Nasal secretion biomarker composition for diagnosing type 2 chronic sinusitis with nasal polyp and application thereof

Abstract

The invention discloses a nasal secretion biomarker composition for diagnosing type 2 chronic sinusitis with nasal polyp and application of the nasal secretion biomarker composition, and relates to the technical field of biological medicine. The nasal secretion biomarker composition comprises a glial cell-derived neurotrophic factor, a cysteine protease inhibitor 5, a transforming growth factor beta 1 and a monocyte chemotactic protein 4. When the nasal secretion biomarker composition is used for diagnosing the intrinsic type of CRSwNP, the nasal secretion is used as a detection sample, polyp biopsy is not needed, the nasal secretion biomarker composition is free of invasiveness, simple, convenient and high in sensitivity and specificity, CRSwNP patients can be classified before an operation under the condition that mucous membranes are not damaged, and the clinical diagnosis efficiency is improved. The method can accurately predict the disease intrinsic type of the CRSwNP patient in advance, and is an ideal non-invasive method.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technology, and in particular to a nasal secretion biomarker composition for diagnosing type 2 chronic sinusitis with nasal polyps and an application thereof. Background Art

[0002] Chronic rhinosinusitis (CRS) is a common inflammatory disease of the upper respiratory tract. According to the clinical phenotype of chronic rhinosinusitis, it is mainly divided into polyp-associated type and non-polyp-associated type. Among them, chronic rhinosinusitis with nasal polyps (CRSwNP) can be further divided into type 2 CRSwNP (also known as Type2 CRSwNP or T2 CRSwNP) and non-type 2 CRSwNP (also known as non-Type2 CRSwNP or non-T2 CRSwNP) according to its intrinsic characteristics. Type 2 CRSwNP usually presents with moderate to severe type 2 helper T cell (Th2) inflammation, accompanied by tissue eosinophilia and elevated IgE (immunoglobulin E) concentrations. The clinical symptoms of this type of patients are more severe, and usually require longer, higher-dose drug treatment and / or repeated surgery. Therefore, predicting the intrinsic type of CRSwNP patients can provide accurate guidance for formulating clinical treatment plans.

[0003] At present, the common method for predicting the intrinsic type of CRSwNP is mainly based on inflammatory factors in nasal polyp tissue. In recent years, different classification indicators have been proposed at home and abroad, and models from simple to complex have been gradually developed to reveal the characteristics of the intrinsic type of CRS. For example, Tomassen et al. identified ten potential intrinsic characteristics of CRS through tissue inflammation biomarkers obtained through surgery (see PMID: 26949058 for details. PMID is the unique identification code of PubMed. It is the literature number in the fields of life sciences and medicine included in the PubMed search engine. The corresponding file can be found by directly entering the PMID number in the PubMed search engine, and each PMID number corresponds to a unique document); Wang et al. divided CRS patients into non-type 2, low type 2, medium type 2, and high type 2 by evaluating the degree of type 2 inflammation. These classifications showed significant differences in inflammatory molecular characteristics and tissue remodeling (see PMID: 36272582 for details). However, these intrinsic diagnostic methods mainly rely on the detection of tissue inflammatory factors and require invasive surgery to obtain nasal polyp tissue, which is subject to various limitations during the operation. At the same time, the concentration of inflammatory factors may be affected by changes in protein expression in specific parts of the sinus cavity, which makes this method difficult to promote and apply on a large scale in clinical practice.

[0004] Therefore, the existing technologies still need to be improved and developed. Summary of the Invention

[0005] Based on the deficiencies of the above-mentioned existing technologies, the objective of the present invention is to provide a nasal secretion biomarker composition for diagnosing type 2 CRSwNP and its application, aiming to solve the problem that the diagnosis of the endotype of CRSwNP mainly relies on inflammatory factors in nasal polyp tissues, and nasal polyp tissues need to be obtained through invasive surgery.

[0006] The technical solution of the present invention is as follows:

[0007] In the first aspect of the present invention, a nasal secretion biomarker composition is provided, wherein the nasal secretion biomarker composition includes glial cell line-derived neurotrophic factor, cystatin 5, transforming growth factor β1, and monocyte chemoattractant protein 4.

