Use of Serum Biomarkers for Non-Disease Diagnosis

Abstract

The present invention discloses serum metabolic markers and their application in evaluating the stability of intracranial artery plaques, belonging to the field of biomedical technology. The serum biomarkers are selected from one or a combination of phosphatidylcholine PC(15:0 / 15:1), phosphatidylcholine PC(14:0 / 18:2), phosphatidylethanol PE(18:2 / 18:3), phosphatidylethanol PE 18:1e / 18:1, 5α-dihydrotestosterone, D-fructose 6-phosphate, phosphatidylcholine PC(14:0 / 18:2), sphingomyelin SM(d28:0 / 14:1), phosphatidylcholine PC(20:3 / 22:6), and phosphatidylcholine PC(20:3 / 20:4), providing a new direction for the treatment or intervention of atherosclerosis.

Description

Technical Field

[0001] The present invention belongs to the technical field of biomedicine, and relates to biological metabolic markers and their application in evaluating the stability of intracranial atherosclerotic plaques in patients with cerebral infarction. Background Art

[0002] A biomarker is a biochemical indicator that can mark changes or possible changes in the structure or function of a system, organ, tissue, cell, and subcellular structure. Research related to biomarkers is extremely popular in the scientific research field. It can balance clinical work and take into account basic research, and has very wide applications. Metabolomics is a newly emerging technology in 1990. It is an extension of genomics, transcriptomics, and proteomics, and is located at the most downstream of systems biology. Its purpose is to examine the change rules of metabolites after a biological system is perturbed by various internal and external environmental factors, reveal what is happening in the body, reflect the molecular results caused by all life phenomena, and reveal the biological processes and possible molecular mechanisms involved by screening phenotypic differences between different groups of different individuals. It is an effective method for screening various biomarkers.

[0003] There are a large number of metabolite species in the human body with a wide concentration range distribution. Currently, hyphenated techniques, namely liquid chromatography-tandem mass spectrometry (LC-MS / MS) techniques, are mostly used to improve the repeatability and resolution of chromatography. Metabolites are defined as small molecules, which are the intermediate products of numerous enzyme-catalyzed metabolic reactions occurring within cells.

[0004] Atherosclerotic disease is a complex metabolic disorder and one of the leading causes of death globally. Increasing evidence shows that intracranial plaque enhancement is closely related to acute ischemic stroke, and plaque enhancement is one of the key factors for atherosclerotic plaque vulnerability. However, there is currently little research on the metabolite changes in plaque enhancement. Gene-environment interactions may lead to the regulation of various metabolic pathways, thereby resulting in changes in the levels of various metabolites. Therefore, identifying metabolites helps to decipher the pathways that may be involved in the pathophysiology of atherosclerosis. With the emergence of the LC-MS / MS method, it is now possible to screen thousands of metabolites, thereby enabling the identification of potential disease metabolic markers.

[0005] Plaque enhancement is an independent risk factor for stroke recurrence, which also promotes the potential imaging marker that obvious plaque enhancement may be used to predict stroke recurrence. In this experiment, serum samples of populations with obvious and non-obvious enhancement of intracranial atherosclerotic plaques were studied. By using ultra-high performance liquid chromatography combined with high-resolution mass spectrometry technology, and a data processing mode combining multivariate statistical analysis with machine learning algorithms, statistically significant differential metabolites between the obvious and non-obvious plaque enhancement groups were found, so as to explore the molecular mechanism of obvious plaque enhancement. The discovery of biomarkers may contribute to the prediction of plaque stability and the risk of cerebral infarction. Flow chart of biomarker screening, such as Figure 1 。

[0006] Metabolomics data generally have a relatively small sample size and complex characteristics such as high dimension and high noise. The data analysis work is restricted to a certain extent. False positive results are prone to occur in the analysis by conventional methods, and specialized mathematics, bioinformatics, and statistics are required to process the data. Extracting valuable information and screening potential biomarkers are the main problems to be solved in metabolomics data processing.

