Biomarker for predicting curative effect of non-small cell lung cancer immunotherapy and application thereof

Abstract

The invention relates to a biomarker for predicting the curative effect of non-small cell lung cancer immunotherapy and application of the biomarker. The biomarker for predicting the curative effect of the non-small cell lung cancer immunotherapy, disclosed by the invention, comprises phosphoglycerate kinase 1 (PGK1), chromogranin A (CHGA), elastin (ELN), heterogeneous ribonucleoprotein H3 (HNRNPH3) and SERPINE1 mRNA binding protein 1 (SERBP1). The biomarkers for predicting the non-small cell lung cancer immunotherapy curative effect can be detected by adopting a mass spectrum method, and the non-small cell lung cancer immunotherapy curative effect of a patient is judged by the markers.

Description

Technical Field

[0001] The present invention relates to biomarkers for predicting the efficacy of immunotherapy for non-small cell lung cancer and applications thereof. Background Art

[0002] Lung cancer is the leading cause of cancer-related morbidity and mortality in my country. Programmed cell death protein 1 / programmed cell death ligand 1 (PD-(L)1) monoclonal antibodies combined with chemotherapy have been approved as first-line treatment for advanced non-small cell lung cancer (NSCLC) (PD-1 monoclonal antibodies include pembrolizumab, carrelizumab, tislelizumab, sintilimab, toripalimab, slulizumab, penampalimab, etc., and PD-L1 monoclonal antibodies include atezolizumab, sugemalimab, etc.), which can enable some patients to achieve long-term survival, but the population that clinically benefits is limited. How to screen more people who benefit from immunotherapy remains a clinical problem that needs to be solved urgently.

[0003] Proteograph proteomics is an innovative plasma proteome high-depth analysis system developed by Seer, USA, which combines ProteographTM with a mass spectrometer to perform unbiased and in-depth detection of proteins in serum plasma. In proteomics research, parallel reaction monitoring (PRM) is often used to perform targeted detection or screening of target proteins. PRM is similar to multiple reaction monitoring (MRM) technology. The difference between the two is that PRM uses a high-resolution mass spectrometer (Q-Orbitrap / timsTOF), which can significantly reduce background impurity interference and improve the selectivity and specificity of detection.

[0004] At present, the main biomarkers for predicting the efficacy of immunotherapy for NSCLC are PD-L1 expression and tumor mutational burden (TMB). Although the former has been verified by prospective clinical studies, its predictive efficacy is limited due to factors such as tissue heterogeneity, detection platform, and detection threshold, and it is not an ideal predictive marker. Similarly, the latter lacks the support of prospective data, and cannot well predict the efficacy of immunotherapy due to difficulties in tissue acquisition, different detection panels, and differences in thresholds. Current clinical practice tells us that PD-L1 and TMB cannot effectively predict the efficacy of immunotherapy. Other biomarkers are needed clinically to expand the population that benefits. Summary of the invention

[0005] The present invention provides novel biomarkers for predicting the efficacy of immunotherapy for non-small cell lung cancer. The biomarkers for predicting the efficacy of immunotherapy for non-small cell lung cancer of the present invention include phosphoglycerate kinase 1 (PGK1), chromogranin A (CHGA), elastin (ELN), heterogeneous nuclear ribonucleoprotein H3 (HNRNPH3) and SERPINE1 mRNA binding protein 1 (SERBP1).

[0006] Specifically, the present invention provides a method for predicting the efficacy of immunotherapy for non-small cell lung cancer, which comprises the step of detecting the contents of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 in the plasma of a subject.

[0007] In one or more embodiments, the detection method comprises: ELISA, suspension chip, MSD electrochemiluminescence technology or mass spectrometry. Preferably, the detection method is mass spectrometry.

[0008] In one or more embodiments, mass spectrometry is used to detect the signal intensity of PGK1, CHGA, ELN, HNRNPH3, and SERBP1 in the subject's plasma.

[0009] In one or more embodiments, after the signal intensities of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 are detected, the signal intensities are substituted into the following formula:

[0010] Score = -0.011 x PGK1 - 0.027 x CHGA - 0.068 x ELN + 0.039 x HNRNPH3 + 0.018 x SERBP1. In one or more embodiments, when the plasma level of the marker is detected using the PRM method, the threshold value is -4.6851.

[0011] If the score is less than or equal to -4.6851, the patient does not respond to immunotherapy for non-small cell lung cancer; if the judgment threshold is greater than -4.6851, the patient responds to immunotherapy for non-small cell lung cancer.

[0012] The present invention also provides the use of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 as biomarkers or their detection reagents in the preparation of reagents or kits for predicting the efficacy of immunotherapy for non-small cell lung cancer.

[0013] In one or more embodiments, the reagent detects the amount of the protein in plasma in the sample.

[0014] In one or more embodiments, the prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma level of the marker is:

[0015] Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

[0016] In one or more embodiments, when the PRM method is used to detect the plasma level of the marker, the threshold value is -4.6851.

[0017] In one or more embodiments, the reagent is a reagent used for detecting proteins using any of the following methods: ELISA, suspension chip, MSD electrochemiluminescence technology or mass spectrometry. Preferably, the detection method is mass spectrometry.

[0018] In one or more embodiments, the reagent is a mobile phase used for mass spectrometry detection, including mobile phase A and mobile phase B.

[0019] In one or more embodiments, the reagents include formic acid in mobile phase A and formic acid acetonitrile in mobile phase B.

[0020] In one or more embodiments, the concentration of the formic acid aqueous solution in the mobile phase A is 0.1%, and the concentration of the formic acid acetonitrile solution in the mobile phase B is 0.1%.

[0021] In one or more embodiments, the agent is an antibody.

[0022] In one or more embodiments, the reagent is a target polypeptide for quantifying protein using a PRM method.

[0023] In one or more embodiments, the reagents further include reagents for separating proteins from plasma.

[0024] The present invention also provides a separated protein combination, which is a marker for the efficacy of immunotherapy for non-small cell lung cancer. The protein combination comprises PGK1, CHGA, ELN, HNRNPH3 and SERBP1.

[0025] In one or more embodiments, the amino acid sequence of PGK1 is as shown in SEQ ID NO:1.