[0008] In the second aspect of the present invention, an application of the nasal secretion biomarker composition as described above in the preparation of a diagnostic product for the endotype of CRSwNP is provided.

[0009] Optionally, the diagnosis of the endotype of CRSwNP includes differentiating type 2 CRSwNP from non-type 2 CRSwNP.

[0010] Optionally, the product includes a reagent or a kit.

[0011] Optionally, when the product is used for diagnosing the endotype of CRSwNP, the sample used is nasal secretion.

[0012] In the third aspect of the present invention, an application of the nasal secretion biomarker composition as described above in constructing a diagnostic model for the endotype of CRSwNP is provided.

[0013] Optionally, a diagnostic model for the endotype of chronic rhinosinusitis with nasal polyps is constructed based on Lasso regression, and the diagnostic model is logit(π) = 0.031 + 1.04×log 2 A + 0.149×log 2 B - 0.791×log 2 C - 0.997×log 2 D + 0.514×E - 0.169×F;

[0014] Among them, logit(π) is the predicted value of the diagnostic model; A is the relative expression level of glial cell line-derived neurotrophic factor in the nasal secretions of CRSwNP subjects to be tested; B is the relative expression level of monocyte chemoattractant protein 4 in the nasal secretions of CRSwNP subjects to be tested; C is the relative expression level of transforming growth factor β1 in the nasal secretions of CRSwNP subjects to be tested; D is the relative expression level of cystatin 5 in the nasal secretions of CRSwNP subjects to be tested; E is the age of CRSwNP subjects to be tested, and F is the gender of CRSwNP subjects to be tested, where 1 is taken for males and 0 is taken for females;

[0015] When the value of logit(π) is greater than or equal to 0.53, the endotype of the CRSwNP subject to be tested is type 2 CRSwNP.

[0016] Beneficial effects: When the nasal secretion biomarker composition in the present invention is used for the diagnosis of CRSwNP endotype, with nasal secretions as the detection sample, no polyp biopsy is required, it is non-invasive, simple and convenient, and has high sensitivity and specificity. It can classify CRSwNP patients before surgery without damaging the mucosa, and predict the disease endotype of CRSwNP patients in advance and accurately. It is an ideal non-invasive method. Using the nasal secretion biomarker composition provided by the present invention to diagnose the CRSwNP endotype helps doctors formulate personalized treatment plans for patients, thereby avoiding unnecessary over-medication or side effects, and improving the curative effect at the same time; reducing the surgical risk. For patients who require surgical intervention, accurate pre-operative typing can help doctors better plan the surgical strategy and reduce the postoperative complication and recurrence rates; improving the efficiency of resource utilization. By identifying high-risk patients earlier, medical resources can be preferentially allocated to the patients who need them most, improving the efficiency of the overall medical service; improving the prognosis of patients. Through more accurate diagnosis and targeted treatment, while reducing the probability of disease progression and recurrence, it can also significantly improve the quality of life of patients. Description of the Drawings

[0017] Figure 1 It is a result graph of different evaluation indexes of the training set in Xgboost, Logistic regression, Lasso regression, random forest and decision tree.

[0018] Figure 2 It is a comparison graph of AUROC values of the test set in Xgboost, Logistic regression, Lasso regression, random forest and decision tree.

[0019] Figure 3 It is a result graph of the five-fold cross-validation AUROC values of Xgboost, Logistic regression, Lasso regression, random forest and decision tree.

[0020] Figure 4 ROC curve analysis diagrams for the training set, test set, and validation set.

[0021] Figure 5 Calibration curve diagram for the Lasso regression diagnostic model.

[0022] Figure 6 Relationship diagram between coefficients and Log(λ).

[0023] Figure 7 Relationship diagram between binomial deviance and Log(λ). Detailed implementation manners

[0024] The present invention provides a nasal secretion biomarker composition and its application. To make the objectives, technical solutions, and effects of the present invention clearer and more definite, the present invention is further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0025] Unless otherwise defined, all technical terms and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. The terms used in the description of the present invention are only for the purpose of describing specific implementation manners and are not intended to limit the present invention.

