Construction method of malignant pleural effusion prediction model, marker combination and application, equipment and medium

Abstract

The invention discloses a construction method of a malignant pleural effusion prediction model, a marker combination and application, equipment and a medium. The construction method comprises the following steps: S1, obtaining characteristics of a biomarker combination; the biomarker combination is composed of three biomarkers of NGAL, CEA and CA50; s2, based on the characteristics of the biomarker combination, adopting Logistic regression analysis to respectively construct regression models corresponding to the three biomarkers, and respectively outputting parameters alpha NGAL, beta NGAL, alpha CEA, beta CEA, alpha CA50 and beta CA50; and S3, based on the output parameters in the step S2, adopting Bayesian analysis to construct a malignant pleural effusion prediction model. The method can be used for predicting the MPE probability in the pleural effusion patient, the identification and prediction accuracy is high, and compared with pleural biopsy and thoracoscopic sampling biopsy, the method has the advantages of being rapid, minimally invasive and the like.

Description

Technical Field

[0001] This invention belongs to the field of disease prediction model technology, specifically relating to a method for constructing a malignant pleural effusion prediction model, a combination of biomarkers and their application, equipment and media. Background Technology

[0002] Pleural effusion can be caused by a variety of diseases, including pneumonia, malignant tumors, tuberculous pleurisy, and heart failure. Pleural effusion caused by malignant tumors is called MPE. Multicenter epidemiological surveys show that MPE accounts for 23.3% of patients with pleural effusion. Common clinical tumor types that cause MPE include lung cancer, breast cancer, ovarian cancer, mesothelioma, and lymphoma. MPE has a poor prognosis, with a median survival of 6–12 months. Therefore, timely and accurate differentiation between MPE and benign pleural effusion (BPE) is crucial for the treatment of MPE patients and is key to improving their prognosis.

[0003] Currently, the gold standard for diagnosing MPE is cytological examination and pleural biopsy. Although cytological examination has advantages such as low cost, speed, and high specificity, its diagnostic sensitivity is only around 0.6, and its accuracy is related to the pathologist's experience. Pleural biopsy can be divided into image-guided pleural biopsy and thoracoscopic biopsy, with thoracoscopic biopsy being the most accurate method for diagnosing MPE. However, both pleural biopsy and thoracoscopic biopsy are invasive and harmful to the patient, and may subsequently lead to complications such as bleeding and infection.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a method for constructing a predictive model for malignant pleural effusion, a combination and application of biomarkers, as well as related equipment and media. The technical problem to be solved by this invention is achieved through the following technical solution: In a first aspect, the present invention provides a combination of biomarkers for the in vitro differential diagnosis of malignant pleural effusion, the combination of biomarkers consisting of NGAL, CEA and CA50, wherein the sample for in vitro identification is pleural effusion.

[0006] Secondly, the present invention provides the application of a reagent for detecting the above-mentioned combination of biomarkers in the preparation of a kit for diagnosing malignant pleural effusion.

[0007] Thirdly, the present invention provides a method for constructing a predictive model for malignant pleural effusion, comprising the following steps: S1. Obtain the biomarker combination of the training set samples, and perform feature extraction on the biomarker combination of the training set samples to obtain the features of the biomarker combination; the biomarker combination consists of three biomarkers: NGAL, CEA and CA50. S2. Based on the characteristics of the biomarker combination, logistic regression analysis was used to construct regression models for the three biomarkers respectively, and the parameters α were output respectively. NGAL β NGAL α CEA β CEA α CA50 and β CA50 ; Where, α NGAL α CEA α CA50 These are the intercept terms of the regression models for the three biomarkers NGAL, CEA, and CA50, respectively, and β is the value of β. NGAL β CEA β CA50 These are the regression coefficients of the regression models corresponding to the three biomarkers NGAL, CEA, and CA50, respectively. S3. Based on the output parameters in step S2, a prediction model for malignant pleural effusion is constructed using Bayesian analysis.