[0026] In one or more embodiments, the amino acid sequence of CHGA is as shown in SEQ ID NO:2.

[0027] In one or more embodiments, the amino acid sequence of ELN is shown in SEQ ID NO:3.

[0028] In one or more embodiments, the amino acid sequence of HNRNPH3 is as shown in SEQ ID NO:4.

[0029] In one or more embodiments, the amino acid sequence of SERBP1 is as shown in SEQ ID NO:5.

[0030] The present invention also provides a reagent for detecting the contents of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 in plasma.

[0031] In one or more embodiments, the reagent is a reagent used for detecting proteins using any of the following methods: ELISA, suspension chip, MSD electrochemiluminescence technology or mass spectrometry. Preferably, the detection method is mass spectrometry.

[0032] In one or more embodiments, the reagent is a mobile phase used for mass spectrometry detection, including mobile phase A and mobile phase B.

[0033] In one or more embodiments, the reagents include formic acid in mobile phase A and formic acid acetonitrile in mobile phase B.

[0034] In one or more embodiments, the concentration of the formic acid aqueous solution in the mobile phase A is 0.1%, and the concentration of the formic acid acetonitrile solution in the mobile phase B is 0.1%.

[0035] In one or more embodiments, the agent is an antibody.

[0036] In one or more embodiments, the reagent is a target polypeptide for quantifying protein using a PRM method.

[0037] In one or more embodiments, the reagents further include reagents for separating proteins from plasma.

[0038] The present invention also provides a kit comprising the reagent described in any embodiment of the present invention.

[0039] In one or more embodiments, the kit further comprises a protein combination as described in any embodiment herein.

[0040] In one or more embodiments, the kit further comprises a reagent for detecting an internal standard.

[0041] Use of the reagent described in any embodiment of the present invention in the preparation of a reagent or kit for predicting the efficacy of immunotherapy for non-small cell lung cancer.

[0042] Use of the reagent described in any embodiment of the present invention and the protein combination described in any embodiment of the present invention in the preparation of a reagent or a kit for predicting the efficacy of immunotherapy for non-small cell lung cancer.

[0043] In one or more embodiments, the kit is a PRM detection kit.

[0044] In one or more embodiments, the reagents include:

[0045] (1) using mass spectrometry to detect the reagents required for the PGK1, CHGA, ELN, HNRNPH3 and SERBP1 and the optional internal standard; preferably, the reagents are mobile phases used for mass spectrometry detection, including mobile phase A and mobile phase B; more preferably, the reagents include a formic acid aqueous solution in mobile phase A and a formic acid acetonitrile solution in mobile phase B; and

[0046] (2) Optionally, a reagent for separating plasma proteins from plasma, preferably selected from alcohol reagents and reagents used for salting out, such as methanol.

[0047] The present invention further provides a device, characterized in that the device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the program, the following steps are implemented:

[0048] (1) Obtain the content of the following proteins in plasma samples: PGK1, CHGA, ELN, HNRNPH3 and SERBP1,

[0049] (2) obtaining a score by constructing a model using the content, and

[0050] (3) Predict the efficacy of immunotherapy for non-small cell lung cancer based on the score.

[0051] In one or more embodiments, the prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma level of the marker is:

[0052] Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

[0053] The present invention also provides a system for predicting the efficacy of immunotherapy for non-small cell lung cancer, characterized in that it comprises:

[0054] The collection device is used to obtain the content of the following proteins in the plasma sample: PGK1, CHGA, ELN, HNRNPH3 and SERBP1,

[0055] A data processing device for obtaining a score by constructing a model using the content,

[0056] A determination device is used to predict the efficacy of immunotherapy for non-small cell lung cancer based on the score.

[0057] In one or more embodiments, the prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma level of the marker is:

[0058] Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

[0059] The present invention also provides a method for constructing a model for predicting the efficacy of immunotherapy for non-small cell lung cancer, comprising the steps of:

[0060] (1) Obtaining the levels of the following proteins in plasma samples: PGK1, CHGA, ELN, HNRNPH3, and SERBP1, and

[0061] (2) Calculate the ROC curve and area under the curve of the protein combination, and establish a model using logistic regression statistical analysis method. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 The protein levels of PGK1, CHGA, ELN, HNRNPH3, and SERBP1 were significantly different between the immunotherapy responder and non-responder groups.

[0063] Figure 2 ROC curve plot constructed for the model.

[0064] Figure 3 ROC curve diagram for model validation. DETAILED DESCRIPTION

[0065] It should be understood that within the scope of the present invention, the above-mentioned technical features of the present invention and the technical features specifically described below (such as embodiments) can be combined with each other to form a preferred technical solution.

[0066] The inventors collected plasma from patients with advanced NSCLC and used non-targeted proteomics for research. Finally, they found biomarkers that can predict the efficacy of immunotherapy for non-small cell lung cancer through machine learning-based algorithms. Specifically, the present invention found that phosphoglycerate kinase 1 (PGK1), chromogranin A (CHGA), elastin (ELN), heterogeneous nuclear ribonucleoprotein H3 (HNRNPH3) and SERPINE1mRNA binding protein 1 (SERBP1) can be used as specific biomarkers for predicting the efficacy of immunotherapy for non-small cell lung cancer. The efficacy of immunotherapy for non-small cell lung cancer in a subject can be diagnosed by detecting the content of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 in the subject's plasma.

[0067] Quantitative detection of proteins is a commonly used method in molecular biology. Through specific technical methods, the content of one or more proteins in a sample is accurately determined. Quantitative detection of proteins is mainly carried out from the perspectives of spectral analysis, chemical analysis, immunological analysis, chromatographic analysis, etc. Specifically, the method of quantitative detection of proteins includes but is not limited to: ELISA, suspension chip, MSD electrochemiluminescence technology, biuret method, BCA method, Lowry method, Bradford method, ultraviolet spectroscopy. For example, in the present invention, a variety of proteins are first screened, and then the signal intensity of the screened PGK1, CHGA, ELN, HNRNPH3 and SERBP1 proteins is determined by mass spectrometry, and then the proteomics data is searched in the library.