[0026] Currently, the common methods for predicting the endotype of CRSwNP mainly rely on inflammatory factors in nasal polyp tissues, but nasal polyp tissues need to be obtained through invasive surgery, which is restricted in the operation process. At the same time, the concentration of inflammatory factors may be affected by changes in protein expression in specific parts of the nasal sinus cavity, making it difficult to widely apply this method in clinical practice. If nasal secretions are used as the test samples, it is not only convenient and not limited by local mucosal inflammation, but also has the advantage of longitudinal dynamic quantitative evaluation. By detecting inflammatory factors in nasal secretions, it is expected to predict the classification of CRSwNP patients in advance, so as to provide more personalized and targeted treatment for CRSwNP patients. Based on this, the embodiments of the present invention provide a nasal secretion biomarker composition, wherein the nasal secretion biomarker composition includes glial cell line-derived neurotrophic factor (GDNF), cystatin 5 (CST5), transforming growth factor β1 (TGFB1), and monocyte chemoattractant protein 4 (MCP-4).

[0027] When the nasal secretion biomarker composition in the present invention is used for the diagnosis of the endotype of CRSwNP, nasal secretion is used as the detection sample. There is no need for polyp biopsy, it is non-invasive, simple and convenient, and has high sensitivity and specificity. It can classify CRSwNP patients before surgery without damaging the mucosa, and accurately predict the disease endotype of CRSwNP patients in advance. It is an ideal non-invasive method. Using the nasal secretion biomarker composition provided by the present invention to diagnose the endotype of CRSwNP helps doctors formulate personalized treatment plans for patients, thereby avoiding unnecessary over-medication or side effects, while improving the curative effect; reducing the surgical risk. For patients who need surgical intervention, accurate pre-operative typing can help doctors better plan the surgical strategy and reduce the postoperative complication and recurrence rates; improving the efficiency of resource utilization. By identifying high-risk patients early, medical resources can be preferentially allocated to the patients most in need, improving the efficiency of the overall medical service; improving the prognosis of patients. Through more accurate diagnosis and targeted treatment, while reducing the probability of disease progression and recurrence, it can also significantly improve the quality of life of patients.

[0028] The embodiment of the present invention also provides an application of the nasal secretion biomarker composition as described above in the preparation of a diagnostic product for the endotype of CRSwNP.

[0029] In some embodiments, the diagnosis of the endotype of CRSwNP includes differentiating type 2 CRSwNP and non-type 2 CRSwNP. The nasal secretion biomarker composition as described above in the present invention can effectively differentiate type 2 CRSwNP and non-type 2 CRSwNP.

[0030] In some embodiments, the product includes a reagent or a kit.

[0031] In some embodiments, when the product is used for diagnosing the endotype of CRSwNP, the sample used is nasal secretion.

[0032] The embodiment of the present invention also provides an application of the nasal secretion biomarker composition as described above in the construction of a diagnostic model for the endotype of CRSwNP.

[0033] In some embodiments, a diagnostic model for the endotype of chronic rhinosinusitis with nasal polyps is constructed based on Lasso regression, and the diagnostic model is logit(π) = 0.031 + 1.04×log 2 A +0.149×log 2 B -0.791×log 2 C -0.997×log 2 D+0.514×E - 0.169×F;

[0034] Among them, logit(π) is the predicted value of the diagnostic model; A is the relative expression level of GDNF in the nasal secretions of CRSwNP subjects to be tested; B is the relative expression level of MCP-4 in the nasal secretions of CRSwNP subjects to be tested; C is the relative expression level of TGFB1 in the nasal secretions of CRSwNP subjects to be tested; D is the relative expression level of CST5 in the nasal secretions of CRSwNP subjects to be tested; E is the age of CRSwNP subjects to be tested; F is the gender of CRSwNP subjects to be tested, where 1 is taken for males and 0 is taken for females; the relative expression levels A, B, C, and D of the above proteins are measured by the Olink Target 96 Inflammation Panel, and the Olink Target 96 Inflammation Panel (including reference proteins) is purchased from Olink Proteomics (Uppsala, Sweden), with the product number 95000B and the batch number B32027B;

[0035] When the value of logit(π) is greater than or equal to 0.53, the endotype of the CRSwNP subject to be tested is type 2 CRSwNP.

[0036] The diagnostic model in the present invention has high accuracy, can effectively distinguish type 2 CRSwNP from non-type 2 CRSwNP, and realizes the diagnosis of the endotype of CRSwNP subjects to be tested.

