A combination of body fluid markers and its application in differentiating breast tumors

Abstract

The present invention discloses a combination of body fluid markers and its application in distinguishing breast tumors. The present invention uses body fluid as a sample, and uses liquid chromatography tandem mass spectrometry to obtain protein expression and blood routine data of patients with benign breast tumors, patients with malignant breast tumors, and healthy people, and establishes a proteomics-based prediction model for distinguishing whether the sample corresponds to a healthy person, a patient with benign breast tumors, and a patient with malignant breast tumors. The prediction method of the present invention has the characteristics of non-invasive and convenient detection, and can accurately distinguish whether the sample corresponds to a healthy person, a patient with benign breast tumors, and a patient with malignant breast tumors.

Description

Technical Field

[0001] The present invention belongs to the field of bioinformatics, and specifically relates to a body fluid marker combination, reagents and kits for detecting the same, a prediction model and system for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, and the application of the system in distinguishing breast tumors. Background Art

[0002] Breast tumors are common diseases in general surgery, mostly occurring in young and middle-aged women, and can have a serious impact on the patient's health. Clinically, breast tumors can be divided into two categories: benign breast tumors and malignant breast tumors (i.e., breast cancer). The treatment plans and prognosis of the two are completely different. If diagnostic confusion leads to treatment failure, serious adverse consequences may occur. Therefore, the differential diagnosis of benign and malignant breast tumors is very important.

[0003] At present, there are five main technologies for diagnosing benign and malignant breast tumors in my country: (1) physical examination; (2) imaging diagnosis (breast mammography, breast ultrasound and breast magnetic resonance imaging); (3) tissue biopsy; (4) breast cancer tumor marker examination (serum cancer antigen 15-3 (CA15-3), serum carcinoembryonic antigen (CEA), serum cancer antigen 125 (CA125), etc.); (5) immunohistochemistry (Ki-67, HER-2, ER, PR, etc.). However, there are the following shortcomings in the detection of benign and malignant breast tumors: (1) clinical physical examination is expensive; (2) imaging diagnosis has a high positive miss rate; (3) the more mature biological targets can only be used for risk assessment and are not suitable for clinical diagnosis; (4) puncture biopsy is first of all invasive, and in addition, it has high requirements for the sampler and puncture site, and it is easy to cause different results due to the influence of external factors. Therefore, there is an urgent need to establish new breast benign and breast malignant tumor examination technologies in this field that can sensitively and stably identify and diagnose benign and breast malignant tumors so as to provide timely and targeted treatment, improve patients' quality of life and survival rate, and reduce medical costs. Summary of the invention

[0004] In view of the defect of the prior art that there is a lack of convenient, sensitive and accurate technology for detecting benign breast tumors and malignant breast tumors, based on high-sensitivity, high-resolution and high-precision modern mass spectrometry technology, combined with bioinformatics analysis and machine learning, the present invention provides a body fluid marker combination, a reagent for detecting the body fluid marker combination, a method for constructing a prediction model for benign breast tumors and malignant breast tumors, a prediction model for benign breast tumors and malignant breast tumors, a system for distinguishing whether the corresponding individual of the sample is a healthy person, a patient with benign breast tumors and a patient with malignant breast tumors, and an application of the body fluid marker combination in preparing a diagnostic product for predicting whether the corresponding individual of the sample is a healthy individual, a patient with benign breast tumors and a patient with malignant breast tumors. The detection method of the present invention has the characteristics of being non-invasive and convenient, and can accurately determine whether the sample to be tested is a benign or malignant breast tumor sample.

[0005] In order to solve the above technical problems, the first aspect of the present invention provides a body fluid marker combination, wherein the body fluid marker combination includes any one or more of the following body fluid markers:

[0006] IGFALS, PON3, ATXN10, COPS4, APOA2, AKAP9, P4HB and neutrophil ratio.

[0007] In some embodiments, the body fluid markers in the body fluid marker combination are derived from serum, plasma, saliva, urine or cerebrospinal fluid of patients with benign breast tumors, serum, plasma, saliva, urine or cerebrospinal fluid of patients with malignant breast tumors, and serum, plasma, saliva, urine or cerebrospinal fluid of healthy people.

[0008] In some preferred embodiments, the body fluid markers in the body fluid marker combination are derived from the serum or plasma of patients with benign breast tumors, patients with malignant breast tumors and healthy people.

[0009] The healthy subjects are individuals who do not suffer from benign breast tumors and malignant breast tumors.

[0010] The second aspect of the present invention provides a reagent for detecting the combination of body fluid markers described in the first aspect of the present invention, wherein the reagent is a mass spectrometry detection reagent, an antibody detection reagent, or a reagent for detection by a routine blood analyzer for identifying the body fluid markers, and the antibody detection reagent is preferably a monoclonal antibody; the reagent for detection by a routine blood analyzer includes a hemolytic agent and a diluent.

[0011] In some embodiments, the benign breast tumor is a fibroadenoma or a papilloma, and the malignant breast tumor, i.e., breast cancer, is breast carcinoma in situ, invasive breast carcinoma, metastatic breast lymph node carcinoma, or metastatic breast carcinoma to distant organs according to the World Health Organization Classification of Breast Tumors; the malignant breast tumor is breast cancer, and the malignant breast tumor is classified according to molecular classification as Luminal A type (ER and / or PR+, HER2-, Ki-67<20%), Luminal B type (HER2 negative type (B1 type): ER+ and / or PR<20%, HER2-, Ki-67≥20%; B2: HER2 positive type (B2 type): ER+ and / or PR+, HER2 overexpression); HER2 positive type (HER2+, ER- and PR-) or basal-like type / triple negative type (HER2-, ER- and PR-); the hemolytic agent is a surfactant, a buffer and an antibacterial agent; the diluent contains sodium chloride, sodium sulfate, a buffer and an antibacterial agent.