[0068] Herein, immunotherapy includes immune checkpoint inhibitor therapy. In an exemplary embodiment, immune checkpoint inhibitor therapy is PD-1 or PD-L1 inhibitor therapy. In an exemplary embodiment, immunotherapy herein includes the use of PD-1 / PD-L1 inhibitors and chemotherapeutic drugs. Chemotherapeutic drugs include, but are not limited to, one or more of pemetrexed, paclitaxel, albumin paclitaxel, gemcitabine, and platinum drugs, for example, pemetrexed + platinum, paclitaxel / albumin paclitaxel + platinum, or gemcitabine + platinum.

[0069] Herein, "amount", "content", "concentration" and the like include both absolute amount and relative amount. Herein, "amount", "content" and "concentration" may be standardized.

[0070] The "marker" described herein can be a nucleic acid molecule or its protein expression product. Methods for detecting protein levels are well known in the art, such as ELISA, suspension chip, MSD electrochemiluminescence technology or mass spectrometry. In certain embodiments, the present invention relates to detecting the level of marker protein from a sample derived from blood.

[0071] Therefore, the present invention relates to reagents for detecting protein expression levels. The reagents used in the above-mentioned methods for detecting protein levels are well known in the art. In one or more embodiments, the reagent is a mobile phase used for mass spectrometry detection, such as formic acid in mobile phase A and formic acid acetonitrile in mobile phase B, and its concentration is conventional in the art. The reagent for detecting protein levels can also be an antibody, such as a specific antibody for PGK1, CHGA, ELN, HNRNPH3 and SERBP1.

[0072] In an exemplary embodiment, the present invention uses the PRM method to detect plasma proteins. The steps and required reagents of conventional PRM methods are known in the art, including mass spectrometry reagents and specific peptides used for quantifying proteins by the PRM method. Designing specific peptides based on protein sequences is a routine technique in the art.

[0073] Therefore, the present invention also relates to a method for pre-treating a sample. Samples derived from blood include, but are not limited to, whole blood, plasma, and serum. Those skilled in the art know methods for pre-treating a sample to obtain components (e.g., nucleic acids, proteins) containing markers to be tested, such as plasma protein extraction kits, Proteograph workstations, etc. Therefore, the reagents herein also include reagents for separating proteins from plasma.

[0074] Herein, the sample is from a mammal, preferably a human. The sample may be from any organ (e.g., lung), tissue (e.g., epithelial tissue), cell (e.g., tumor cell), or body fluid (e.g., blood, plasma, serum, tissue fluid, urine). In an exemplary embodiment, the sample is plasma.

[0075] The present invention also relates to a kit for predicting the efficacy of immunotherapy for non-small cell lung cancer, comprising the reagents described herein. The kit may also include the markers described herein as internal standards or positive controls. In addition to the reagents, the kit also includes other reagents required for detecting proteins. Exemplarily, the kit may include one or more of the following: tools and / or reagents for collecting samples from subjects, tools and / or reagents suitable for long-term storage of samples and / or controls, tools and / or reagents for pre-treating samples and / or controls.

[0076] Mass spectrometry

[0077] The term "mass spectrometer" refers to a mass spectrometer (MS), which is an instrument for weighing the "mass" of molecules. In proteomics research, the "mass" of molecules that need to be weighed by mass spectrometry generally includes the mass-to-charge ratio (m / z) of the precursor ion and the mass-to-charge ratio of the fragment ion after the precursor ion is broken down. The appearance of mass spectrometers from different manufacturers and different types varies, but their internal structure mainly includes three parts: ion source, mass analyzer, and detector. The function of the ion source is to charge the molecules to be tested and enter the mass spectrometer for scanning. In proteomics research, the main ionization method used is electrospray ionization (ESI); the mass analyzer is the core component of the mass spectrometer, and its function is to distinguish the mass size (m / z) of the molecules to be tested so as to separate or scan them; common mass analyzers include quadrupole, ion trap, time-of-flight (TOF), and electrostatic field orbital trap (Orbitrap), among which TOF and Orbitrap are high-resolution mass analyzers; the function of the detector is to convert and record the signals of the molecules to be tested to obtain the raw data of the mass spectrometer.

[0078] According to the ion source, mass spectrometers can be divided into electron impact ion source (EI) mass spectrometer, chemical ionization (CI) mass spectrometer, electrospray ionization (ESI) mass spectrometer, atmospheric pressure chemical ionization (APCI) mass spectrometer, matrix-assisted laser desorption ionization (MALDI) mass spectrometer, etc.; according to the mass analyzer, they can be divided into magnetic mass analyzer mass spectrometer, quadrupole mass analyzer mass spectrometer, ion trap mass analyzer mass spectrometer, time-of-flight mass analyzer (TOF) mass spectrometer, Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer; according to the application field, they can be divided into organic mass spectrometer, inorganic mass spectrometer, biological mass spectrometer; according to the scanning mode, they can be divided into full scan mode, selected ion monitoring mode (SIM), multiple reaction monitoring mode (MRM), parallel reaction monitoring mode (PRM), etc.

[0079] In the field of mass spectrometry, there are also different data acquisition modes and types, such as DDA data (Data-Dependent Acquisition), DIA data (Data-Independent Acquisition), SIM data (Selected Ion Monitoring Data), MRM data (Multi-Reaction Monitoring Data), and PRM data (Parallel Reaction Monitoring Data). According to the purpose of analysis, sample type, and requirements for data quality, you can choose the appropriate data acquisition mode to obtain the corresponding data.

[0080] In this article, the LC-MS / MS mass spectrometer is used, which is an analytical instrument that combines liquid chromatography (LC) with tandem mass spectrometry (MS / MS): the sample is separated by liquid chromatography, which uses the difference in the distribution coefficients of different substances between the stationary phase and the mobile phase to separate the complex sample mixture into various components, so that they flow out in sequence according to different retention times. Each component flowing out of the liquid chromatography enters the ion source of the mass spectrometer and is ionized, producing ions with a certain charge and different mass numbers. These ions enter the tandem mass spectrometry system, and the mass analyzer separates the ions according to different mass-to-charge ratios (m / z), obtaining a mass spectrum arranged in order of mass-to-charge ratio. By analyzing and processing the mass spectrum, the qualitative and quantitative results of the sample can be obtained.