[0037] The present invention will be further described below through specific examples.

[0038] The research in the following examples was approved by the Scientific Research Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University (No. 2023-211). And all subjects signed a written consent form before the operation. The diagnosis of CRSwNP in the subjects was based on medical history, physical examination, nasal endoscopy, and sinus CT scan, following the 2020 European Position Paper on Rhinosinusitis and Nasal Polyps (EPOS2020) (see PMID: 32077450).

[0039] In the following examples, the meanings of English and English abbreviations are as follows:

[0040] TNF-α: Tumor necrosis factor-α;

[0041] IL-1β: Interleukin-1β;

[0042] IL-5: Interleukin-5;

[0043] IL-6: Interleukin-6;

[0044] IFN-γ: Interferon-γ;

[0045] IL-13: Interleukin-13;

[0046] IL-17A: Interleukin-17A;

[0047] ECP: human eosinophil cationic protein;

[0048] Periostin: periostin;

[0049] ALB: albumin;

[0050] IL-8: Interleukin-8;

[0051] MPO: myeloperoxidase.

[0052] Example 1

[0053] Exclusion criteria for CRSwNP patients included: (1) posterior choanal polyps of the maxillary sinus, cystic fibrosis, allergic fungal rhinosinusitis; (2) malignant tumors, severe systemic chronic diseases, infectious diseases; (3) autoimmune diseases and other systemic immune diseases; (4) long-term use of immunosuppressive drugs; (5) age less than 18 years old; (6) use of steroids, antibiotics or non-steroidal anti-inflammatory drugs within 4 weeks before surgery. That is, CRSwNP patients who met any of the above were excluded.

[0054] (1) Obtain baseline demographic and general clinical characteristic information of 82 CRSwNP patients

[0055] The baseline demographic and general clinical characteristic information of 82 CRSwNP patients used for model construction is shown in Table 1.

[0056] Table 1. Baseline demographic and general clinical characteristic information of CRSwNP patients

[0057]

[0058]

[0059] Note: Mann-Whitney U test was used for continuous variables, and chi-square test was used for categorical variables. HPF represents high-power field of view; LM score represents Lund-Mackay sinus CT score; MLK score represents modified Lund-Kennedy nasal endoscopy score; SNOT-22 represents nasal sinus outcome test 22; VAS represents visual analogue scale.

[0060] In addition, the data forms involved in Table 1 are: median [minimum value, maximum value] or frequency (percentage %). Take "blood neutrophil count, 10 9Taking " / L" as an example, 3.55[0.300, 5.80] therein indicates that 3.55 is the median (that is, among non-type 2 CRSwNP patients, the median blood neutrophil count is 3.55×10 9 / L), the minimum value is 0.300 (that is, among non-type 2 CRSwNP patients, the minimum blood neutrophil count is 0.300×10 9 / L), and the maximum value is 5.80 (that is, among non-type 2 CRSwNP patients, the maximum blood neutrophil count is 5.80×10 9 / L). Additionally, taking "asthma, yes (%)" as an example, 2(4.8) therein indicates that 2 is the frequency (that is, among 40 non-type 2 CRSwNP patients, 2 have asthma), and 4.8 is 4.8% (that is, 4.8% of non-type 2 CRSwNP patients have asthma). The meanings of other data are to be inferred by analogy.

[0061] Among them, the type 2 CRSwNP patients among the 82 CRSwNP patients in Table 1 were determined by the following method, which specifically includes the following steps:

[0062] a. Protein quantification was performed using enzyme immunoassay technology (i.e., enzyme-linked immunosorbent assay, Elisa) and multi-factor liquid chip technology (Luminex) to measure the protein concentrations of 12 cytokines in the nasal polyp tissues of 82 CRSwNP patients. Based on the protein concentration data of these 12 biomarkers, consensus clustering analysis was performed to identify and distinguish the immune endotypes of the patients, which specifically includes the following steps:

[0063] Collect nasal tissue biopsies from the nasal polyps of CRSwNP patients and store them frozen. Weigh the frozen tissue samples and homogenize them. Take 0.1 g of the homogenized tissue, dissolve it in 1 mL of 0.9% NaCl solution, and use it for detection after centrifugation at 1500 g for 10 minutes.