[0012] In some embodiments, the buffer is ADA, TRIS, HEPES or imidazole; and the antibacterial agent is a penicillin-streptomycin dual antibody solution.

[0013] The third aspect of the present invention provides a kit, which comprises the reagent as described in the second aspect of the present invention.

[0014] A fourth aspect of the present invention provides a method for constructing a prediction model for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor. The method comprises:

[0015] The protein expression data or routine blood test data are input into a RandomForest model for machine learning to construct a prediction model for distinguishing the individuals corresponding to the samples as healthy people, patients with benign breast tumors, and patients with malignant breast tumors; the sources of the protein expression data include serum, plasma, saliva, urine, or cerebrospinal fluid of patients with benign breast tumors, serum, plasma, saliva, urine, or cerebrospinal fluid of patients with malignant breast cancer, and serum, plasma, saliva, urine, or cerebrospinal fluid of healthy people; the protein expression data or routine blood test data include protein expression data or routine blood test data of the body fluid marker combination as described in the first aspect of the present invention.

[0016] In some embodiments, the protein expression data is obtained by LC-MS technology and collected using DIA detection method; and / or, the data collected by the DIA detection method is peptide matched by Firmiana software; the routine blood data is obtained by analyzing the subjects through clinical blood sampling using a routine blood analysis kit and / or a routine blood analyzer.

[0017] In some embodiments, the peptide matching database is the UniProt human protein database; and / or, DIA-NN is used to analyze the protein expression data after Firmiana processing: the iBAQ method is used for protein quantification, the FOT of each protein is calculated, and the FOT of each protein is input into the RandomForest model as the protein expression data.

[0018] In some embodiments, the RandomForest model includes a model training process, and the detailed code of the model training process is:

[0019]

[0020] In some preferred embodiments, the RandomForest model further includes a data partitioning process, and the detailed code of the data partitioning process is:

[0021] 1. Data partition: data = pd.read_csv('D: / data / data analysis / OC_xgboost_classifier1.csv', index_col = 0)

[0022] data_y = data['label']

[0023] data_x=data.drop(['label'],axis=1)

[0024] x_train,x_test,y_train,y_test=

[0025] train_test_split(data_x, data_y, test_size=0.3, random_state=1).

[0026] In some embodiments, before the protein expression data or blood routine data are input into the RandomForest model, the healthy human samples, benign breast tumor samples and malignant breast tumor samples are first divided into a training set and a validation set;

[0027] And / or, the protein expression data input into the RandomForest model meet the following conditions: the expression in benign breast tumor samples / the expression in malignant breast tumor samples>1 and the expression in benign breast tumor samples / the expression in healthy human samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the expression in malignant breast tumor samples / the expression in benign breast tumor samples>1 and the expression in malignant breast tumor samples / the expression in healthy human samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the expression in healthy human samples / the expression in malignant breast tumor samples>1 and the expression in healthy human samples / the expression in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test is less than 0.05;

[0028] And / or, the routine blood test data input into the RandomForest model meet the following conditions: the routine blood test level in benign breast tumor samples / the routine blood test level in malignant breast tumor samples>1 and the routine blood test level in benign breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.05.

[0029] In some embodiments, the sample is human serum, plasma, saliva, urine, or cerebrospinal fluid.

[0030] In some embodiments, the sample is human serum or plasma.

[0031] In some embodiments, according to the construction method described in the fifth aspect of the present invention, the ratio of the training set to the validation set is 7:3.

[0032] The sixth aspect of the present invention provides a prediction model for distinguishing corresponding individuals as healthy people, patients with benign breast tumors and patients with malignant breast tumors. The prediction model is constructed by the construction method described in the fourth aspect of the present invention.

[0033] A seventh aspect of the present invention provides a system for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, the system comprising:

[0034] (1) a data receiving module, used to receive or input protein expression data and blood routine data from body fluid samples of healthy people, patients with benign breast tumors, and patients with malignant breast tumors, wherein the protein expression data and blood routine data include the protein expression data and blood routine data of the body fluid marker combination as described in the first aspect of the present invention;

[0035] In some embodiments, the body fluid sample is human serum, plasma, saliva, urine or cerebrospinal fluid;

[0036] In some embodiments, the body fluid sample is human serum or plasma.

[0037] (2) a judgment and output module, which is used to judge the input data after the receiving or inputting is completed by using the prediction model as described in the fifth aspect of the present invention, and obtain the result that the individual corresponding to the sample is a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor;

[0038] The criteria for judging the results are:

[0039] When the input data meets "Judgment Condition 1", the output judgment result is "Yes, good", and the output prediction result is "suffering from benign breast tumor";

[0040] When the input data meets "Judgment Condition 2", the output judgment result is "Yes, Malignant", and the output prediction result is "Suffering from breast malignant tumor";

[0041] When the input data meets "Judgment Condition 3", the output judgment result is "No", and the output prediction result is "No benign breast tumor and malignant breast tumor", that is, the individual corresponding to the sample is a healthy person;

[0042] The system determines that the predicted probability that the individual corresponding to the sample is a healthy person is p1, the predicted probability that the individual corresponding to the sample suffers from a benign breast tumor is p2, and the predicted probability that the individual corresponding to the sample suffers from a malignant breast tumor is p3;

[0043] The judgment condition 1 is p2>p1 and p2>p3; the judgment condition 2 is p3>p1 and p3>p2; the judgment condition 3 is p1>p2 and p1>p3;

[0044] The protein expression data or blood routine data include the protein expression data or blood routine data of the body fluid marker combination as described in the first aspect of the present invention;

[0045] In a preferred embodiment of the present invention, the system further comprises a sample extraction module, which extracts sample data and transmits it to the marker combination verification module.