[0081] Generally, mass spectrometry can be used to detect the abundance of ion responses of phosphoglycerate kinase 1 (PGK1), chromogranin A (CHGA), elastin (ELN), heterogeneous nuclear ribonucleoprotein H3 (HNRNPH3) and SERPINE1mRNA binding protein 1 (SERBP1) in the plasma of the subject. For example, peripheral blood of the subject can be obtained, and plasma can be separated by conventional methods. Afterwards, the protein in the plasma can be removed. Methods for removing protein from plasma include, but are not limited to, salting out, organic solvent precipitation, and pH adjustment. For example, neutral salts such as ammonium acetate can be used to remove the hydration layer of the protein and aggregate and precipitate. Alternatively, polar organic solvents can be used to reduce the dielectric constant to cause protein agglutination and precipitation. Commonly used organic solvents include alcohol organic solvents, such as monohydric alcohols with 1-4 carbon atoms such as methanol and ethanol, and also include acetone, acetonitrile, etc. In certain embodiments, sulfosalicylic acid can also be used to precipitate protein. When using sulfosalicylic acid, the pH value needs to be adjusted to be lower than the pI value of plasma protein, so that the protein is positively charged, thereby combining with salicylic acid radicals to form insoluble salts and precipitate. Since plasma proteins are all negatively charged, the pH of plasma is adjusted to below the lowest isoelectric point of plasma proteins. For example, the isoelectric point of plasma albumin is 4.7-4.9, so the pH can be adjusted to below 4.7 to achieve protein precipitation and separation.

[0082] After the protein in the plasma is precipitated, the supernatant is obtained and mass spectrometry can be performed. Mass spectrometry techniques known in the art can be used for detection. In an embodiment of the present invention, a TOF mass spectrometer is used for mass spectrometry detection. The chromatographic column can use a conventional chromatographic column, such as a chromatographic column provided by a mass spectrometry company. Different mobile phases can be selected according to different chromatographic columns. For example, in an embodiment of the present invention, the present invention uses an acetonitrile-water-formic acid system, wherein mobile phase A is a 0.1% formic acid aqueous solution and phase B is a 0.1% formic acid acetonitrile solution. After the chromatographic column is balanced with 100% A phase, the sample is directly loaded onto the chromatographic column by an automatic sampler, and then separated by gradient separation of the chromatographic column, with a flow rate of 300nL / min and a gradient duration of 60min. Mobile phase B ratio: 2% for 0min, 2-22% for 45min, 22-37% for 5min, 37-80% for 5min, 80% for 5min.

[0083] Other mass spectrometry operations can be performed according to the operation manual, thereby obtaining the ion response abundance of PGK1, CHGA, ELN, HNRNPH3 and SERBP1.

[0084] In this paper, during the sample preparation stage, the DDA data of the sample is first collected to obtain full scan information with a large amount of data, which includes the daughter ion information of the selected parent ion, which is helpful for the structural identification and analysis of the compound.

[0085] The term "parallel reaction monitoring" (PRM) refers to a targeted quantitative analysis technique based on high-resolution mass spectrometry. In PRM, the characteristic ions of the target compound must first be determined, including parent ions and daughter ions. Parent ions are the initial ions formed after the target compound is ionized, and daughter ions are fragment ions produced by processes such as collision-induced dissociation (CID) of the parent ion in the collision chamber of the mass spectrometer. PRM monitors the transition from parent ions to daughter ions of the target compound within a selected mass window. When the target compound in the sample enters the mass spectrometer, the instrument automatically monitors these specific ion reactions to achieve specific detection of the target compound. This monitoring mode is similar to multiple reaction monitoring (MRM), but PRM uses a high-resolution mass spectrometer (Q-Orbitrap / timsTOF), which can significantly reduce background impurity interference and improve the selectivity and specificity of the detection. Usually, 1-3 specific peptides are selected for the protein to be tested, and 3-6 fragment ions are selected for each target peptide for quantification.

[0086] The construction of the PRM method includes: analyzing the collected DDA mass spectrometry data, and then using SpectroDive software to generate the corresponding library to determine the specific peptides and fragment ions that need to be detected, completing the construction of the PRM method. Then the PRM data is collected. This set of data can effectively eliminate matrix interference and accurately detect target compounds. At the same time, it also has high resolution and mass accuracy, and can collect more accurate ion mass information.

[0087] In this paper, proteins were first enriched, enzymatically hydrolyzed, desalted, and concentrated from plasma, followed by nanoLC-MS / MS detection, DDA library search identification, PRM method construction, and quantitative analysis. The PRM method construction was completed using SpectroDive software: 1) The collected DDA data was matched to the target peptide sequence and the library was searched using SpectroDive software to generate the corresponding library; 2) Panel screening peptides and confirmation of quantitative secondary b / y ions were generated based on the library and target peptides. A maximum of 3 specific peptides were selected for each protein, and 3 to 6 fragment ions were selected for each target peptide for quantification. In an exemplary embodiment, in the PRM detection method, the peptide segment of PGK1 is a fragment of PGK1 comprising the sequence shown in SEQ ID NO:1, preferably as shown in SEQ ID NO:1; the peptide segment of CHGA is a fragment of CHGA comprising the sequence shown in SEQ ID NO:2, preferably as shown in SEQ ID NO:2; the peptide segment of ELN is a fragment of ELN comprising the sequence shown in SEQ ID NO:3, preferably as shown in SEQ ID NO:3; the peptide segment of HNRNPH3 is a fragment of ELN comprising the sequence shown in SEQID NO:4, preferably as shown in SEQ ID NO:4; the peptide segment of SERBP1 is a fragment of SERBP1 comprising the sequence shown in SEQ ID NO:5, preferably as shown in SEQ ID NO:5.

[0088] Bioinformatics analysis

[0089] The term "bioinformatics analysis" is a discipline that uses mathematics, computer science and biology to process, analyze and interpret biological data. Usually biological data refers to data related to physiological and biochemical processes such as genomic data, transcriptomic data, proteomic data, and metabolomic data. Sequence alignment, gene annotation, expression analysis, cluster analysis or pathway analysis of these data are helpful for disease research and diagnosis, drug development, and the promotion of medical device development.