[0007] Random forest (RF) is an ensemble classifier containing multiple decision trees. The algorithm advantage of RF lies in the visualization of sample differences, the judgment of the importance of features, with reliable classification prediction ability and high accuracy, not prone to overfitting, can screen high-dimensional data, and does not require prior dimensionality reduction processing, and has a high tolerance for noise and abnormal data. RF obtains the optimal parameters of the model by repeating 10-fold cross-validation. Similar to logistic regression in traditional statistical methods, the RF model can also effectively explain the dependent variable Y corresponding to the independent variable X. That is, when constructing the classification tree, the model uses the Bootstrap resampling technique, randomly samples with replacement, randomly selects independent and dependent variables, and creates different classification trees according to the importance scores of mean decrease accuracy and Gini index of mean decrease Gini, while ensuring that the branches and leaves of each tree can grow maximally. In the process of constructing the model, according to the importance of variables, the importance of predictive variables is sorted to eliminate some invalid variables, improve the classification accuracy of the model, and obtain potential differential metabolites.

[0008] Least absolute shrinkage and selection operator (LASSO) regression is a generalized linear regression model first proposed by Robert Tibshirani in 1996. By adding an L1 penalty function on the basis of least squares (L1

[0009] penalization), effectively solve the influence of collinearity, make the model more stable and have stronger generalization ability. By compressing the coefficients of some meaningless or less meaningful independent variables to 0 and including the variables with non-zero regression coefficients into the final model, according to 10-fold cross-validation and the optimal λ value, a simplified model is obtained to further narrow the range of target metabolites. If the sample size cannot be much larger than the number of variables and most variables have little relationship with the result, the least squares multiple regression will not obtain good prediction results. In these cases, the LASSO regression model may not only improve the prediction accuracy but also simplify the model through variable screening to determine important target metabolites. Summary of the Invention

[0010] The object of the present invention is to provide several serum metabolic markers and methods for using them to evaluate the stability of intracranial atherosclerotic plaques.

[0011] The application of serum biomarkers in non-disease diagnosis for evaluating the stability of intracranial artery plaques, wherein the serum biomarkers are selected from one or a combination of phosphatidylcholine PC(15:0 / 15:1), phosphatidylcholine PC(14:0 / 18:2), phosphatidylethanol PE(18:2 / 18:3), phosphatidylethanol PE 18:1e / 18:1, 5α-dihydrotestosterone, D-fructose 6-phosphate, phosphatidylcholine(14:0 / 18:2), sphingomyelin SM(d28:0 / 14:1), phosphatidylcholine PC(20:3 / 22:6), phosphatidylcholine PC(20:3 / 20:4).

[0012] A further improvement of the technical solution of the present invention lies in that: the serum biomarkers are a combination of 5α-dihydrotestosterone, D-fructose 6-phosphate and phosphatidylcholine PC(14:0 / 18:2).

[0013] A further improvement of the technical solution of the present invention lies in that: the serum biomarkers are from the serum of the subject.

[0014] A further improvement of the technical solution of the present invention lies in that: a decrease in the content of any one of the 6 serum metabolites, namely phosphatidylcholine PC(15:0 / 15:1), phosphatidylcholine PC(14:0 / 18:2), phosphatidylethanol PE(18:2 / 18:3), PE(18:1e / 18:1), 5α-dihydrotestosterone, D-fructose 6-phosphate, indicates that the intracranial artery plaques of the subject are likely to show significant enhancement and there is a risk of plaque instability.

[0015] A further improvement of the technical solution of the present invention lies in that an increase in the content of any one of the three serum metabolites, namely sphingomyelin SM (d28:0 / 14:1), PC (20:3 / 22:6), and PC (20:3 / 20:4), indicates that the intracranial artery plaque of the subject is prone to obvious enhancement and there is a risk of plaque instability.

[0016] A further improvement of the technical solution of the present invention lies in that a decrease in the combined serum biomarker concentration of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC (14:0 / 18:2) indicates that the intracranial artery plaque of the subject is prone to obvious enhancement and there is a risk of plaque instability.

[0017] A further improvement of the technical solution of the present invention lies in that a decrease in the combined serum biomarker concentration of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC (14:0 / 18:2) indicates that the intracranial artery plaque of the subject is prone to obvious enhancement. As a marker of plaque instability and the risk of stroke, plaque enhancement can help identify the recurrence of ischemic events.

[0018] A further improvement of the technical solution of the present invention lies in that the serum biomarker described in claim 1 or 2, after adjusting for confounding factors, obvious enhancement of the cerebral artery plaque is an independent risk factor for the recurrence of cerebral infarction and is used for the early diagnosis of predicting the recurrence of cerebral infarction in the subject.