[0046] In some embodiments of the present invention, the system further comprises a data processing module, and the data processing module is used to collect protein expression data or blood routine data in body fluid samples.

[0047] In some embodiments of the present invention, the data are collected by LS-MA technology, and the protein expression data are collected by DIA detection.

[0048] In some embodiments of the present invention, the protein expression data collected by the DIA detection method are matched with peptides using Firmiana software.

[0049] In some embodiments of the present invention, the peptide matching database is the UniProt human protein database; and / or, DIA-NN is used to analyze the protein expression data processed by Firmiana: the iBAQ method is used for protein quantification, the FOT of each protein is calculated, and the FOT of each protein is input into the data receiving module as the protein expression data.

[0050] An eighth aspect of the present invention provides a computer-aided method for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, the method comprising the following steps:

[0051] Step 1: receiving or inputting protein expression data or blood routine data in a body fluid sample, wherein the protein expression data includes protein expression data or blood routine data of the body fluid marker combination as described in the first aspect of the present invention;

[0052] Step 2: Input the protein expression data or blood routine data received or input in step 1 into the prediction model as described in the fifth aspect of the present invention, and output a judgment result on whether the individual corresponding to the body fluid sample is a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor.

[0053] The ninth aspect of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can implement the functions of the system as described in the seventh aspect of the present invention, or implement the steps of a computer-assisted prediction method for distinguishing whether a sample corresponds to an individual as a healthy person, a patient with benign breast tumor, and a patient with malignant breast tumor as described in the eighth aspect of the present invention.

[0054] In the present invention, the readable storage medium may more specifically include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0055] In a possible implementation, the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the functions of the system as described in the seventh aspect of the present invention, or to implement the steps of a computer-assisted prediction method for distinguishing whether a sample corresponds to an individual as a healthy person, a patient with benign breast tumor, and a patient with malignant breast tumor as described in the eighth aspect of the present invention.

[0056] The program code for executing the present invention may be written in any combination of one or more programming languages, and may be executed entirely on a user device, partially on a user device, as an independent software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0057] The tenth aspect of the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the functions of the system as described in the seventh aspect of the present invention, or to implement the steps of a computer-assisted prediction method for distinguishing whether a sample corresponds to an individual as a healthy person, a patient with benign breast tumor, and a patient with malignant breast tumor as described in the eighth aspect of the present invention.

[0058] On the basis of being in accordance with the common sense in the art, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present invention.

[0059] The reagents and raw materials used in the present invention are commercially available.

[0060] The positive and progressive effects of the present invention are:

[0061] The present invention uses body fluid as a sample, and utilizes liquid chromatography tandem mass spectrometry to obtain protein phenotypes and blood routine data of patients with benign breast tumors, patients with malignant breast tumors, and healthy controls, and establishes a proteomics-based prediction model for distinguishing whether the sample corresponds to a healthy person, a patient with benign breast tumors, or a patient with malignant breast tumors. The detection method of the present invention has the characteristics of non-invasive and convenient detection, and can accurately distinguish whether the sample corresponds to a healthy person, a patient with benign breast tumors, or an individual with malignant breast tumors. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 Flow chart of the method for detecting benign breast tumors and malignant breast tumors.

[0063] Figure 2 Cumulative curve diagram of serum proteome identification.

[0064] Figure 3 Confusion matrix of healthy people, patients with benign breast tumors and patients with malignant breast tumors when the training set is used as internal validation. Healthy people: Healthy donors, HD; patients with benign breast tumors: Benign mammary disease, BD; patients with malignant breast tumors: Breast cancer, BC.

[0065] Figure 4 Schematic diagram of the area under the random forest ROC curve when the training set is used as internal validation.

[0066] Figure 5 Confusion matrix of healthy people, patients with benign breast tumors and patients with malignant breast tumors when the validation set is used as external validation. Healthy donors: HD; Benign mammary disease: BD; Breast cancer: BC.

[0067] Figure 6 Schematic diagram of the area under the random forest ROC curve when the validation set is used as external validation.

[0068] Figure 7 Confusion matrix of healthy donors, benign breast tumor patients, and breast malignant tumor patients when validating the prediction model for an independent validation set. Healthy donors: HD; Benign mammary disease patients: BD; Breast cancer patients: BC.

[0069] Figure 8 Illustration of the area under the random forest ROC curve when validating the prediction model for an independent validation set.

[0070] Fig. 9 The structure diagram of a system for distinguishing whether a sample corresponds to a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor.

[0071] Fig.10 A schematic diagram of the structure of an electronic device. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0073] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0074] The embodiment uses body fluid as sample and utilizes liquid chromatography tandem mass spectrometry to obtain protein phenotype and blood routine data of patients with benign breast tumors and malignant breast tumors, and establishes prediction models for benign breast tumors and malignant breast tumors.

[0075] Figure 1 The flowchart of the method for detecting benign breast tumors and malignant breast tumors in the embodiment comprises the following steps:

[0076] Example 1 Selection of samples

[0077] Based on body fluid samples from healthy people, patients with benign breast tumors and patients with malignant breast tumors, training sets, validation sets and independent validation sets were established for proteomic analysis and routine blood analysis.

[0078] Serum samples from 322 patients with benign and malignant breast tumors and healthy people were selected. Among these samples, 56 were from healthy people, 112 were from patients with benign breast tumors (91 from patients with breast fibroadenoma and 21 from patients with papilloma), and 154 were from patients with malignant breast tumors (41 patients with breast carcinoma in situ, 68 patients with invasive breast cancer, and 45 patients with breast lymph node metastatic cancer or breast distant organ metastatic cancer).