[0090] Commonly used analysis methods in the field of bioinformatics include: (1) Principal Component Analysis (PCA): multiple variables of the original data are converted into a few principal components through linear transformation. These principal components are linear combinations of the original variables; (2) Hierarchical Clustering Analysis (HCA): cluster dendrograms are constructed based on similarity or distance metrics between data points (such as genes, proteins or samples); (3) K-means Clustering: This method is suitable for rapid classification of a large number of biological samples and classification of gene expression data; (4) Survival Analysis: In cancer research, by analyzing the clinical characteristics and gene expression data of patients, the survival prognosis of patients is evaluated or the effects of different treatments on patient survival are compared; (5) Gene Enrichment Analysis: By comparing the observed number of genes in the gene set with the expected number, it is determined whether the gene enrichment is statistically significant. It can usually be used to determine biological pathways and functional categories related to diseases and help screen potential drug targets; (6) Support Vector Machine (SVM) Support Vector Machine (SVM): is a supervised machine learning algorithm used for classification and regression analysis.

[0091] ROC Curve

[0092] The term "ROC curve" refers to the Receiver Operating Characteristic Curve, which is a tool used to evaluate the performance of classification models and is widely used in bioinformatics and medicine. The ROC curve helps us understand the performance of the classifier at different thresholds by plotting the relationship between the True Positive Rate (TPR) and the False Positive Rate (FPR). The True Positive Rate (TPR), or sensitivity, represents the proportion of samples correctly identified as positive to all actual positive samples, and its calculation formula is:

[0093]

[0094] Among them, TP (True Positive) is the number of samples that are actually positive and predicted as positive by the model, and FN (False Negative) is the number of samples that are actually positive but predicted as negative by the model. The false positive rate (FPR) indicates the proportion of samples that are mistakenly identified as positive to all actual negative samples, and its calculation formula is:

[0095]

[0096] Among them, FP (False Positive) is the number of samples that are actually negative examples but predicted as positive examples by the model, and TN (True Negative) is the number of samples that are actually negative examples and predicted as negative examples by the model.

[0097] The TPR and FPR plotted by changing the classification threshold, the closer the curve is to the upper left corner, the better the performance of the classifier. Ideally, the ROC curve will pass through the (0,1) point, which means that the model is able to distinguish between positive and negative examples, that is, FPR = 0 (no false positives) and TPR = 1 (all positive examples are correctly identified). If the ROC curve is a diagonal line from the origin (0,0) to the point (1,1), it means that the model's predictions are random and have no actual classification ability. In this case, for any given true positive rate, the false positive rate is the same, similar to random guessing.

[0098] The ROC curve provides an intuitive way to evaluate the performance of classification models (such as the classification of whether a gene is a disease-related gene or a protein is a functional protein in bioinformatics). By observing the position of the curve in the coordinate plane, the quality of the model can be quickly judged. For example, in cancer biomarker screening, the ROC curve can be used to evaluate the model that distinguishes cancer patients from healthy people based on the concentration or signal intensity of biomarkers in the blood. If the ROC curve is close to the upper left corner, it means that a higher true positive rate can be obtained at a lower false positive rate, and the model can identify cancer patients well. In practical applications, it is necessary to determine a suitable classification threshold to divide samples into positive and negative examples. The ROC curve can help researchers understand the performance under different thresholds. For example, in microbial identification, a classification model is constructed based on the gene sequence characteristics of the microorganism. Through the ROC curve, the changes in the true positive rate and false positive rate of the model in identifying microbial species under different sequence similarity thresholds can be observed. According to actual needs (such as the tolerance for false positives and true positives), a suitable threshold can be selected from the ROC curve for classification.

[0099] AUC (Area Under the Curve) is the area under the ROC curve. It is a quantitative indicator used to measure the performance of the model more intuitively. The value range of AUC is from 0.5 to 1. The closer the AUC is to 1, the stronger the model's ability to distinguish. For example, in the prediction of gene regulatory networks, if a model predicting the interaction between transcription factors and target genes has a high AUC value, it means that it can better identify truly interacting gene pairs and non-interacting gene pairs. After a series of true positive rate (TPR) and false positive rate (FPR) data points have been calculated to draw the ROC curve, the trapezoidal method can be used to approximate the area under the curve. This is based on the principle of dividing the area under the ROC curve into multiple small trapezoids and then calculating the sum of the areas of these trapezoids. It is now usually combined with a computer to calculate the AUC value using built-in functions of statistical software or programming languages ​​(such as R language and Python language).

[0100] In one or more embodiments, when compared with the control sample, the expression level of the subject sample increases or decreases. The expression of the measured protein is mathematically analyzed to obtain a score. For the sample detected, when the score is greater than the threshold, the result is determined to be positive, that is, the immunotherapy effect of non-small cell lung cancer is better (responding to the immunotherapy of non-small cell lung cancer), otherwise it is negative, that is, the immunotherapy effect of non-small cell lung cancer is poor (not responding to the immunotherapy of non-small cell lung cancer). The method of conventional mathematical analysis and the process of determining the threshold value are known in the art, and an exemplary method is a mathematical model, such as a logistic regression model, a support vector machine and a random forest model. Those skilled in the art know the conventional method of constructing a logistic regression model. For example, for differentially expressed markers, a logistic regression model, a support vector machine model, or a random forest model is constructed for two groups of samples, and the accuracy, sensitivity and specificity of the test results and the area under the predictive characteristic curve (ROC) (AUC) of the model statistics are used to statistically test the test set sample prediction score.

[0101] An exemplary logistic regression model is as follows (wherein each protein name indicates its plasma content):

[0102] Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

[0103] When the plasma content of the marker is detected using the PRM method, the threshold for determining the efficacy of immunotherapy for non-small cell lung cancer is -4.6851. That is, if the score is less than or equal to -4.6851, the patient does not respond to immunotherapy for non-small cell lung cancer; if the score is greater than -4.6851, the patient responds to immunotherapy for non-small cell lung cancer.