[0008] In one embodiment of the present invention, in step S1, the training set samples include malignant pleural effusion samples and benign pleural effusion samples; The malignant pleural effusion samples include lung cancer pleural effusion samples, malignant pleural mesothelioma pleural effusion samples, breast cancer pleural effusion samples, gastric cancer pleural effusion samples, and lymphoma pleural effusion samples; The benign pleural effusion samples include those from parapneumonia, tuberculous pleural effusion, heart failure, pulmonary embolism, mixed connective tissue disease, cirrhosis, pneumothorax, idiopathic pleural fibroelastosis, interstitial lung disease, pulmonary thromboembolism, and hypoalbuminemia.

[0009] In one embodiment of the present invention, in step S1, the biomarker combination is characterized by the concentrations of NGAL, CEA and CA50 in the training set samples, respectively. The concentration of NGAL was obtained using Luminex liquid phase chip technology. The concentrations of CEA and CA50 were obtained by chemiluminescent immunoassay.

[0010] In one embodiment of the present invention, in step S3, the prediction probability expression of the Bayesian model is: P = 0.31 × LR CEA ×LR NGAL×LR CA50 / (1+0.31×LR CEA ×LR NGAL ×LR CA50 ); where LR CEA LR NGAL LR CA50 These are the likelihood ratios for CEA, NGAL, and CA50, respectively.

[0011] Fourthly, the present invention provides a method for processing isolated pleural effusion data, comprising: The characteristics of a combination of biomarkers in a sample to be tested are detected, the combination of biomarkers consisting of NGAL, CEA and CA50; By inputting the features of the biomarker combination into the malignant pleural effusion prediction model constructed above, the risk value of malignant pleural effusion can be obtained.

[0012] Fifthly, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-described method for constructing a malignant pleural effusion prediction model when it calls the computer program in the memory.

[0013] In a sixth aspect, the present invention provides a storage medium storing a computer program for executing the steps of the above-described method for constructing a malignant pleural effusion prediction model.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a biomarker combination for pleural effusion, consisting of neutrophil gelatinase-associated lipocalin (NGAL), carcinoembryonic antigen (CEA), and carbohydrate antigen 50 (CA50), for the identification of malignant pleural effusion (MPE) and the prediction of the probability of MPE in patients with pleural effusion. It has high accuracy in identification and prediction and has advantages such as speed and minimal invasiveness compared with pleural biopsy and thoracoscopic biopsy.

[0015] This invention provides a method for constructing a predictive model using a combination of biomarkers (composed of NGAL, CEA and CA50). The method employs a Bayesian algorithm to calculate the posterior probability based on the likelihood ratio of the biomarkers, which helps in risk prediction and thus assists in making more accurate clinical decisions.

[0016] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a schematic diagram comparing the NGAL concentrations in the MPE and BPE groups provided in this embodiment of the invention. Figure 2 This is a schematic diagram comparing the CEA concentrations in the MPE group and the BPE group provided in this embodiment of the invention; Figure 3 This is a schematic diagram comparing the CA50 concentrations in the MPE and BPE groups provided in this embodiment of the invention; Figure 4 This is the ROC curve of NGAL, CEA, and CA50 jointly predicting the MPE probability provided in the embodiments of the present invention; Figure 5 This is the ROC curve of NGAL and CA50 jointly predicting the MPE probability provided in the embodiments of the present invention; Figure 6 This is a statistical graph showing the likelihood ratios of the biomarkers NGAL, CEA, and CA50 at different concentrations, provided in this embodiment of the invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following, in conjunction with the accompanying drawings and specific embodiments, provides a detailed description of the construction method, biomarker combination and application, equipment and media of a malignant pleural effusion prediction model proposed according to the present invention.

[0019] The foregoing and other technical contents, features, and effects of the present invention will be clearly presented in the following detailed description of specific embodiments in conjunction with the accompanying drawings. Through the description of the specific embodiments, a more in-depth and concrete understanding can be gained of the technical means and effects adopted by the present invention to achieve its intended purpose. However, the accompanying drawings are for reference and illustration only and are not intended to limit the technical solutions of the present invention.

[0020] The terms “comprising,” “including,” or any other variation thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed.