[0079] Specifically, 70% of the samples were randomly selected from different types of samples to form the training set, and the remaining 30% of the samples were used as the validation set. The training set consisted of 225 samples, including 39 healthy samples, 108 benign breast tumor samples (88 patients with breast fibroadenomas and 20 patients with papilloma), and 78 malignant breast tumor samples (21 patients with breast carcinoma in situ, 34 patients with breast invasive carcinoma, and 23 patients with breast lymph node metastasis / metastasis to distant organs of the breast); the validation set consisted of 97 samples, including 17 healthy samples, 4 benign breast tumor samples (3 patients with breast fibroadenomas and 1 patient with papilloma), and 76 malignant breast tumor samples (20 patients with breast carcinoma in situ, 34 patients with breast invasive carcinoma, and 22 patients with breast lymph node metastasis / metastasis to distant organs of the breast). The independent validation set consisted of 113 samples, including 57 healthy subjects, 29 benign breast tumor samples (21 patients with breast fibroadenoma and 8 patients with papilloma), and 27 malignant breast tumor samples (7 patients with breast carcinoma in situ, 14 patients with breast invasive carcinoma, and 6 patients with breast lymph node metastasis / metastasis to distant organs of the breast). The samples in the training set, validation set, and independent validation set did not overlap.

[0080] The design and implementation of this study were approved and supervised by the Medical Ethics Committee through an ethics vote. Written informed consent was obtained from all patients.

[0081] Example 2 Sample preparation and mass spectrometry analysis

[0082] Body fluid extraction, body fluid protein extraction, mass spectrometry analysis, peptide identification and protein quantification, and protein identification are completed through retrieval and comparison.

[0083] Blood samples were drawn from each patient and serum was separated. First, serum protein was extracted. (1) 2 μL serum was mixed with 98 mL 50 mM Ammonium Bicarbonate (NH4HCO3) to form a 100 μL system and heated at 95°C for 3 min (Eppendorf); (2) After the sample returned to room temperature, it was digested with trypsin (Promega) at an enzyme mass ratio of 1:50 at 37°C overnight; (3) Peptides were extracted and dried (SpeedVac, Eppendorf).

[0084] Secondly, the extracted serum proteins were subjected to mass spectrometry analysis. An EASY-nLC 1200 ultrahigh pressure system liquid chromatography system (Thermo Fisher Scientific) was used. Peptides were separated in 10 min with a flow rate of 600 nL / min on a 150 μm I.D. × 8 cm column with a laser-pulled electrospray emitter packed with 1.9 μm ReproSil-Pur 120C18-AQ particles. Mobile phases A and B were water and acetonitrile (0.1 vol% FA), respectively. Mobile phase B was linearly increased from 15% to 30% in 10 min. Samples were analyzed on a Q-EX-HF mass spectrometer (MS) using a nano-electrospray ion source (Thermo Fisher Scientific). For the generation of ion mobility enhanced libraries, the mass spectrometer was run in data-independent mode. In general, 75% of the samples were injected. The peptide was dissolved in 12 μL (0.1% formic acid) buffer and 9 μL was loaded onto a 100 μm ID × 2.5 cm C18 trapping column with a maximum pressure of 280 bar and 14 μL of solvent A (0.1% formic acid). The DIA method consisted of an MS1 ​​scan from 300-1400 m / z at 60k resolution (AGC target 4e5 or 50ms). 30 DIA fragments were then acquired at 15k resolution with an AGC target of 5e4 and a maximum injection time of 22ms. The setting "Inject all available parallel time" was enabled. HCD fragmentation was set to a normalized collision energy of 30%. Spectra were recorded in profile mode. The default charge state for MS2 was set to 3.

[0085] Finally, peptide identification and protein quantification. All data were processed using the Firmiana one-stop proteomics cloud platform independently developed by the laboratory (officially launched in 2017). FragPipe (v.12.1) and MSFragger (2.2) were used to search the UniProt human protein database (updated on 2019.12.17, 20406 entries). The mass tolerance of the precursor was 20ppm, and the mass tolerance of the product ion was 50mmu. A maximum of two gaps were allowed to be missing. The search engine used cysteine ​​carbamidomethylation as a fixed modification and methionine N-acetylation and oxidation as variable modifications. The precursor ion score charge was limited to +2, +3, and +4. The results of the DIA data were combined into a spectral library using SpectraST software. All libraries were used as reference spectral libraries. DIA data were analyzed using DIANN (v.1.7.0). DIA-NN was performed with default settings: Precursor FDR: 5%, Loglev: 1, Mass accuracy: 20ppm, MS1 accuracy: 10ppm, Scan window: 30, Implicit proteingroup: genes, Quantification strategy: robust LC (high accuracy). The quantification of identified peptides was calculated as the average of the peak areas of chromatographic fragment ions in all reference libraries. Label-free protein quantification was calculated using the label-free, intensity-based absolute quantification (iBAQ) method. Peak area values ​​of the corresponding proteins were calculated. The fraction of total protein (FOT) was used to represent the normalized abundance of a specific protein in a sample. FOT is defined as the iBAQ of a protein divided by the total iBAQ of all identified proteins in the sample. For ease of representation, FOT values ​​were multiplied by 10e5, and missing values ​​were calculated as 10e-5.

[0086] (2) The acquisition of blood routine indicators includes blood sample collection, blood cell count, blood cell morphology examination, and biochemical index examination. After the blood sample of each patient is drawn and the serum is separated, the blood routine analyzer is used for automated analysis.

[0087] The collected blood samples are automatically calculated by the blood routine analyzer, including red blood cell count, white blood cell count, platelet count, hematocrit rate, mean corpuscular volume, mean corpuscular hemoglobin content and other indicators. The blood routine analyzer detection reagents include a hemolytic agent and a diluent, the hemolytic agent is a surfactant, a buffer and an antibacterial agent; the diluent contains sodium chloride, sodium sulfate, a buffer and an antibacterial agent.