[0064] Referring to the biomarkers used in the existing methods (PMID: 26949058, PMID: 36272582), 12 biomarkers (Table 2) were selected in this example, and these biomarkers reflect the inflammatory patterns observed in CRSwNP. Specifically, an ELISA assay kit (purchased from Abbexa) was used to detect ECP (Cat# abx055137). An ELISA kit (purchased from R&D Systems) was used to detect MPO (Cat# DY3174), ALB (Cat# DY1455), and Periostin (Cat# DY3548B). All other biomarkers were evaluated using the Luminex 100 system (purchased from R&D Systems) to obtain the protein concentrations of these 12 biomarkers.

[0065] Table 2. Names and Significances of Biomarkers for Tissue Cluster Analysis

[0066]

[0067] b. Based on the protein concentration data of the above 12 biomarkers, perform consensus clustering analysis to identify and distinguish the immune subtypes of CRSwNP patients, specifically including the following steps:

[0068] Use ConsensusClusterPlus (R package) to construct a consensus matrix, and set the following parameters for clustering the samples:'maxK = 10, reps = 50, pItem = 0.8, oFeature = 1, clusterAlg = 'pam', seed = 123456, distance = 'Euclidean'. Based on similar features and variable order, perform inflammatory subtyping (here the subtyping is defined as the dependent variable) on CRSwNP patients based on tissue biomarker cluster analysis, and group and label each of the 82 CRSwNP patients. Among them, there are 42 non-type 2 CRSwNP patients and 40 type 2 CRSwNP patients.

[0069] (2) Aspirate the nasal secretions of 82 CRSwNP patients and perform Olink proteomics technology detection

[0070] As a clinical sample, the protein concentration in nasal secretions may be low and is affected by various factors (such as collection time, collection method, etc.). The present invention combines the high sensitivity and precision of high-throughput Olink technology, making it very suitable for the analysis of such complex samples, capable of effectively detecting low-abundance or difficult-to-detect proteins related to type 2 CRSwNP, and can achieve accurate prediction with only 1 μL of sample volume.

[0071] Collect nasal secretions using the nasal secretion collection device (also known as Nasalfluids absorption device, denoted as NFAD) and sampling method disclosed in the inventor's previous patent (CN116019495A). Then, use the Olink immunoassay method, that is, use the Olink Target 96 inflammation panel (also known as Olink Target 96 inflammation panel, purchased from Olink Proteomics, Uppsala, Sweden, catalog number 95000B, batch number B32027B, containing reference proteins) to quantitatively detect 96 proteins in nasal secretions, and obtain the relative expression levels of each protein. After the relative expression levels of each protein are log2-transformed, the arbitrary unit normalized protein expression (NPX) of each protein is obtained, that is, NPX is the logarithm of the protein relative expression level with base 2. The higher the NPX, the higher the protein concentration. The difference in the expression protein levels is analyzed using the unpaired t-test (Mann-Whitney U test), and p < 0.1 is considered statistically significant.

[0072] Then, using the endotype of CRSwNP as the dependent variable and the nasal secretion biomarkers detected by Olink technology as the independent variables, a prediction model is constructed. Five different machine learning regression algorithms are used, combined with the five-fold cross-validation method, for model training, optimization, and testing. The performance evaluation of the model includes indicators such as the area under the receiver operating characteristic curve (also known as the area under the ROC curve, denoted as AUROC), precision, balanced precision, F-means (F1 score, representing the harmonic mean of precision and recall), NPV (negative predictive value), PPV (positive predictive value), PR AUC (area under the precision-recall curve), precision, recall, sensitivity, and specificity. The DeLong's test is used to evaluate the difference in AUROC values between models, and the model with the highest average indicator is considered the best. To determine the best classifier performance, the Youden's Index is used to determine the optimal cut-off point for positive class prediction in the training dataset. Based on the variable combination selected by the optimal model (for Lasso regression, a diagnostic model is constructed with GDNF, CST5, TGFB1, and MCP-4 and has the best diagnostic performance, see the following results for details), a diagnostic model for type 2 CRSwNP is constructed, specifically:

[0073] Using the stratified sampling method, an 82-CRSwNP patient dataset was randomly divided into a training set (for model training) and a test set (for model testing) in a 7:3 ratio. Five different machine learning methods, namely Logistic regression, Lasso regression, decision tree, random forest (RF), and extreme gradient boosting (Xgboost), were used. The endotype of CRSwNP was used as the dependent variable, and nasal secretion biomarkers detected by Olink technology were used as independent variables to construct a diagnostic model. Each machine learning method utilized its specific feature selection strategy to identify the best predictors. The five-fold cross-validation method and Bayesian optimization for hyperparameter tuning were adopted to reduce overfitting and improve model performance. After the classifier training was completed, the average of the five-fold cross-validation AUROC and the AUROC values of the training set and test set were used for comprehensive performance evaluation. DeLong's test was used to evaluate the statistical differences in AUROC values between models, and the model with the highest average metric was considered the best. The Youden index was used to determine the optimal cut-off point for positive class prediction in the training dataset.

[0074] For information on the relevant parameters of the machine learning algorithms, please refer to Table 3.

[0075] Table 3. Relevant parameter information of machine learning algorithms

[0076]

[0077]

[0078] Among them, the results of different evaluation metrics of the training set in Logistic regression, Lasso regression, decision tree, random forest, and Xgboost are as Figure 1 shown, and the AUROC value results of the test set in Logistic regression, Lasso regression, decision tree, random forest, and Xgboost are as Figure 2 shown; using the training set, the five-fold cross-validation AUROC value results of the diagnostic model in Logistic regression, Lasso regression, decision tree, random forest, and Xgboost are as Figure 3 shown.

[0079] When the above five machine learning algorithms were used to construct the diagnostic model, different biomarkers were adopted. However, from the above results, it can be seen that the Lasso regression diagnostic model constructed with GDNF, CST5, TGFB1, and MCP-4 had the best performance in predicting type 2 CRSwNP. As Figure 1As shown, the performance evaluations of the Lasso regression diagnostic model on the training set are as follows: AUROC = 0.91, precision = 0.86, balanced precision = 0.86, F-means = 0.86, NPV = 0.86, PPV = 0.86, PR AUC = 0.9, accuracy = 0.86, recall = 0.86, sensitivity = 0.86, and specificity = 0.86.

[0080] As Figure 2 shown, the AUROC value of the Lasso regression diagnostic model on the test set is 0.91. As Figure 3 shown, the five-fold cross-validation of the Lasso regression diagnostic model shows that its average AUROC reaches the highest value of 0.85. The above results indicate that the Lasso regression diagnostic model has strong predictive ability and stability. Additionally, the 95% confidence intervals (the AUROC value of the 95% confidence interval for the training set is 0.83 - 0.98; the AUROC value of the 95% confidence interval for the test set is 0.79 - 1, see Figure 4 ) further support the reliability and statistical significance of the Lasso regression diagnostic model under different data partitions.

[0081] In addition, 30 additional CRSwNP patients (14 of whom are type 2 CRSwNP and 16 are non-type 2 CRSwNP) were included as the validation set. These 30 CRSwNP patients are different patient samples from the above 82 CRSwNP patients.

[0082] Using the 30 CRSwNP patient samples as the validation set to further evaluate the accuracy of the Lasso regression diagnostic model, its AUROC value is 0.92 (as Figure 4 shown, the AUROC value of the 95% confidence interval for the validation set is 0.82 - 1). The above-constructed Lasso regression diagnostic model also has good diagnostic effects in the validation set, which further shows that the above-constructed Lasso regression diagnostic model has high diagnostic efficiency and clinical diagnostic significance.

[0083] Furthermore, the calibration plot of the Lasso regression diagnostic model shows that the diagonal between the predicted probability and the actual occurrence frequency is close to ( Figure 5 ), indicating that the prediction results are highly consistent with the actual observed results, and the model has good diagnostic ability.