[0021] The term "pleural effusion" refers to an ex vivo specimen or sample. In the description of this invention, it can be understood that the pleural effusion collected from a patient with pleural effusion is a pleural effusion sample.

[0022] Compared with cytological examination and pleural biopsy, the detection of soluble markers in pleural effusion has advantages such as low cost, speed, ease of implementation, objectivity and minimal invasiveness, and is therefore a promising diagnostic method for MPE.

[0023] Neutrophil gelatinase-associated lipocarboxin (NGAL), also known as Lipocalin-2 (LCN2), is a small protein secreted by inflammatory-activated neutrophils or macrophages. Previous studies have suggested that blood NGAL is a biomarker of acute kidney injury.

[0024] Bayesian analysis is widely used in diagnostic research because it combines prior probabilities with likelihood ratios to dynamically update disease probabilities. Compared to traditional threshold classification methods, Bayesian analysis can fully utilize information from continuous variables. Therefore, this study uses a Bayesian theoretical framework, combined with continuous likelihood ratios, to systematically evaluate the combined diagnostic efficacy of biomarkers NGAL, CEA, and CA50 in pleural effusion. This study follows the Standard for Reporting Diagnostic Accuracy (STARD).

[0025] The solution provided by the present invention will be further described below with reference to specific embodiments.

[0026] Example 1: Sample Acquisition and Analysis The samples used in this study were patients treated at the Affiliated Hospital of Inner Mongolia Medical University between September 2018 and August 2021. Exclusion criteria were as follows: (i) patients with an undetermined diagnosis at discharge; (ii) pregnant women; (iii) patients under 18 years of age; (iv) patients who developed pleural effusion during hospitalization; and (v) pleural effusion caused by trauma. This study was approved by the Ethics Committee of the Affiliated Hospital of Inner Mongolia Medical University (No. 2018011). All participants signed informed consent forms. The Declaration of Helsinki was strictly followed in conducting this study.

[0027] The diagnostic criteria are as follows: Malignant pleural effusion (MPE) is diagnosed based on a positive pleural biopsy or positive pleural fluid cytology. In some patients with a high probability of MPE but negative cytology, and who are unwilling or unable to undergo pleural biopsy, the diagnosis of MPE is based on the primary tumor, after ruling out other causes of pleural effusion. Tuberculous pleural effusion is diagnosed based on a positive pleural biopsy, positive Mycobacterium tuberculosis culture, or positive acid-fast staining. In some patients with a high probability of tuberculous pleural effusion but negative etiological evidence, and who are unwilling to undergo pleural biopsy, the diagnosis of tuberculous pleural effusion is based on a pleural fluid adenosine deaminase (ADA) activity greater than 35 U / L, good responsiveness to anti-tuberculosis treatment, and after ruling out other causes of pleural effusion. Peripapillary pleural effusion is diagnosed using a mixed clinical criteria, including the patient's imaging characteristics, microbial culture, physical examination, laboratory tests, medical history, and responsiveness to antibiotic treatment. The diagnostic criteria for pleural effusion in heart failure include clinical manifestations, imaging features, serum NT-proBNP, and response to anti-heart failure treatment.

[0028] Based on the above exclusion and diagnostic criteria, pleural effusion samples from 153 patients were ultimately included in the analysis. This included 66 cases of malignant pleural effusion and 87 cases of benign pleural effusion (including 23 cases of heart failure pleural effusion, 32 cases of parapneumonia pleural effusion, 20 cases of tuberculous pleural effusion, and 12 cases of pleural effusion caused by other etiologies: 4 cases of pulmonary embolism pleural effusion, 2 cases of mixed connective tissue disease pleural effusion, 1 case of cirrhosis pleural effusion, 1 case of pneumothorax pleural effusion, 1 case of idiopathic pleural fibroelastosis pleural effusion, 1 case of interstitial lung disease pleural effusion, 1 case of pulmonary thromboembolism pleural effusion, and 1 case of hypoalbuminemia pleural effusion). The patients included in the analysis were divided into a malignant pleural effusion group (MPE group) and a benign pleural effusion group (BPE group).