[0088] Blood cell morphology examination involves making thin slices of the collected blood specimens and observing and counting the blood cell morphology, size, color and other blood cell morphological characteristics in the blood sample under a microscope.

[0089] The biochemical indicators collected in the blood samples are tested, and the testing range includes hemoglobin concentration, average hemoglobin content of red blood cells, white blood cell classification count, and average platelet volume.

[0090] Example 3 Screening of protein markers

[0091] Analyze the expression of protein markers and select effective proteins for identification.

[0092] The number of proteins identified in each of the 322 samples ranged from 1,700 to 2,300. Figure 2 The cumulative curve shown in Figure 1 shows that a total of 8,955 proteins were identified in 322 samples. The present invention uses the ratio of the expression level of each protein in a specific sample to the expression level of all proteins in the sample (fraction of total, FOT) as its normalized expression level. The FOT value is multiplied by 10e5 and then logarithmized to base 10.

[0093] Proteins with significant differences were selected from the training set (FOT difference fold greater than 1, that is, the expression level in benign breast tumor samples / the expression level in malignant breast tumor samples>1 and the expression level in benign breast tumor samples / the expression level in healthy human samples>1, and the p value of the Kruskal-Willis test was less than 0.05; or, the expression level in malignant breast tumor samples / the expression level in benign breast tumor samples>1 and the expression level in malignant breast tumor samples / the expression level in healthy human samples>1, and the p value of the Kruskal-Willis test was less than 0.05; or, the expression level in healthy human samples / the expression level in malignant breast tumor samples>1 and the expression level in healthy human samples / the expression level in benign breast tumor samples>1, and the p value of the Kruskal-Willis test was less than 0.05.

[0094] A total of 47 proteins were used as candidate proteins: F11R, CNOT1, ARHGAP25, EPS15L1, BLOC1S4, BCAN, RPS10, EML4, ARSA, MMP10, NDUFS3, AKAP12, RMDN3, AGPS, UBA3, BCAM, MAP4K4, GPRC5C, ASAP2, RAVER1, STK4, AOC2, RPL17, PARS2, P4HB, SMC2, RAB32, CELA2A, EIF4B, MMS19, RCN3, PROCR, PRKACB, APOA2, SP140L, FDPS, CDC123, PVR, PPIG, DCTN4, CDH6, PDK1, COPS4, AKAP9, PON3, IGFALS and ATXN10 (Table 1 shows the differential expression of the 47 candidate proteins).

[0095] Table 1 The difference fold of FOT detection values ​​of each protein marker

[0096]

[0097]

[0098] Example 4 Establishing a prediction model

[0099] The protein markers and / or routine blood indicators of the training set were divided into a healthy control group, a benign breast tumor group, and a malignant breast tumor, i.e., breast cancer group, and the prediction model was trained using a machine learning algorithm based on the RandomForest model.

[0100] First, the training set was divided into a healthy control group, a benign breast tumor group, and a malignant breast tumor (i.e., breast cancer) group. The expression level in benign breast tumor samples / the expression level in malignant breast tumor samples>1 and the expression level in benign breast tumor samples / the expression level in healthy samples>1 were selected, and the p-value of the Kruskal-Willis test was less than 0.05; or, the expression level in malignant breast tumor samples / the expression level in benign breast tumor samples>1 and the expression level in malignant breast tumor samples / the expression level in healthy samples>1, and the p-value of the Kruskal-Willis test was less than 0.05; or, the expression level in healthy samples / the expression level in malignant breast tumor samples>1 and the expression level in healthy samples / the expression level in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test was less than 0.05;

[0101] And / or, the routine blood test data input into the RandomForest model meet the following conditions: the routine blood test level in benign breast tumor samples / the routine blood test level in malignant breast tumor samples>1 and the routine blood test level in benign breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; the routine blood test indicators and proteins are used as candidate markers. A total of 71 proteins and / or blood routine indicators were screened as candidate markers (F11R, CNOT1, ARHGAP25, EPS15L1, BLOC1S4, BCAN, RPS10, EML4, ARSA, MMP10, NDUFS3, AKAP12, RMDN3, AGPS, UBA3, BCAM, MAP4K4, GPRC5C, ASAP2, RAVER1, STK4, AOC2, RPL17, PARS2, P4HB, SMC2, RAB32, CELA2A, EIF4B, MMS19, RCN3, PROCR, PRKACB, APOA2, SP140L, FDPS, CDC1 23, PVR, PPIG, DCTN4, CDH6, PDK1, COPS4, AKAP9, PON3, IGFALS, ATXN10, WBC count, RBC count, Hb concentration, Hct, mean corpuscular volume, mean corpuscular Hb content, mean corpuscular Hb concentration, platelet count, neutrophil ratio, lymphocyte ratio, monocyte ratio, eosinophil ratio, basophil ratio, neutrophil count, lymphocyte count, monocyte count, eosinophil count, basophil count, RBC distribution width CV, RBC distribution width SD, platelet distribution width, mean platelet volume, large platelet ratio and platelet hematocrit).

[0102] The prediction model was established based on machine learning method. Eight important features (IGFALS, PON3, ATXN10, COPS4, APOA2, AKAP9, P4HB and neutrophil ratio) were screened out using RandomForest algorithm and the model was established.

[0103] Specifically, based on the RandomForest machine learning algorithm, the numerical values ​​of the 71 screened candidate markers and blood routine indicators were input into the sklearn integrated script shown below to establish a prediction model to predict the prediction probability of benign and malignant breast cancer patients and healthy people.