[0104] When other detection methods are used to detect the absolute or relative content of the protein in plasma, the judgment threshold will change due to the different presentation of the test results. However, the specific statistical relationship reflected by the model and the judgment results for the same patient will not change. When other methods are used to detect concentrations or relative concentrations, the threshold can be easily determined by those skilled in the art. For example, the threshold can be obtained by the following method: a) Collect samples of patients in the known non-small cell lung cancer immunotherapy response group and non-response group, and obtain the relative content of the combination of the 5 proteins in the above samples; b) Use the logistic regression statistical analysis method to construct an immunotherapy discrimination model for the protein combination, at which time the technicians in the field will obtain a model containing specific parameters and a judgment threshold; the more samples, the greater the sensitivity and specificity of the obtained model, and the more accurate the diagnostic results of the model. Then, the patient's responsiveness to immunotherapy for non-small cell lung cancer can be judged by the following steps: c) Obtain the relative content of the protein combination of the sample to be tested, the method may be the same as or different from a); d) According to the corresponding logistic regression model obtained in step b), calculate the model score of the sample, and then judge the responsiveness of immunotherapy for non-small cell lung cancer.

[0105] For the logistic regression statistical analysis method, the model and parameters obtained based on the content measured by the same batch of experimental results are unique. Moreover, based on the statistical relationship and statistical principles obtained in this application, it can be seen that the parameters (including coefficients, sensitivity, specificity, and accuracy) of the model obtained based on the content measured by different batches of experimental results may change slightly, but the specific statistical relationship reflected by the model will not change, that is, it has the ability to discriminate the efficacy of immunotherapy for non-small cell lung cancer.

[0106] After knowing the gene combination, technicians in this field can obtain a logistic regression model based on any non-small cell lung cancer immunotherapy sample, and the model also has the function of identifying the efficacy of immunotherapy for non-small cell lung cancer.

[0107] According to statistical principles, for a fixed gene combination, the statistical relationship of the model constructed by technicians in this field based on the above method after testing any samples that respond or do not respond to immunotherapy for non-small cell lung cancer is the same. Even if the parameters may change slightly, the model can still identify the efficacy of immunotherapy for non-small cell lung cancer.

[0108] Herein, the inventors found that the plasma levels of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 were related to the efficacy of immunotherapy for non-small cell lung cancer. After further research, a mathematical model was constructed to determine the efficacy of immunotherapy for non-small cell lung cancer. In one or more embodiments, the logistic regression prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma levels of the markers is:

[0109] Score = 4.988-0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1

[0110] The content of each protein is the relative concentration detected by PRM. The peptide levels of these five proteins were jointly analyzed and the scores of each sample were calculated using the above formula. See Table 4 for ROC curve analysis. Figure 2 , AUC was 0.884, threshold was 0.3029, specificity was 88.9%, and sensitivity was 90.5%.

[0111] The present invention also provides a device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the program: (1) obtaining the content of the following proteins in a plasma sample: PGK1, CHGA, ELN, HNRNPH3, and SERBP1, (2) obtaining a score using the content by constructing a prediction model for the efficacy of immunotherapy for non-small cell lung cancer as described herein, and (3) predicting the efficacy of immunotherapy for non-small cell lung cancer based on the score.

[0112] The present invention also provides a system for predicting the efficacy of immunotherapy for non-small cell lung cancer, comprising: a collection device for obtaining the content of the following proteins in a plasma sample: PGK1, CHGA, ELN, HNRNPH3 and SERBP1, a data processing device for obtaining a score using the content by constructing the non-small cell lung cancer immunotherapy efficacy prediction model described in this article, and a determination device for predicting the efficacy of immunotherapy for non-small cell lung cancer based on the score.

[0113] The present invention also provides a method for constructing a model for predicting the efficacy of immunotherapy for non-small cell lung cancer, comprising the steps of: (1) obtaining the content of the following proteins in a plasma sample: PGK1, CHGA, ELN, HNRNPH3 and SERBP1, and (2) calculating the ROC curve and the area under the curve value of the protein combination, and establishing a model using a logistic regression statistical analysis method.

[0114] It should be understood that the efficacy of immunotherapy for non-small cell lung cancer described in the present invention is relative. If the patient's detection threshold is less than or equal to -4.6851, it indicates that the patient's non-small cell lung cancer immunotherapy efficacy is poor (does not respond to immunotherapy), or relative to the subject whose judgment threshold is greater than -4.6851, the non-small cell lung cancer immunotherapy efficacy is better (responds to immunotherapy). When the patient's detection threshold is greater than -4.6851, but the patient shows symptoms of poor efficacy, it can be tested in combination with other conventional efficacy detection methods in the field as needed.

[0115] In addition, the above calculation formula and threshold value of the present invention are obtained based on the use of 250 μL plasma samples. Therefore, when the amount of plasma sample used is different, the coefficients and threshold values ​​of the above calculation formula will be different, but the corresponding coefficients and threshold values ​​can be easily determined by those skilled in the art according to the method disclosed in the specification of this application.

[0116] The present invention will be described below in the form of specific examples. It should be understood that these examples are merely illustrative and are not intended to limit the scope of the present invention. The methods and reagents used in the examples are, unless otherwise stated, conventional methods and reagents in the art.

[0117] Example

[0118] Example 1: High-depth screening of plasma protein corona proteomics

[0119] 1. Sample pretreatment

[0120] Prepare 250ul plasma sample and put it into the sample tube. Start the Proteograph automated workstation. The machine will automatically operate and complete the pre-treatment process, including: 1) nanoparticle NP-plasma incubation and protein corona formation; 2) magnetic bead purification of NP protein corona; 3) protein digestion and peptide purification.

[0121] 2. Mass spectrometry test

[0122] The desalted and lyophilized peptides were redissolved in 0.1% formic acid aqueous solution and analyzed by LC-MS / MS. The entire system was a timsTOF Pro 2 mass spectrometer (Bruker Daltonics) connected in series with an UltiMate 3000 system (Thermo Fisher Scientific, MA, USA). A total of 200 ng of sample was loaded (analytical column: 25 cm × 75 μm.d., Ion Opticks), and the sample was separated by a gradient of 80 min at a column temperature of 50 °C. A total of 2.5 μL of sample was loaded, and the column flow rate was controlled at 300 nL / min. The gradient started from 4% of phase B, increased to 28% in 45 min, increased to 44% in 10 min, and reached 90% in 10 min, maintained for 7 min, and 4% equilibrated for 8 min.