[0019] A further improvement of the technical solution of the present invention lies in the application of the serum biomarker described in claim 1 or 2 for identifying and evaluating the effect of drug and / or surgical treatment on the treatment of the recurrence of cerebral infarction.

[0020] Due to the adoption of the above technical solution, the technical effects achieved by the present invention are as follows:

[0021] The present invention provides serum metabolic markers that can evaluate the stability of intracranial atherosclerotic plaques. By detecting the concentrations of one or several of phosphatidylcholine PC(15:0 / 15:1), phosphatidylcholine PC(14:0 / 18:2), phosphatidylethanol PE(18:2 / 18:3), phosphatidylethanol PE(18:1e / 18:1), 5α-dihydrotestosterone, D-fructose 6-phosphate, phosphatidylcholine PC(14:0 / 18:2), sphingomyelin SM(d28:0 / 14:1), phosphatidylcholine PC(20:3 / 22:6), and phosphatidylcholine PC(20:3 / 20:4) in the blood samples of subjects, using the technology of ultra-high performance liquid chromatography (UHPLC) combined with high-resolution mass spectrometry, and using the data processing mode of multivariate statistical analysis combined with machine learning algorithms, a simple, efficient and easily acceptable serological screening method is found to judge the stability of intracranial atherosclerotic plaques in subjects, early identify vulnerable plaques, provide a new direction for the treatment or intervention of atherosclerosis, provide clues for finding the recurrence of cerebral infarction, and provide a basis for formulating strategies for preventing the recurrence of cerebral infarction. Brief Description of the Drawings

[0022] Figure 1 is the flow chart for screening serum metabolic markers of the present invention;

[0023] Figure 2 is the experimental flow chart of the present invention;

[0024] Figure 3 is the clustering heat map of the changes of 9 differential metabolites of the present invention;

[0025] Figure 4 is the ROC curve of serum biomarkers significantly related to the enhancement of cerebral artery plaques;

[0026] Figure 5 is the best combination of markers for significant enhancement of intracranial atherosclerotic plaques. Detailed Description of the Invention

[0027] To elaborate on the technical content, structural features, achieved objectives and effects of the present invention, the following will be described in detail with reference to the accompanying drawings of the specification.

[0028] The present invention studies the serum samples of the population with significant enhancement and non-significant enhancement of intracranial atherosclerotic plaques, uses the technology of ultra-high performance liquid chromatography (UHPLC) combined with high-resolution mass spectrometry, and uses the data processing mode of multivariate statistical analysis combined with machine learning algorithms to find differentially expressed metabolites with statistical significance between the two groups, so as to explore the molecular mechanism of plaque enhancement. The discovery of biomarkers may help improve cerebrovascular risk prediction and early disease screening.

[0029] 1. Medical record selection and sample preparation

[0030] Inclusion criteria for the research subjects: ① Patients aged 35 - 80 years old, with the first acute cerebral infarction confirmed by Diffusion-weighted imaging (DWI) within 7 days of onset, and having ≥50% stenosis of the intracranial large artery; ② The general vital signs of the patients are stable.

[0031] Exclusion criteria for the research subjects: ① There is ≥50% stenosis of the cervical vessels on the same side of the lesion; ② There is atrial fibrillation, valvular disease or cardiogenic embolism caused by other reasons; ③ Patients with non-atherosclerotic cerebral infarction such as intracranial artery dissection, arteritis, etc.; ④ Patients with transient ischemic attack with negative DWI; ⑤ There are relative / absolute contraindications related to the safety of magnetic resonance imaging such as in vivo metal implants and claustrophobia.

[0032] The signal intensities of the plaque and the pituitary stalk after enhancement in HR-MRI were measured, and the ratio of the two indicators was calculated. The degree of plaque enhancement was divided into 3 grades: Grade 0 (no enhancement): The degree of enhancement is similar to or less than that of the normal intracranial arterial wall; Grade 1 (mild enhancement): The degree of enhancement is greater than Grade 0 but less than that of the pituitary infundibulum; Grade 2 (obvious enhancement): The degree of enhancement is similar to or higher than that of the pituitary infundibulum. In this analysis, Grade 2 enhancement was defined as the obvious enhancement group (HE group), and Grades 0 and 1 were defined as the non-obvious enhancement group (non-HE group). The correlations between parameters such as the degree of stenosis of the responsible vessel, plaque burden, and plaque enhancement grading and the recurrence events of cerebral infarction were analyzed, and a recurrence risk model of cerebral infarction was established.