[0104] Use the RandomForest Classifier function to train the model; use the random_forest.fit function to verify the model in the validation set; the detailed script code of the generalized linear regression model is as follows:

[0105] 1. Data partition: data = pd.read_csv('D: / data / data analysis / OC_xgboost_classifier1.csv', index_col = 0)

[0106] data_y = data['label']

[0107] data_x=data.drop(['label'],axis=1)

[0108] x_train,x_test,y_train,y_test=train_test_split(data_x,data_y,test_size=0.3,random_state=1)

[0109] 2. Model training: sklearn.ensemble.RandomForestClassifier(n_estimators=100,criterion='gini',max_depth=None,min_samples_split=2,min_samples_leaf=1,min_weight_fraction_leaf=0.0,max_features='auto',max_leaf_no des=None, min_impurity_decrease=0.0, min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=None, random_state=None, verbose=0, warm_start=False, class_weight=None, ccp_alpha=0.0, max_samples=None)

[0110] 3. Model prediction: random_forest.fit(x_train,y_train)

[0111] y_pred=random_forest.predict(x_train)

[0112] 4. Model evaluation: y_test_bin = label_binarize(y_test, classes = np.unique(y_test))

[0113] n_classes=y_test_bin.shape[1]

[0114] classes = np.unique(y_test)

[0115] fpr = dict()

[0116] tpr=dict()

[0117] roc_auc = dict()

[0118] for i in range(n_classes):

[0119] fpr[i],tpr[i],_=roc_curve(y_test_bin[:,i],y_score[:,i])

[0120] roc_auc[i]=auc(fpr[i],tpr[i])

[0121] fpr["micro"],tpr["micro"],_=roc_curve(y_test_bin.ravel(),y_score.ravel())

[0122] roc_auc["micro"]=auc(fpr["micro"],tpr["micro"])

[0123] Table 2 shows eight important features in the prediction model.

[0124] Table 2 Eight important features in the prediction model

[0125] name Kruskal-Willis test p-value IGFALS 1.09E-07 PON3 6.87E-08 ATXN10 9.40E-04 COPS4 5.88E-22 APOA2 5.39E-10 AKAP9 2.31E-09 P4H 1.11E-02 Neutrophil ratio 1.34E-05

[0126] During the model validation process, the training set samples are first used as internal validation. Figure 3 is the confusion matrix for internal validation, Figure 4 Figure 2 is a schematic diagram of the area under the ROC curve for internal validation. Figure 3 As shown in Table 3, the internal validation model includes 225 samples, and the accuracy of the diagnostic model is 99.6%, the precision is 99.6%, the recall is 99.1%, and the F1 value is 99.3%. Figure 4 As shown, the ordinate represents the true positive rate of the prediction model; the abscissa represents the false positive rate of the prediction model, and the area under the ROC-micro curve is 0.99.

[0127] Table 3 Internal validation model indicators

[0128] Model indicators value Accuracy 99.6％ Accuracy 99.6％ Recall 99.1％ F1-score 99.3％

[0129] External validation uses validation set samples. Figure 5 is the confusion matrix for external validation, Figure 6 Figure 2 is a schematic diagram of the area under the ROC curve for external validation. Figure 5 As shown in Table 4, the model includes 97 samples, and the accuracy of the diagnostic model is 72.2%, the precision is 77.2%, the recall is 71.6%, and the F1 value is 74.3%. Figure 6 As shown in Figure 2, the area under the ROC-micro curve is 0.84. This result suggests that the protein prediction model has high sensitivity and specificity and has potential application value in the clinical prediction of benign and malignant breast tumors.

[0130] Table 4 External validation model indicators

[0131] Model indicators value Accuracy 72.2％ Accuracy 77.2％ Recall 71.6％ F1-score 74.3％

[0132] Then use the independent validation set samples for verification. Figure 7 is the confusion matrix verified by the independent validation set, Figure 8 The figure is a schematic diagram of the area under the ROC curve verified by the independent validation set. Figure 7 As shown in Table 5, the model includes 113 samples, and the accuracy of the diagnostic model is 45.1%, the precision is 38.1%, the recall is 35.2%, and the F1 value is 36.6%. Figure 8 As shown in Figure 2, the area under the ROC-micro curve is 0.59. This result suggests that the protein prediction model has high sensitivity and specificity and has potential application value in the clinical prediction of benign and malignant breast tumors.

[0133] Table 5 Independent verification model indicators

[0134] Model indicators value Accuracy 45.1％ Accuracy 38.1％ Recall 35.2％ F1-score 36.6％

[0135] Example 5 Prediction system for distinguishing whether a sample corresponds to a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor

[0136] The prediction system 61 for distinguishing whether the sample corresponds to a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor includes a data processing module 52 and a judgment and output module 53, and also includes a data receiving module 51 ( Fig. 9 ).

[0137] The data receiving module 51 is used to receive or input protein expression data and blood routine data from body fluid samples of healthy people, patients with benign breast tumors and patients with malignant breast tumors, and transmit them to the data processing module; the protein expression data and blood routine data include the protein expression data and blood routine data of the body fluid marker combination described in Table 2.

[0138] The data processing module 52 is used to input the received or input protein expression data and blood routine data of the body fluid marker combination into the judgment and output module 53. The protein expression data and blood routine data of the body fluid marker combination can be collected by the data collection module 51, and the protein expression data and blood routine data of the body fluid marker combination can also be obtained from other sources.

[0139] The judgment and output module 53 is used to judge whether the calculation result meets the preset judgment condition through the prediction model as described in Example 4 after the reception or input is completed. The judgment condition is: when the input data meets "judgment condition 1", the output judgment result is "yes, good", and the output prediction result is "suffering from benign breast tumor"; when the input data meets "judgment condition 2", the output judgment result is "yes, bad", and the output prediction result is "suffering from malignant breast tumor"; when the input data meets "judgment condition 3", the output judgment result is "no", and the output prediction result is "not suffering from benign breast tumor and malignant breast tumor", that is, the individual corresponding to the sample is a healthy person.