[0123] The mass spectrometer uses diaPASEF mode for DIA data acquisition, with a scan range of 349-1229 m / z and an isolation window width of 40 Da. During the PASEF MSMS scan, the collision energy increases linearly with the ion mobility from 59 eV (1 / K0 = 1.6 Vs / cm 2 ) rises to 20eV (1 / K0 = 0.6Vs / cm 2 ).

[0124] 3. Library search and protein quantification

[0125] The original data files were searched and analyzed using the software DIA-NN (version 1.8.1).

[0126] Table 1 Database matching parameter settings

[0127]

[0128]

[0129] 4. Experimental Results

[0130] By testing the baseline plasma samples of 5 patients with advanced NSCLC who were treated with PD-(L)1 monoclonal antibody combined with chemotherapy (responder, including complete response (CR) and partial response (PR)) and 5 patients who were non-responder (including stable disease (SD) and progressive disease (PD)), it was found that the levels of up to 222 proteins were different between the two groups. PGK1, CHGA, ELN, HNRNPH3 and SERBP1 were selected for subsequent experiments ( Figure 1 ).

[0131] Example 2, PRM quantitative proteomics validation

[0132] 1. Sample Preparation

[0133] 1.1 Protein enrichment

[0134] The protein enrichment and enzymatic hydrolysis steps were performed using the EasyPeptide Blood Low-Abundance Protein Enrichment and Pretreatment Kit (OSFP0002).

[0135] Before use, take out reagent H and disperse the magnetic beads evenly by blowing or ultrasound, transfer 20 μL of reagent H to a new EP tube (Note: 2 ml flat-bottom tube is recommended), place it on a magnetic stand and let it stand for 2 minutes, and discard the supernatant after magnetic separation;

[0136] Add 50 μL of reagent G, place on a constant temperature mixer, shake and wash the magnetic beads for 5 minutes (1300 rpm, room temperature), place on a magnetic stand and let stand for 2 minutes, magnetically separate and discard the supernatant;

[0137] Resuspend the magnetic beads with 50 μL of reagent G, add 50 μL of plasma, place on a constant temperature mixer, and incubate at 37°C for 1 h (1300 rpm);

[0138] Place on a magnetic stand and let stand for 2 minutes, discard the supernatant by magnetic separation, add 150 μL of reagent G, shake and wash for 5 minutes, repeat the washing three times, and the low-abundance protein enriched on the magnetic beads.

[0139] 1.2 Proteolysis

[0140] Add 20 μL of reagent A to the washed magnetic beads and mix well by pipetting;

[0141] Heating at 95°C for 5 min (95°C, 1000 rpm);

[0142] After heating, when the sample returns to room temperature, add 7.5 μL of reagent B, mix well, and perform enzymatic hydrolysis at 37°C for 2 h (37°C, 1300 rpm);

[0143] After the enzymatic hydrolysis is completed, add 1.5 μL of reagent C, mix well, and terminate the enzymatic hydrolysis reaction. At this time, a precipitate will appear. Centrifuge at 17000g for 2 minutes and then aspirate the supernatant for subsequent desalting.

[0144] 1.3 Peptide desalting

[0145] Desalting column activation: 100 μL of methanol was added to the desalting column and centrifuged at 700 g for 1 min;

[0146] Desalting column to remove impurities: 100 μL Condition buffer was added to the desalting column and centrifuged at 700 g for 1 min;

[0147] Desalting column balance: 100 μL Wash buffer was added to the desalting column and centrifuged at 700 g for 1 min;

[0148] Loading samples: Load all samples into the desalting column, centrifuge at 700g for 1 min, and repeat the operation once; (make sure pH <3 before loading samples, you can use 20% TFA to adjust pH <3 before loading samples)

[0149] Desalting column washing: add 100 μL Wash buffer to the desalting column, centrifuge at 700 g for 1 min, and repeat the operation once;

[0150] Sample elution: 30 μL of Elution buffer was added to the desalting column, centrifuged at 700 g for 1 min, and eluted twice, with a final sample volume of 60 μL;

[0151] Concentrate the samples using a vacuum refrigerated centrifugal concentrator.

[0152] 2. nanoLC-MS / MS detection

[0153] 3 μL of total peptides were taken from each sample and separated by nano-UPLC liquid phase system nanoElute2, and then connected to a mass spectrometer (timsTOF Pro2) equipped with a nanoliter ion source for data acquisition. Chromatographic separation was performed using a 75 μm ID×15 cm reverse phase column (PePSep C18, 1.9 μm, 75 μm×15 cm, Bruker, Germany). The mobile phase used an acetonitrile-water-formic acid system, in which mobile phase A was 0.1% formic acid aqueous solution and phase B was 0.1% formic acid acetonitrile solution. After the chromatographic column was equilibrated with 100% phase A, the sample was directly loaded onto the chromatographic column by an autosampler and then separated by gradient separation of the chromatographic column, with a flow rate of 300 nL / min and a gradient time of 60 min. The proportion of mobile phase B: 2% for 0 min, 2-22% for 45 min, 22-37% for 5 min, 37-80% for 5 min, and 80% for 5 min.

[0154] The mass spectrometer uses DDA PaSEF mode for DDA data acquisition, with a scan range of 100-1700 m / z. During the PASEFMS / MS scan, the collision energy increases linearly with the ion mobility, from 20 eV (1 / K0 = 0.6 Vs / cm 2 ) rises to 59eV (1 / K0 = 1.6Vs / cm 2 ).

[0155] 3. Library search and protein quantification

[0156] 3.1DDA database search and identification

[0157] The raw data files were searched using SpectroDive (4.1.230421.52329; Biognosys AG 2013) software with the Pulsar search engine, and qualitative analysis was performed after the search.

[0158] Table 2 Database matching parameter settings

[0159]

[0160] 3.2 PRM method construction and quantitative analysis

[0161] The PRM method was constructed using SpectroDive software: 1) The collected DDA data was matched to the target peptide sequence and the library was searched using SpectroDive software to generate the corresponding library. 2) Panel screening peptides and confirmation of quantitative secondary b / y ions were generated based on the library and target peptides. A maximum of 3 specific peptides were selected for each protein, and 3 to 6 fragment ions were selected for quantification for each target peptide.