[0029] Pleural fluid samples were collected from 153 patients with pleural effusion and placed in test tubes without any anticoagulants. After being left at room temperature for 4 hours, the samples were centrifuged. The centrifuged samples were aliquoted and stored in a refrigerator at -80°C to -70°C for later use.

[0030] Example 2: Detection of biomarkers in samples The concentrations of carcinoembryonic antigen (CEA), carbohydrate antigen 50 (CA50), and neutrophil gelatinase-associated lipocalin (NGAL) in pleural effusion samples were measured.

[0031] (1) Detection of carcinoembryonic antigen (CEA) and carbohydrate antigen 50 (CA50) The concentration of CEA in pleural fluid was detected using an Abbott I2000SR immunoassay analyzer via chemiluminescent immunoassay (CLIA).

[0032] The concentration of CA50 in pleural fluid was measured using a New Industries MAGLUMI 4000 chemiluminescence immunoassay analyzer.

[0033] (2) Detection of neutrophil gelatinase-associated lipocalin (NGAL) The concentration of NGAL in various pleural effusion samples was detected using Luminex liquid chromatography-array technology. The instrument used was a Luminex® 100 / 200™, and the reagents were two human premixed multianalyte kits from a custom R&D system. The kits included premixed magnetically encoded microspheres (with NGAL-specific antibodies pre-coated on them), biotin-labeled detection antibodies, streptavidin-phycoerythrin (SA-PE) conjugates, calibrators, diluents, and wash buffers.

[0034] The specific testing includes the following steps: Step 1: Take the frozen sample to be tested, place it on ice to thaw slowly, centrifuge and collect the supernatant.

[0035] Step 2: Vortex the premixed magnetically encoded microspheres for 30 seconds, and add 50 μL of the premixed microsphere suspension to each well of the 96-well microplate. The 96-well microplate is equipped with detection wells, standard curve determination wells, blank wells (with diluent added only) and quality control wells.

[0036] Step 3: According to the kit instructions, add a series of calibrators to the standard curve determination wells; add the supernatant of the pleural effusion sample to be tested to the detection wells; and add the set concentration of calibrator to the quality control wells as a positive control.

[0037] After the addition is complete, cover the 96-well microplate with a sealing film, place it on a flatbed shaker, and incubate at room temperature in the dark.

[0038] Step 4, Washing: After incubation, remove the sealing membrane; place the 96-well microplate on a magnetic separator and let it stand for 1 minute to allow the magnetic microspheres to adhere to the bottom of the wells. Carefully aspirate the liquid from the wells, being careful not to touch the magnetic beads. Add 100-200 μL of washing buffer to each well, remove the 96-well microplate from the magnetic separator, vortex to resuspend the microspheres, and then place the 96-well microplate on the magnetic separator again and let it stand for 1 minute to allow the magnetic microspheres to adhere to the bottom of the wells. Discard the supernatant washing buffer and perform a second wash.

[0039] Repeat the washing steps above twice more.

[0040] Step 5: After washing, add 50 μL of biotin-labeled detection antibody working solution to each well of the 96-well microplate, cover with a new sealing film, and incubate on a plate shaker at room temperature in the dark. After incubation, wash twice according to the washing procedure in Step 4.

[0041] Step 6: After washing, add 50 μL of streptavidin-phycoerythrin (SA-PE) working solution to each well; cover with a sealing membrane and incubate on a plate shaker at room temperature in the dark. Repeat the washing procedure in step 4 twice. After washing, add 100-150 μL of detection buffer to each well to resuspend the microspheres.

[0042] Step 7: Place the 96-well microplate into the sample loader of the Luminex® 100 / 200™ analyzer for testing.

[0043] Step 8: Detection process and data statistics The detection process mainly includes microsphere identification and signal reporting. Microsphere identification: The first light-emitting diode (LED) excites the dye inside the microsphere, and the specific analyte NGAL represented by each microsphere is identified according to the fluorescence code. Signal reporting: The second light-emitting diode (LED) excites the phycoerythrin on SA-PE, and the fluorescence intensity (MFI) emitted is measured. This intensity is proportional to the amount of NGAL analyte bound to the microsphere.