[0140] The system determines that the predicted probability that the individual corresponding to the sample is a healthy person is p1, the predicted probability that the individual corresponding to the sample suffers from a benign breast tumor is p2, and the predicted probability that the individual corresponding to the sample suffers from a malignant breast tumor is p3; the judgment condition 1 is p2>p1 and p2>p3; the judgment condition 2 is p3>p1 and p3>p2; the judgment condition 3 is p1>p2 and p1>p3.

[0141] Embodiment 6 Electronic device

[0142] The present embodiment provides an electronic device, which can be expressed in the form of a computing device (for example, a server device), including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the purpose of using the prediction model described in the present invention to distinguish between healthy people, patients with benign breast tumors, and patients with malignant breast tumors in Example 4 of the present invention can be achieved.

[0143] Fig.10 The hardware structure diagram of this embodiment is shown, and the electronic device specifically includes:

[0144] At least one processor 91, at least one memory 92, and a bus 93 for connecting different system components (including the processor 91 and the memory 92), wherein:

[0145] The bus 93 includes a data bus, an address bus, and a control bus.

[0146] The memory 92 includes a volatile memory, such as a random access memory (RAM) 921 and / or a cache memory 922 , and may further include a read-only memory (ROM) 923 .

[0147] The memory 92 also includes a program having a set (at least one) of program modules 924, and / or a program tool 925, such program modules 924 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.

[0148] The processor 91 executes various functional applications and data processing by running the computer programs stored in the memory 92, such as the data analysis method of embodiment 3 of the present invention.

[0149] The electronic device can further communicate with one or more external devices 94 (e.g., keyboards, pointing devices, etc.). Such communication can be performed through an input / output (I / O) interface 95. In addition, the electronic device can also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter 96. The network adapter 96 communicates with other modules of the electronic device through a bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.

[0150] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.

[0151] Embodiment 7 Computer readable storage medium

[0152] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps of the method for distinguishing between healthy people, patients with benign breast tumors, and patients with malignant breast tumors in Embodiment 4 of the present invention are implemented.

[0153] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0154] In a possible implementation, the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of the method for distinguishing between healthy people, patients with benign breast tumors and patients with malignant breast tumors in Example 4 of the present invention.

[0155] The program code for executing the present invention may be written in any combination of one or more programming languages, and may be executed entirely on a user device, partially on a user device, as an independent software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0156] Finally, although the embodiments of the present invention have been shown and described above, the above embodiments are exemplary and are only used to illustrate the technical solutions of the present invention. They cannot be understood as limiting the present invention. A person of ordinary skill in the art can change, modify, replace and modify the above embodiments within the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A combination of body fluid markers for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, characterized in that: The body fluid marker combination consists of IGFALS, PON3, ATXN10, COPS4, APOA2, AKAP9, P4HB and neutrophil ratio.

2. The body fluid marker combination according to claim 1, characterized in that: The body fluid markers in the body fluid marker combination are derived from the serum, plasma, saliva, urine or cerebrospinal fluid of patients with benign breast tumors, the serum, plasma, saliva, urine or cerebrospinal fluid of patients with malignant breast tumors, and the serum, plasma, saliva, urine or cerebrospinal fluid of healthy people.

3. The body fluid marker combination according to claim 2, characterized in that: The body fluid markers in the body fluid marker combination are derived from the serum or plasma of patients with benign breast tumors, patients with malignant breast tumors and healthy people.

4. A reagent for detecting the body fluid marker combination according to any one of claims 1 to 3, characterized in that: The reagent is a mass spectrometry detection reagent, an antibody detection reagent or a reagent used for routine blood analyzer detection for identifying the body fluid marker combination; the reagent used for routine blood analyzer detection includes a hemolytic agent and a diluent.

5. The reagent according to claim 4, characterized in that The antibody detection reagent is a monoclonal antibody.

6. The reagent according to claim 5, characterized in that The benign breast tumors include fibroadenomas and papillomas, and the malignant breast tumors include breast carcinoma in situ, breast invasive carcinoma, breast lymph node metastatic carcinoma and breast distant organ metastatic carcinoma; the hemolytic agent is a surfactant, a buffer and an antibacterial agent; the diluent contains sodium chloride, sodium sulfate, a buffer and an antibacterial agent.

7. The reagent according to claim 6, characterized in that The buffer is ADA, TRIS, HEPES or imidazole; the antibacterial agent is penicillin-streptomycin dual antibiotic solution.

8. A kit, characterized in that: The kit comprises the reagent according to any one of claims 4 to 7.

9. A method for constructing a prediction model for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, characterized in that: The construction method comprises: The protein expression data or routine blood data are input into a RandomForest model for machine learning to construct a prediction model for distinguishing the individuals corresponding to the samples as healthy people, patients with benign breast tumors, and patients with malignant breast tumors; the sources of the protein expression data include serum, plasma, saliva, urine, or cerebrospinal fluid of patients with benign breast tumors, and serum, plasma, saliva, urine, or cerebrospinal fluid of patients with malignant breast tumors, and serum, plasma, saliva, urine, or cerebrospinal fluid of healthy people; the protein expression data or routine blood data are the protein expression data or routine blood data of the body fluid marker combination as described in claim 1.

10. The construction method according to claim 9, characterized in that: The protein expression data are obtained by LC-MS technology and collected using DIA detection method; and / or, the data collected by the DIA detection method are peptide matched by Firmiana software; the routine blood data are obtained by analyzing the subjects with a routine blood analysis kit and / or a routine blood analyzer after clinical blood collection.