[0033] Thirty cases were randomly selected from each of the HE group and the non-HE group for metabolomics analysis. Peripheral blood samples were collected with vacuum tubes containing EDTA, left to stand for 30 minutes and then centrifuged (4℃, 3000 rpm, 15 min). The separated serum was frozen at -80℃. The original files (.raw) obtained from mass spectrometry detection were imported into Compound Discoverer 3.1 (CD3.1) software. Simple screening of parameters such as retention time and mass-to-charge ratio was performed for each metabolite, and then peak alignment of different samples was performed by setting a retention time deviation of 0.2 min and a mass deviation of 5 ppm to make the identification more accurate. Subsequently, peak extraction was performed by setting information such as a mass deviation of 5 ppm, a signal intensity deviation of 30%, a signal-to-noise ratio of 3, the minimum signal intensity, and sum ions. At the same time, the peak area was quantified, the target ions were integrated, and then the molecular formula was predicted through the molecular ion peak and fragment ions and compared with

[0034] The data was compared with the mzCloud (https: / / www.mzcloud.org / ), mzVault, and Masslist databases. Background ions were removed using blank samples, and the original quantitative results were normalized. Finally, metabolite identification and relative quantification results were obtained. Three databases were used for database searching, such as the KEGG database (https: / / www.genome.jp / kegg / pathway.html), the LIPID Maps database (http: / / www.lipidmaps.org / ), and the HMDB database (https: / / hmdb.ca / metabolites) to annotate the identified metabolites and determine the types of differential metabolites. A total of 1404 different metabolites were detected in the HE and non-HE groups. The experimental flow chart is as shown in Figure 2 .

[0035] 2. Serum biomarker screening and prediction model establishment

[0036] Due to the high correlation among variables in metabolomic data, traditional univariate statistical methods cannot accurately mine data information. Multivariate statistical methods need to be used to perform dimensionality reduction and regression analysis on the collected multi-dimensional data to screen for differential metabolites. After converting the data using the metaX software, partial least squares discriminant analysis (PLS-DA) was performed. PLS-DA is a supervised discriminant analysis statistical method that uses partial least squares regression to establish a relationship model between metabolite expression levels and sample categories to achieve the prediction of sample categories. The variable importance in the projection (VIP) value of the first principal component of each metabolite was obtained through PLS-DA. VIP represents the influence intensity and explanatory ability of each metabolite on the classification discrimination between the two groups. VIP > 1 was used as the screening criterion for differential metabolites. In the univariate analysis part, the statistical significance (P value) of each metabolite between the two groups was calculated based on the t-test, and the fold change (FC) value of the metabolite between the two groups was calculated. A positive FC value indicates an increase in the relative content of the metabolite; a negative value indicates a decrease in the relative content of the metabolite. The screening criteria for differential metabolites were VIP > 1, P value < 0.05, and FC > 1.2 or < 0.833. A total of 169 differential endogenous metabolites in the ESI+ mode and 110 metabolites in the ESI- mode were identified, with a total of 279 variables, of which 125 variables were up-regulated and 154 variables were down-regulated.

[0037] Standard analysis obtained relatively more differential metabolites. Further, a data processing method using machine learning algorithms was adopted. Through supervised machine learning models, namely Random Forests (RF) and Least absolute shrinkage and selection operator (LASSO) regression analysis, and using the method of cross-validation, differential metabolites with statistical significance between the two groups were found, further reducing the number of metabolites. A binary logistic regression was used to construct a prediction model for the enhancement risk of cerebral artery plaques. Taking the subjects with non-obvious plaque enhancement as the reference objects, the predictive factors jointly screened by machine learning and non-machine learning were included, and potential biomarkers with high specificity and sensitivity for predicting obvious plaque enhancement were found. The ROC curve was plotted to evaluate the prediction effect of the model, and metabolite combinations with strong discriminant ability were found. The biological significance of the metabolites was explained through metabolic pathways to determine the biomarkers.