[0044] In instrumental analysis, at least 50 effective microspheres of the analyte NGAL are counted in each well, and the median fluorescence intensity (MFI) is recorded. A standard curve is generated using analytical software such as xPONENT® or Belysa® based on the concentration of the standard and its corresponding MFI value. The MFI value of each sample is then substituted into the standard curve to calculate the concentration of NGAL in that sample.

[0045] Example 3: Construction and Analysis of Predictive Model Continuous variables are represented using the median and interquartile range (IQR), such as Figures 1-3 This study compared the levels of three biomarkers in the MPE and BPE groups. The results showed that the levels of the three biomarkers were significantly higher in the MPE group than in the BPE group.

[0046] The Kolmogorov-Smirnov test was used to assess the normality of biomarker concentrations, with a significance level set at 0.10. Depending on the data distribution, ANOVA (for normally distributed data) or the Kruskal-Wallis H test (for skewed data) was used to compare differences in pleural effusion biomarker concentrations among patients with different types of pleural effusion. Independent samples t-tests (for normally distributed data) or Mann-Whitney U tests (for skewed data) were used for comparisons between two groups. Chi-square tests or Fisher's exact test were used to compare categorical data. All statistical analyses and graphs were performed using R software. A p-value < 0.05 was considered statistically significant.

[0047] (1) Constructing and analyzing the Logistic regression model Step 1: Construct a logistic regression model using three biomarkers: NGAL, CEA, and CA50. Logit(P=MPE) = β0 + (β1 × [NGAL]) + (β2 × [CEA]) + (β3 × [CA50]), where β0 = -2.48, β1 = 0.029, β2 = 0.043, β3 = 0.108, and [NGAL], [CEA], and [CA50] represent the concentrations of neutrophil gelatinase-associated lipotransferase, carcinoembryonic antigen, and carbohydrate antigen 50, respectively. Step 2: Plot the ROC curve of the combined prediction of MPE probability using the logistic regression model with NGAL, CEA, and CA50. See the results below. Figure 4 See Table 1.

[0048] Table 1. ROC curve parameters of pleural effusion biomarkers for identifying MPE

[0049] According to Table 1 and Figure 4The AUC of the ROC curve for the model constructed by combining three biomarkers was 0.88, with a sensitivity of 0.83 and a specificity of 0.85. Compared with the AUC of CEA and CA50 biomarkers alone, the AUC of the biomarker combination (composed of NGAL, CEA, and CA50) was significantly improved, indicating that the biomarker combination has higher diagnostic performance for MPE and BPE. Compared with the ROC curve using NGAL alone, the AUC improved from 0.85 to 0.88, which is a clinically significant improvement, indicating that the combined biomarker combination (composed of NGAL, CEA, and CA50) has a stronger ability to distinguish between diseased and healthy individuals. In addition, the lower limit of the AUC confidence interval for the biomarker combination (0.82) is higher than the upper limit of the AUC for NGAL alone (0.80), indicating that the biomarker combination has better robustness and reliability in performance evaluation. Compared to using NGAL alone for ROC curves, while achieving a significant increase in sensitivity, the specificity was adjusted from 0.97 (0.93-1.00) to 0.85 (0.77-0.92). However, a specificity of 0.85 remains very high, far exceeding the commonly used threshold for general diagnostic tests (e.g., 0.75-0.80), which is sufficient to ensure a low false positive rate. In other words, a higher false positive rate can be verified by subsequent, more precise (but potentially more expensive or invasive) gold standard methods. The consequences of missed diagnoses (false negatives), however, are usually irreversible. This invention successfully optimizes the diagnostic model from an extreme of 'high specificity, low sensitivity' to a more balanced and practical system of 'high sensitivity, high specificity'.

[0050] At the same time, such as Figure 5 As shown, the ROC curve of CA50 and NGAL combined was plotted using the same logistic regression analysis method. The AUC of CA50 and NGAL combined was 0.795.

[0051] (2) Bayesian analysis The triple biomarker combination of NGAL, CEA, and CA50 described in this invention demonstrated superior overall diagnostic efficacy compared to single biomarkers in ROC analysis (AUC: 0.88 vs. 0.85), particularly with a significantly improved sensitivity (0.83 vs. 0.71). However, the ROC curve provides an overall performance evaluation and cannot directly answer the question of the MPE prediction probability when facing the test results of a specific sample in clinical practice.