11. The construction method according to claim 10, characterized in that: The peptide matching database is the UniProt human protein database; and / or, DIA-NN is used to analyze the protein expression data after Firmiana processing: the iBAQ method is used for protein quantification, the FOT of each protein is calculated, and the FOT of each protein is input into the RandomForest model as the protein expression data.

12. The construction method according to claim 9, characterized in that: Before the protein expression data or routine blood test data are input into the RandomForest model, the samples of healthy subjects, samples of benign breast tumors, and samples of malignant breast tumors are first divided into a training set and a validation set; And / or, the protein expression data input into the RandomForest model meet the following conditions: the expression in benign breast tumor samples / the expression in malignant breast tumor samples>1 and the expression in benign breast tumor samples / the expression in healthy human samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the expression in malignant breast tumor samples / the expression in benign breast tumor samples>1 and the expression in malignant breast tumor samples / the expression in healthy human samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the expression in healthy human samples / the expression in malignant breast tumor samples>1 and the expression in healthy human samples / the expression in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; And / or, the routine blood test data input into the RandomForest model meet the following conditions: the routine blood test level in benign breast tumor samples / the routine blood test level in malignant breast tumor samples>1 and the routine blood test level in benign breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1, and the p-value of the Kruskal-Willis test is less than 0.05; or, the routine blood test level in malignant breast tumor samples / the routine blood test level in benign breast tumor samples>1 and the routine blood test level in malignant breast tumor samples / the routine blood test level in healthy person samples>1, and the p-value of the Kruskal-Willis test is less than 0.

05.

13. The construction method according to claim 12, characterized in that: The sample is human serum, plasma, saliva, urine or cerebrospinal fluid.

14. The construction method according to claim 13, characterized in that: The sample is human serum or plasma.

15. The construction method according to claim 14, characterized in that: The ratio of the number of samples in the training set and the validation set is 7:

3.

16. A prediction model for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, characterized in that: The prediction model is constructed by the construction method described in any one of claims 9 to 15.

17. A prediction system for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, characterized in that: The system comprises: (1) a data receiving module, used to receive or input protein expression data and blood routine data from body fluid samples of healthy people, patients with benign breast tumors, and patients with malignant breast tumors, wherein the protein expression data and blood routine data include the protein expression data and blood routine data of the body fluid marker combination according to claim 1; (2) a judgment and output module, which is used to judge the received or input data after the receiving or inputting is completed by using the prediction model according to claim 16, and obtain the result that the individual corresponding to the sample is a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor; The criteria for judging the results are: When the input data meets "judgment condition 1", the output judgment result is "yes, good", and the output prediction result is "suffering from benign breast tumor"; When the input data meets "Judgment Condition 2", the output judgment result is "Yes, Malignant", and the output prediction result is "Suffering from breast malignant tumor"; When the input data meets "Judgment Condition 3", the output judgment result is "No", and the output prediction result is "No benign breast tumor and malignant breast tumor", that is, the individual corresponding to the sample is a healthy person; The system determines that the predicted probability that the individual corresponding to the sample is a healthy person is p1, the predicted probability that the individual corresponding to the sample suffers from a benign breast tumor is p2, and the predicted probability that the individual corresponding to the sample suffers from a malignant breast tumor is p3; The judgment condition 1 is p2>p1 and p2>p3; the judgment condition 2 is p3>p1 and p3>p2; the judgment condition 3 is p1>p2 and p1>p3; The protein expression data or blood routine data include the protein expression data or blood routine data of the body fluid marker combination according to claim 1.

18. The system of claim 17, wherein: The body fluid sample is human serum, plasma, saliva, urine or cerebrospinal fluid.

19. The system of claim 18, wherein: The body fluid sample is human serum or plasma.

20. The system according to any one of claims 17 to 19, characterized in that The system further comprises a sample extraction module, which extracts sample data and transmits it to the marker combination verification module.

21. The system of claim 20, wherein: The system also includes a data processing module, which is used to collect protein expression data or blood routine data in body fluid samples.

22. The system of claim 21, wherein: The data were obtained by LS-MA technology, and protein expression data were collected using DIA detection.

23. The system of claim 22, wherein: The protein expression data collected by the DIA detection method were matched with peptides using Firmiana software.

24. The system of claim 23, wherein: The peptide matching database is the UniProt human protein database; and / or, DIA-NN is used to analyze the protein expression data processed by Firmiana: the iBAQ method is used for protein quantification, the FOT of each protein is calculated, and the FOT of each protein is input into the data receiving module as the protein expression data.

25. A computer-aided method for distinguishing whether a sample corresponds to a healthy person, a patient with a benign breast tumor, or a patient with a malignant breast tumor, characterized in that: The method comprises the following steps: Step 1: receiving or inputting protein expression data or blood routine data in a body fluid sample, wherein the protein expression data includes the protein expression data or blood routine data of the body fluid marker combination according to claim 1; Step 2: Input the protein expression data or blood routine data received or input in step 1 into the prediction model as described in claim 16, and output a judgment result on whether the individual corresponding to the body fluid sample is a healthy person, a patient with benign breast tumor, or a patient with malignant breast tumor.

26. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, can implement the functions of the system as claimed in claim 17, or implement the steps of a computer-assisted prediction method for distinguishing whether samples correspond to healthy people, patients with benign breast tumors, and patients with malignant breast tumors as claimed in claim 25.

27. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the functions of the system according to claim 17, or to implement the steps of the method according to claim 25.

28. Use of the body fluid marker combination as described in any one of claims 1-3, the reagent as described in any one of claims 4-7, the kit as described in claim 8, the system as described in any one of claims 17-24, the computer-readable storage medium as described in claim 26 and / or the electronic device as described in claim 27 in preparing a product for distinguishing whether the individual corresponding to the sample is a healthy person, a patient with benign breast tumor and a patient with malignant breast tumor.

Images