection reagent thereof in the preparation of a product for predicting the clinical mortality risk of patients with aortic dissection, wherein the first biomarker includes RCN1, SERPINA3 and CPN1.

[0126] Example 7

[0127] A kit for predicting the clinical mortality risk of patients with aortic dissection, the kit comprising a reagent for detecting the expression level of a second biomarker in a plasma sample from a subject; the second biomarker includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

[0128] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A biomarker associated with aortic dissection, characterized in that: The biomarkers include a first biomarker and / or a second biomarker; the first biomarker is used to diagnose aortic dissection, and the first biomarker includes RCN1, SERPINA3 and CPN1; the second biomarker is used to predict the clinical mortality risk of patients with aortic dissection, and the second biomarker includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

2. Use of a first biomarker or a detection reagent thereof in the preparation of a product for diagnosing aortic dissection, characterized in that: The first biomarkers include RCN1, SERPINA3 and CPN1.

3. A kit for diagnosing aortic dissection, characterized in that: The kit comprises a reagent for detecting the expression level of a first biomarker in a sample; the first biomarker comprises RCN1, SERPINA3 and CPN1; and the sample is blood or plasma of a subject.

4. Use of a second biomarker or a detection reagent thereof in the preparation of a product for predicting the clinical mortality risk of patients with aortic dissection, characterized in that: The second biomarkers include SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5.

5. A kit for predicting the clinical mortality risk of patients with aortic dissection, characterized in that: The kit comprises a reagent for detecting the expression level of a second biomarker in a sample; the second biomarker comprises SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5; and the sample is blood or plasma of a subject.

6. A method for constructing an aortic dissection diagnostic model, characterized in that: The following steps are involved: The plasma protein data of healthy people are obtained to construct a healthy sample group, and the plasma proteome data of patients with aortic dissection are obtained to construct a disease sample group; the disease sample group includes the plasma proteome data of patients with aortic dissection whose clinical outcome is death and the sample group data with good prognosis; each sample in the healthy sample group and the disease sample group includes the expression level of each protein feature in the plasma proteome and the annotation information of the auxiliary examination of whether the corresponding subject is diagnosed; The healthy sample group and the disease sample group are randomly divided into a training set and a test set; The training set is subjected to differential protein analysis, and differential proteins are screened using t-test and P value, so as to obtain protein features whose expression level difference between the plasma proteome of healthy people and patients with aortic dissection satisfies P<0.05 as the optimal marker combination for diagnosing aortic dissection; the optimal marker combination for diagnosing aortic dissection includes RCN1, SERPINA3 and CPN1; Taking the expression level of the optimal marker for diagnosing aortic dissection of each sample as feature information, using training set and test set data, and constructing an aortic dissection diagnosis model based on a machine learning algorithm; The aortic dissection diagnostic model is used to obtain risk parameters for aortic dissection in a subject based on the expression level of the optimal marker combination for diagnosing aortic dissection.

7. A diagnostic model for aortic dissection, characterized in that: The method for constructing an aortic dissection diagnostic model according to claim 6 is used to construct the model.

8. A method for constructing a clinical mortality risk prediction model for patients with aortic dissection, characterized by: The following steps are involved: The plasma proteome data of patients with aortic dissection whose clinical outcome was good were obtained as the sample group with good prognosis, and the plasma proteome data of patients with aortic dissection whose clinical outcome was death were obtained as the sample group with poor prognosis; wherein each sample includes the clinical characteristic information of the corresponding patient, the expression level of each protein feature in the plasma proteome, and the annotation information of whether the corresponding patient died; The good prognosis sample group and the poor prognosis sample group are randomly divided into a training set and a test set; The training set is subjected to differential protein analysis, and differential proteins are screened using t-test and P value, so as to obtain protein features in the plasma proteome whose expression level difference satisfies P<0.05 as the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection; the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection includes SYTL1, B2M, LYVE1, TTR, PRSS2, CSTF1 and H2BC5; The expression level and clinical characteristic information of the optimal marker for predicting the clinical mortality risk of patients with aortic dissection of each sample are used as characteristic information, a clinical mortality risk prediction model for patients with aortic dissection is constructed based on the training set using a machine learning algorithm, and the optimal clinical feature combination is determined based on the feature importance of each clinical feature in the predictive performance of the clinical mortality risk prediction model for patients with aortic dissection; Using the expression level of the optimal marker for predicting the clinical mortality risk of patients with aortic dissection and the information of the optimal clinical characteristics of each sample in the test set data to evaluate the predictive performance of the clinical mortality risk prediction model for patients with aortic dissection; The clinical mortality risk prediction model for patients with aortic dissection is used to obtain the risk parameters of clinical mortality in patients with aortic dissection based on the expression level of the optimal marker combination for predicting the clinical mortality risk of patients with aortic dissection and the optimal clinical feature combination.

9. A clinical mortality risk prediction model for patients with aortic dissection, characterized by: The method for constructing a clinical mortality risk prediction model for patients with aortic dissection as described in claim 8 is used to construct the model.

10. A system for diagnosing aortic dissection or predicting the clinical mortality risk of patients with aortic dissection, characterized by: Includes processor and display; The processor is configured to predict the risk of aortic dissection in a subject or the clinical risk of clinical death in aortic dissection patients using an aortic dissection diagnostic model or a clinical risk of death prediction model for aortic dissection patients based on the expression level of the biomarker associated with aortic dissection; the display is used to present the risk parameter of aortic dissection in the subject or the clinical risk of death predicted by the processor; The biomarker associated with aortic dissection is the biomarker associated with aortic dissection as described in claim 1.

Images