[0038] In this application, 16 differential metabolites were screened by the method of random forest and cross-validation, and 24 differential metabolites were selected by the LASSO regression model. The specific differential metabolites are shown in Table 1 and Table 2.

[0039] Table 1 16 differential metabolites after cross-validation by random forest

[0040]

[0041]

[0042] Table 2 24 differential metabolites selected by LASSO regression

[0043]

[0044]

[0045] Further, 9 differential metabolites with FC > 2 or < 0.5 were selected from them: 4 phosphatidylcholines, 1 sphingomyelin, 2 phosphatidylethanolamines, 1 androgen, and 1 carbohydrate. Compared with the non-HE group, the contents of 6 serum metabolites, namely phosphatidylcholine (PC) (15:0 / 15:1), PC (14:0 / 18:2), phosphatidylethanol (PE) (18:2 / 18:3), PE (18:1e / 18:1), 5α-Dihydrotestosterone, and D-Fructose 6-phosphate, in the serum samples of patients in the HE group decreased, while the contents of 3 serum metabolites, namely sphingomyelin (SM) (d28:0 / 14:1), PC (20:3 / 22:6), and PC (20:3 / 20:4), increased. See Table 3.

[0046] Table 3 Serum metabolic markers related to significant enhancement of intracranial atherosclerotic plaques between two groups

[0047]

[0048] Hierarchical clustering analysis was performed using the screened Biomarkers. The relative quantitative values of the Biomarkers were normalized and clustered to verify their discriminant ability. For the clustering heatmap of the expression differences of 9 differential metabolites, see Figure 3 . The potential biomarkers discovered in this application are involved in glycerophospholipid metabolism, androgen transport and metabolism, glycolysis, and sphingolipid metabolism, providing a new direction for a more in-depth understanding of the metabolic mechanism of intracranial atherosclerotic plaque enhancement.

[0049] After normalizing the relative quantitative values of the metabolites by Z-score, binary logistic regression analysis was performed. Logistic regression analysis can obtain the weights of independent variables, and these weights can be used to measure the ability of biomolecules to distinguish different groups. The effects of 2 clinical variables and 9 serum metabolite levels on the significant enhancement of plaques were corrected. Through forward stepwise variable selection analysis (Wald test), it was found that 5α-dihydrotestosterone, PC(14:0 / 18:2), and D-fructose 6-phosphate were significantly correlated with the significant enhancement of intracranial atherosclerotic plaques. With the increase in the concentrations of 5α-dihydrotestosterone (OR: 0.044; 95% CI: 0.004 - 0.542), D-fructose 6-phosphate (OR: 0.062; 95% CI: 0.005 - 0.789), and PC(14:0 / 18:2) (OR: 0.031; 95% CI: 0.001 - 0.894), it was less likely for the plaques to show significant enhancement. See Table 4 for details.

[0050] Table 4 Predictive factors for significant enhancement of plaques in patients with intracranial atherosclerotic plaques

[0051]

[0052] In summary, when comparing the HE group and the non-HE group in the population with cerebral infarction, a decrease of less than 0.5 times in the content of one of the metabolites phosphatidylcholine PC(15:0 / 15:1), PC(14:0 / 18:2), phosphatidylethanolamine PE(18:2 / 18:3), PE(18:1e / 18:1), 5α-Dihydrotestosterone, and D-Fructose 6-phosphate indicates that the intracranial artery plaque in the subject is prone to significant enhancement and there is a risk of plaque instability. An increase of more than 2 times in the content of the three serum metabolites sphingomyelin SM(d28:0 / 14:1), PC(20:3 / 22:6), and PC(20:3 / 20:4) indicates that the intracranial artery plaque in the subject is prone to significant enhancement and there is a risk of plaque instability. After systematic screening, the more obvious the decrease in the combined serum biomarkers of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC(14:0 / 18:2), the more likely the intracranial artery plaque in the subject is to show significant enhancement.