[0052] To address the aforementioned issues and more closely integrate diagnostic performance with clinical decision-making, this invention employs Likelihood Ratio (LR) and Bayesian Analysis. Specifically, it includes: Step 10: Fit a Logistic regression model using data from 153 samples (NGAL concentration as the independent variable and disease state "malignant / benign" as the dependent variable).

[0053] Logistic (P) NGAL ) = α NGAL +β NGAL ×[NGAL], the model outputs coefficient α NGAL β NGAL α NGAL This represents the log-odds predicted by the model when the biomarker concentration is 0; β NGAL This represents how much the risk of MPE increases on average for every unit increase in NGAL concentration.

[0054] Using the same method, construct Logistic(P) CEA ) = α CEA +β CEA ×[CEA],Logistic(P) CA50 ) = α CA50 +β CA50 ×[CA50]; obtain α CEA β CEA and α CA50 β CA50 .

[0055] S20. Calculate the likelihood ratio (LR). NGAL =exp[β NGAL ×(x NGAL -x′ NGAL )], where x NGAL Let x′ be the concentration of NGAL in the sample. NGAL This represents the concentration of NGAL when DOR is 1. LR CEA =exp[β CEA ×(x CEA -x′ CEA )], where x CEA Let x′ be the concentration of CEA in the sample. CEA This represents the concentration of CEA when DOR is 1. LR CA50 =exp[β CA50 ×(x CA50 -x′ CA50 )], where x CA50 x′ represents the concentration of CA50 in the sample. CA50 This represents the concentration of CA50 when DOR is 1.

[0056] For example, in this embodiment, the Bayesian parameters of CA50, NGAL, and CEA are shown in Table 2; the likelihoods of each biomarker at different concentrations are as follows: Figure 6 As shown.

[0057] Table 2. Bayesian parameters of CA50, NGAL, and CEA

[0058] S30. The prior probability of malignant pleural effusion (MPE) is 23.7% in patients with pleural effusion, according to a multicenter study (Tian P, Qiu R, Wang M, Xu S, Cao L, Yang P, Li W: Prevalence, Causes, and Health Care Burden of Pleural Effusions Among Hospitalized Adults in China. JAMA Netw Open 2021, 4(8):e2120306). The prior ratio of MPE is calculated based on the prior probability: prior ratio = prior probability / (1 - prior probability).

[0059] S40. Based on the likelihood ratio LR obtained in step S20 and the prior ratio obtained in step S30, calculate the posterior ratio: posterior ratio = prior ratio × LR NGAL ×LR CEA ×LR CA50 .

[0060] The posterior probability is calculated based on the posterior ratio: posterior probability = posterior ratio / (1 + posterior ratio).

[0061] Bayesian analysis can reveal the contribution of each biomarker's detection value in calculating the posterior probability within a biomarker combination. The following example illustrates this using a pleural effusion sample with CEA at 45 ng / ml, CA50 at 15 IU / mL, and NGAL at 55 ng / mL.

[0062] In the pleural effusion sample, CEA was 45 ng / ml, CA50 was 15 IU / mL, and NGAL was 55 ng / mL.

[0063] Calculate the likelihood ratio: LR NGAL =exp[0.059(55-33.054)]=3.649; LR CEA =exp[0.059(45-15.907)]=5.568; LR CA50=exp [0.161(15-5.397)]=4.695. By calculating the likelihood ratio, the effect of the difference between the concentration measurement of each biomarker and x′ on the predicted probability of MPE can be determined. In other words, in this sample, when GAL is 55 ng / mL, the likelihood ratio represents the exponent that the probability of MPE increases due to the increase in GAL concentration. The larger the likelihood ratio value, the larger the probability exponent and the greater the predicted probability.

[0064] Calculate the prior ratio: Prior ratio = 0.237 / (1 - 0.237) = 0.31.