[0084] Among them, the specific information of the above-constructed Lasso regression diagnostic model is as follows:

[0085] Set α (i.e., alpha) = 1, and optimize the regularization parameter (λ) (as Figure 6 and Figure 7As shown in [figure], four key predictive variables (i.e., biomarkers), namely GDNF, MCP-4, TGFB1, and CST5, were successfully screened out, and these variables are closely related to the prediction of the above Lasso regression diagnostic model. This process first optimized the regularization parameter through cross-validation to ensure that the model can maintain high prediction performance while preventing overfitting. The optimization of the regularization parameter helps to balance the complexity and fitting degree of the model, especially by constraining the model through L1 regularization (Lasso) and L2 regularization (Ridge), thereby reducing the introduction of unnecessary variables and enhancing the generalization ability of the model. These variables showed significant predictive ability in the model, indicating their close biological relevance or statistical association with type 2 CRSwNP. In the Lasso regression diagnostic model, after adjusting for gender and age, the final formula of the constructed Lasso regression diagnostic model is: logit(π) = 0.031 + 1.04×log 2 A + 0.149×log 2 B - 0.791×log 2 C - 0.997×log 2 D + 0.514×E - 0.169×F;

[0086] where logit(π) is the predicted value of the diagnostic model; A is the relative expression level of GDNF in the nasal secretions of CRSwNP patients to be tested; B is the relative expression level of MCP-4 in the nasal secretions of CRSwNP patients to be tested; C is the relative expression level of TGFB1 in the nasal secretions of CRSwNP patients to be tested; D is the relative expression level of CST5 in the nasal secretions of CRSwNP patients to be tested; E is the age of CRSwNP patients to be tested; F is the gender of CRSwNP patients to be tested, where male is taken as 1 and female is taken as 0; the optimal cut-off value based on the Youden index is 0.53.

[0087] In summary, the present invention provides a nasal secretion biomarker composition for diagnosing type 2 CRSwNP and its application. When the nasal secretion biomarker composition of the present invention is used for diagnosing the endotype of CRSwNP, taking nasal secretions as the detection sample, there is no need for polyp biopsy, it is non-invasive, simple and convenient, and has high sensitivity and specificity. It can classify CRSwNP patients before surgery without damaging the mucosa, and can predict the disease endotype of CRSwNP patients in advance and accurately. It is an ideal non-invasive method.

[0088] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or changes can be made according to the above description, and all such improvements and changes should fall within the protection scope of the appended claims of the present invention.

Claims

1. A nasal secretion biomarker composition, characterized in that: The nasal secretion biomarker composition includes glial cell line-derived neurotrophic factor, cystatin 5, transforming growth factor β1 and monocyte chemoattractant protein 4.

2. Use of the nasal secretion biomarker composition according to claim 1 in the preparation of an intrinsic diagnostic product for chronic sinusitis with nasal polyps.

3. The use according to claim 2, characterized in that: The diagnosis of CRSwNP intrinsic type includes differentiating between type 2 CRSwNP and non-type 2 CRSwNP.

4. The use according to claim 2, characterized in that: The product comprises a reagent or a kit.

5. The use according to claim 2, characterized in that: When the product is used to diagnose chronic sinusitis with intrinsic nasal polyps, the sample used is nasal secretions.

6. Use of the nasal secretion biomarker composition according to claim 1 in constructing an intrinsic diagnostic model for chronic sinusitis with nasal polyps.

7. The use according to claim 6, characterized in that: Based on Lasso regression, an intrinsic diagnostic model for chronic sinusitis with nasal polyps was constructed, and the diagnostic model was logit(π)=0.031+1.04×log2 A +0.149×log2 B -0.791×log2 C -0.997×log2 D +0.514×E-0.169×F; Among them, logit(π) is the predicted value of the diagnostic model; A is the relative expression of glial cell line-derived neurotrophic factor in the nasal secretions of subjects with chronic sinusitis and nasal polyps; B is the relative expression of monocyte chemoattractant protein 4 in the nasal secretions of subjects with chronic sinusitis and nasal polyps; C is the relative expression of transforming growth factor β1 in the nasal secretions of subjects with chronic sinusitis and nasal polyps; D is the relative expression of cysteine ​​proteinase inhibitor 5 in the nasal secretions of subjects with chronic sinusitis and nasal polyps; E is the age of subjects with chronic sinusitis and nasal polyps; F is the gender of subjects with chronic sinusitis and nasal polyps, where males are 1 and females are 0; When the logit(π) value is greater than or equal to 0.53, the intrinsic type of the subject with chronic sinusitis with nasal polyps is type 2 chronic sinusitis with nasal polyps.

Images