[0053] To evaluate the diagnostic efficacy of metabolites for the two groups, according to the machine learning model confirmed by Biomarker, the area under the curve (AUC) of the corresponding receiver operating characteristic curve was generated to analyze and evaluate the predictive ability of biomarkers. The criteria for using the AUC value to evaluate the classification model are as follows: for AUC = 1, the model can perfectly classify, but in most cases, perfect classification of the predicted data does not exist; for 0.5 < AUC < 1, the established model is better than random guessing, and this model can have predictive value by properly adjusting parameters; for AUC = 0.5, the model has no predictive value, and for AUC < 0.5, the established model is worse than random guessing. The abscissa represents the false positive rate (1 - specificity), and the ordinate represents the true positive rate (sensitivity). When the AUC value = 0.5, the biomarker has no effect on predicting the occurrence of the event and has no predictive value. When the AUC value > 0.5, the closer the AUC value is to 1, the higher the prediction accuracy. Generally speaking, when the AUC value is between 0.5 and 0.7, the prediction accuracy is relatively low; when the AUC value is between 0.7 and 0.9, there is a certain prediction accuracy; when the AUC value is above 0.9, there is a high prediction accuracy. The larger the AUC value, the higher the sensitivity of the metabolite in distinguishing between the HE group and the non-HE group. For the 9 markers, the AUC values are all greater than 0.75, indicating good predictive ability. Taking significant plaque enhancement as the state variable and 3 independent risk factors as the test variables, a combined ROC curve was obtained, with an AUC of 0.980. See Figure 4 。

[0054] Application of 3 Serum Metabolic Markers in Evaluating the Stability of Intracranial Atherosclerotic Plaques in Patients with Cerebral Infarction

[0055] Collect blood samples from the subjects for non-targeted metabolomics analysis. According to current research, in the acute cerebral infarction cohort, the HE group and the non-HE group were compared. The content of one of the metabolites phosphatidylcholine PC(15:0 / 15:1), PC(14:0 / 18:2), phosphatidylethanolamine PE(18:2 / 18:3), PE(18:1e / 18:1), 5α-Dihydrotestosterone, D-Fructose 6-phosphate decreased by less than 0.5 times, indicating that the intracranial artery plaque in the subject was prone to obvious enhancement and there was a risk of plaque instability. The content of 3 serum metabolites, sphingomyelin SM(d28:0 / 14:1), PC(20:3 / 22:6), and PC(20:3 / 20:4), increased by more than 2 times, indicating that the intracranial artery plaque in the subject was prone to obvious enhancement and there was a risk of plaque instability. After systematic screening, the more obvious the decrease in the combined serum biomarkers of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC(14:0 / 18:2), the more likely the intracranial artery plaque in the subject was to show obvious enhancement.

[0056] According to previous research evidence, the enhancement of the vascular wall may be the result of endothelial cell injury, inflammatory response, neovascularization within the plaque, or thrombosis caused by plaque rupture. Plaque enhancement is generally considered a characteristic of plaque instability. Our study found that the annual recurrence risk of cerebral infarction was significantly correlated with obvious plaque enhancement. Furthermore, our study found that the above-mentioned serum biomarkers could be used to predict the recurrence risk of cerebral infarction. In the secondary prevention of cerebral infarction, the images of HR-MRI have been used as a reliable basis for determining the etiology, pathogenesis, and selecting individual prevention and treatment measures for patients with cerebral infarction. However, there are still disadvantages such as high cost, small popularization range, and the need for intravenous injection of contrast agents, which affect its routine screening and implementation in patients with cerebral infarction. Our study found that serum biomarkers could also predict the stability of intracranial atherosclerotic plaques and the recurrence of cerebral infarction, and were simple, efficient, low-cost, and highly sensitive. Therefore, serum biomarkers could be used as markers for intracranial plaque instability and the risk of cerebral infarction.

[0057] The above-described embodiments are merely descriptions of the preferred embodiments of the present invention and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. Use of serum biomarkers for non-disease diagnosis, characterized in that: For evaluating the stability of intracranial artery plaques, the serum biomarker is a combination of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC(14:0 / 18:2); the serum biomarker is derived from the serum of a subject.

2. The application of the serum biomarker for non-disease diagnosis according to claim 1, wherein: A decrease in the combined serum biomarker content of 5α-dihydrotestosterone, D-fructose 6-phosphate, and phosphatidylcholine PC(14:0 / 18:2) indicates that the intracranial artery plaques of the subject are prone to significant enhancement and there is a risk of plaque instability.

Images