[0065] Calculate the posterior ratio: Posterior ratio = Prior ratio × LR CEA ×LR NGAL ×LR CA50 =0.31×3.649×5.568×4.695=29.571.

[0066] Calculate the posterior probability: Posterior probability = posterior ratio / (1 + posterior ratio) = 29.571 / (1 + 29.571) = 0.967; that is, the MPE prediction probability of this sample is 96.7%.

[0067] Another embodiment of the present invention provides a storage medium storing a computer program for executing the steps of the methods described in the above embodiments. Yet another aspect of the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor, when invoking the computer program in the memory, implements the steps of the methods described in the above embodiments. Specifically, the integrated modules implemented as software functional modules can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0068] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A combination of biomarkers for the in vitro differential diagnosis of malignant pleural effusion, characterized in that, The biomarker combination consists of NGAL, CEA and CA50, and the in vitro identified sample is pleural effusion.

2. The use of the reagent for detecting the combination of biomarkers according to claim 1 in the preparation of a kit for diagnosing malignant pleural effusion.

3. A method for constructing a predictive model for malignant pleural effusion, characterized in that, Includes the following steps: S1. Obtain the biomarker combination of the training set samples, and perform feature extraction on the biomarker combination of the training set samples to obtain the features of the biomarker combination; the biomarker combination consists of three biomarkers: NGAL, CEA and CA50. S2. Based on the characteristics of the biomarker combination, logistic regression analysis was used to construct regression models for the three biomarkers respectively, and the parameters α were output respectively. NGAL β NGAL α CEA β CEA α CA50 and β CA50 ; Where, α NGAL α CEA α CA50 These are the intercept terms of the regression models for the three biomarkers NGAL, CEA, and CA50, respectively, and β is the value of β. NGAL β CEA β CA50 These are the regression coefficients of the regression models corresponding to the three biomarkers NGAL, CEA, and CA50, respectively. S3. Based on the output parameters in step S2, a prediction model for malignant pleural effusion is constructed using Bayesian analysis.

4. The method for constructing a predictive model for malignant pleural effusion according to claim 3, characterized in that, In step S1, the training set samples include malignant pleural effusion samples and benign pleural effusion samples; The malignant pleural effusion samples include lung cancer pleural effusion samples, malignant pleural mesothelioma pleural effusion samples, breast cancer pleural effusion samples, gastric cancer pleural effusion samples, and lymphoma pleural effusion samples; The benign pleural effusion samples include those from parapneumonia, tuberculous pleural effusion, heart failure, pulmonary embolism, mixed connective tissue disease, cirrhosis, pneumothorax, idiopathic pleural fibroelastosis, interstitial lung disease, pulmonary thromboembolism, and hypoalbuminemia.

5. The method for constructing a predictive model for malignant pleural effusion according to claim 4, characterized in that, In step S1, the biomarker combination is characterized by the concentrations of NGAL, CEA, and CA50 in the training set samples, respectively. The concentration of NGAL was obtained using Luminex liquid phase chip technology. The concentrations of CEA and CA50 were obtained by chemiluminescent immunoassay.

6. The method for constructing a predictive model for malignant pleural effusion according to claim 3, characterized in that, In step S3, the prediction probability expression of the Bayesian model is: P = 0.31 × LR CEA ×LR NGAL ×LR CA50 / (1+0.31×LR CEA ×LR NGAL ×LR CA50 ); Among them, LR CEA LR NGAL LR CA50 These are the likelihood ratios for CEA, NGAL, and CA50, respectively.

7. A method for processing isolated pleural effusion data, characterized in that, include: The characteristics of a combination of biomarkers in a sample to be tested are detected, the combination of biomarkers consisting of NGAL, CEA and CA50; By inputting the features of the biomarker combination into the malignant pleural effusion prediction model constructed according to any one of claims 3-6, the risk value of malignant pleural effusion is obtained.

8. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor calls the computer program in the memory, it implements the steps of the method for constructing the malignant pleural effusion prediction model according to any one of claims 3-6.

9. A storage medium, characterized in that, The storage medium stores a computer program for executing the steps of the method for constructing the malignant pleural effusion prediction model according to any one of claims 3-6.

Images