[0162] 4. Experimental Results

[0163] 4.1 Model construction

[0164] The baseline plasma specimens of 21 patients with advanced NSCLC in the responder group and 9 patients in the nonresponder group who were treated with PD-(L)1 monoclonal antibody combined with chemotherapy were subjected to PRM targeted detection of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 (Table 3), and the model was constructed as follows:

[0165] Score = 4.988-0.011×var1-0.027×var2-0.068×var3+0.039×var4+0.018×var5 Table 3 PRM targeted detection results

[0166]

[0167]

[0168] The peptide levels of these five proteins were jointly analyzed using the above formula to calculate the scores of each sample, see Table 4, and the ROC curve analysis is shown in Figure 2 , AUC was 0.884, the judgment standard was a score greater than 0.3029, 19 of the 21 response groups were positive, and 1 of the 9 non-response groups was positive, with a specificity of 88.9% and a sensitivity of 90.5%. Table 4 Scores of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 combined for the construction of a prediction model for the efficacy of NSCLC immunotherapy

[0169] Group Score Group Score Responder 3.6523 Responder 2.3789 Responder 0.7480 Responder 1.7738 Responder 1.8807 Responder 1.9432 Responder -0.9653 Responder 3.9640 Responder 0.4446 Responder 3.8758 Responder -0.2832 Responder 0.9094 Responder 0.8982 Nonresponder 0.0585 Responder 0.4875 Nonresponder 0.3029 Responder 2.8451 Nonresponder 0.2835 Responder 2.0797 Nonresponder 2.0484 Responder 1.5112 Nonresponder -2.3240 Responder 2.6530 Nonresponder -1.5421 Responder 1.1992 Nonresponder -0.5624 Responder 2.6182 Nonresponder -0.0990 Responder 3.4278 Nonresponder -0.8753

[0170] 4.2 Model Validation

[0171] The baseline plasma samples of 18 responders and 12 nonresponders of advanced NSCLC patients treated with PD-(L)1 monoclonal antibody combined with chemotherapy were subjected to PRM targeted detection of PGK1, CHGA, ELN, HNRNPH3 and SERBP1. The peptide levels of these five proteins were jointly analyzed and calculated using the above formula. The scores of each sample are shown in Table 5, and the ROC curve analysis is shown in Figure 3 , AUC is 0.935. According to the interpretation standard provided by the present invention (score greater than 0.3029), 16 of the 18 response groups were positive, and 2 of the 12 non-response groups were positive, with a specificity of 83.3% and a sensitivity of 88.9%.

[0172] Table 5 Scores of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 combined for validation of the prediction model for the efficacy of immunotherapy in NSCLC

[0173]

[0174]

[0175] Sequence of this article:

[0176] SEQ ID NO:1PGK1

[0177] IQLINNMLDK

[0178] SEQ ID NO:2CHGAELQDLALQGAK

[0179] SEQ ID NO:3ELN AGYPTGTGVGPQAAAAAAAK

[0180] SEQ ID NO:4HNRNPH3

[0181] DGMDNQGGYGSVGR

[0182] SEQ ID NO:5SERBP1

[0183] SAAQAAAQTNSNAAGK.

Claims

1. A separated protein combination, which is a marker for the efficacy of immunotherapy for non-small cell lung cancer, the protein combination comprising PGK1, CHGA, ELN, HNRNPH3 and SERBP1, Preferably, The amino acid sequence of PGK1 is shown in SEQ ID NO: 1, and / or The amino acid sequence of CHGA is shown in SEQ ID NO: 2, and / or The amino acid sequence of ELN is shown in SEQ ID NO: 3, and / or The amino acid sequence of HNRNPH3 is shown in SEQ ID NO: 4, and / or The amino acid sequence of SERBP1 is shown in SEQ ID NO:

5.

2. A reagent for detecting the content of PGK1, CHGA, ELN, HNRNPH3 and SERBP1 in blood, Preferably, the reagent is a reagent used for detecting proteins using any of the following methods: ELISA, suspension chip, MSD electrochemiluminescence technology or mass spectrometry, More preferably, the reagent is a reagent used for quantifying protein using the PRM method, More preferably, the reagent also includes a reagent for separating proteins from blood.

3. A kit comprising the reagent according to claim 2, Preferably, the kit further comprises the protein combination according to claim 1, Preferably, the kit further comprises a reagent for detecting an internal standard.

4. Use of phosphoglycerate kinase 1 (PGK1), chromogranin A (CHGA), elastin (ELN), heterogeneous nuclear ribonucleoprotein H3 (HNRNPH3) and SERPINE1 mRNA binding protein 1 (SERBP1) as biomarkers or their detection reagents in the preparation of reagents or kits for predicting the efficacy of immunotherapy for non-small cell lung cancer, Preferably, the reagent detects the content of the protein or the transcription level of the coding sequence of the protein.

5. The use according to claim 4, characterized in that The prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma content of the marker is: Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

6. A device, characterized in that: The device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the program: (1) Obtain the content of the following proteins in plasma samples: PGK1, CHGA, ELN, HNRNPH3 and SERBP1, (2) obtaining a score by constructing a model using the content, and (3) Predict the efficacy of immunotherapy for non-small cell lung cancer based on the score.

7. The device according to claim 6, characterized in that The prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma content of the marker is: Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

8. A system for predicting the efficacy of immunotherapy for non-small cell lung cancer, characterized in that: include: The collection device is used to obtain the content of the following proteins in the plasma sample: PGK1, CHGA, ELN, HNRNPH3 and SERBP1, A data processing device for obtaining a score by constructing a model using the content, A determination device is used to predict the efficacy of immunotherapy for non-small cell lung cancer based on the score.

9. The system according to claim 8, characterized in that The prediction model for the efficacy of immunotherapy for non-small cell lung cancer based on the plasma content of the marker is: Score = -0.011×PGK1-0.027×CHGA-0.068×ELN+0.039×HNRNPH3+0.018×SERBP1.

10. A method for constructing a model for predicting the efficacy of immunotherapy for non-small cell lung cancer, comprising the steps of: (1) Obtaining the levels of the following proteins in plasma samples: PGK1, CHGA, ELN, HNRNPH3, and SERBP1, and (2) Calculate the ROC curve and area under the curve of the protein combination, and establish a model using logistic regression statistical analysis method.

Images