Biomarker and diagnostic system for lung cancer

By screening and utilizing biomarkers such as Cyfra21-1, CEA, CA125, Pro-SFTPB, combined with data analysis modules, a detection system for lung cancer diagnosis was developed, which solved the lack of sensitivity and specificity of lung cancer diagnosis markers in the prior art. The problem of this paper achieves accurate distinction and early prediction of primary and metastatic lung cancer, and improves the treatment effect.

CN118937675APending Publication Date: 2024-11-12HANGZHOU GUANGKEANDE BIOTECHNOLOGY CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202410272729.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2023-05-15
Filing Date
2023-05-31
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing lung cancer diagnosis methods lack effective markers, which leads to more than 70% of lung cancer patients missing the best treatment opportunity when they are diagnosed. The sensitivity and specificity of commonly used markers such as CEA and CA125 are not ideal, making it difficult to accurately distinguish the primary Sexual and metastatic lung cancer.

Method used

Cyfra21-1, CEA, CA125, Pro-SFTPB and other biomarkers were screened out. Through blood sample detection, combined with data analysis modules and detection systems, reagents and diagnostic models for distinguishing primary and metastatic lung cancer were developed, and the relative Abundance and combination markers were diagnosed.

Benefits of technology

It realizes early risk prediction and metastatic judgment of lung cancer patients, improves the sensitivity and specificity of diagnosis, can more accurately distinguish between primary and metastatic lung cancer, provides early intervention and treatment strategies, and improves the control rate before surgery and long-term survival rates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118937675A_ABST
    Figure CN118937675A_ABST
Patent Text Reader

Abstract

The invention provides biomarkers, namely CA125 and Pro-SFTPB, and the biomarkers and a combination of the biomarkers and other existing markers can be used for judging whether primary lung cancer is transferred to other tissues or organs or not, so that canceration of the other tissues or organs is caused.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application is a divisional application of Chinese application number: 202310631059.X, filing date: May 31, 2023. This application claims priority of the prior Chinese application, application number: 2023105286767, filing date: May 11, 2023;

[0002] 2023105433569, filing date: priority on May 15, 2023, all contents of which are made part of the present invention. Technical Field

[0003] The present invention relates to the field of medicine, and more specifically, to the use of proteomics to screen lung cancer biomarkers and use them for the diagnosis of lung cancer, and more particularly to a new biomarker for distinguishing whether lung cancer is primary lung cancer or metastatic lung cancer and its use, and also to a marker that can distinguish whether primary lung cancer has metastasized and its use. Background Art

[0004] Proteomics is the science of studying the composition, localization, changes and interaction patterns of proteins in cells, tissues or organisms, including the study of protein expression patterns and proteome functional patterns. With the development of mass spectrometry technology, liquid chromatography-mass spectrometry (LC-MS / MS) has become the most important tool in proteomics research. The development of proteomics is of great significance for finding diagnostic markers for diseases, screening drug targets, toxicology research, etc., and is therefore widely used in medical research.

[0005] Lung cancer is one of the most common malignant tumors in clinical practice. It is highly malignant and has a rapid course of disease. Its morbidity and mortality rates rank first among malignant tumors and are increasing year by year.

[0006] Accurate diagnosis of lung cancer is the key to reducing mortality, but there is currently a lack of effective diagnostic methods. More than 70% of lung cancer patients have missed the best time for treatment when they are diagnosed. At present, there are two main methods for diagnosing lung cancer: histology and imaging. However, both methods have certain limitations. With the development of immunology and molecular biology, tumor-related protein markers have shown increasingly important clinical value in the diagnosis and treatment of lung cancer, and have become an indispensable biological indicator for auxiliary diagnosis, observation of efficacy and judgment of prognosis.

[0007] Resection of the primary tumor is an essential treatment for lung cancer to be cured, but surgery itself may also promote postoperative recurrence in the following ways: inducing perioperative micrometastasis, removing anti-angiogenic signals from the tumor, inducing the secretion of tumor growth factors, and inducing postoperative cell-mediated immunosuppression. Therefore, neoadjuvant therapy that reduces the viability of tumor cells in micrometastases and intervenes earlier becomes an attractive treatment strategy, which can improve the complete control rate of tumor patients before surgery and greatly improve the long-term survival and cure rate of patients.

[0008] Clinically, a variety of tumor markers have been discovered that can be used for lung cancer diagnosis, pathological classification and clinical staging, and for judging prognosis and efficacy. However, the diagnostic efficacy of the commonly used lung cancer markers (CEA, CA125) is not ideal, and a specific tumor marker with high sensitivity and specificity for lung cancer diagnosis has not yet been found.

[0009] Therefore, it is of great clinical value to find new lung cancer diagnosis-related markers, combine multiple markers, and adopt appropriate lung cancer diagnosis prediction models. Summary of the invention

[0010] In view of the problems existing in the prior art, the present invention provides a biomarker for lung cancer detection, and screens out a series of new biomarkers that can early predict the risk of lung cancer. In particular, it can distinguish whether lung cancer patients have primary lung cancer and metastatic lung cancer. In addition, it can effectively distinguish primary and metastatic lung cancer, so that effective treatment can be carried out for different pathogenesis. In particular, for primary lung cancer, it can also predict or judge in advance whether the primary lung cancer will metastasize, which has positive significance for surgical resection and prognosis.

[0011] On the one hand, the present invention provides a use of a biomarker in preparing a reagent for distinguishing whether a lung cancer patient has primary lung cancer or metastatic lung cancer, wherein the biomarker is selected from one or more of the following: Cyfra21-1, CEA, CA125, Pro-SFTPB. In fact, it is to find that traditional markers have new uses.

[0012] In some methods, the markers that can be used to distinguish whether lung cancer is primary or metastatic lung cancer can be used to prepare detection reagents with biomarkers as detection targets, such as sample pretreatment reagents, antigens or antibodies, and other biological reagents and kits suitable for the detection of the biomarkers; they can also be developed into standardized reagents or kits suitable for LC-UV or LC-MS detection of the biomarkers.

[0013] Furthermore, the biomarkers include Cyfra21-1, CEA, CA125, and Pro-SFTPB.

[0014] Furthermore, the reagent is used to detect biomarkers in a body fluid sample, and the body fluid sample includes any one of blood, urine, saliva, and sweat.

[0015] In some embodiments, the biomarkers of the present invention are obtained by screening blood samples, and are particularly suitable for developing blood test reagents or kits for lung cancer prediction, especially in lung cancer patients, and can distinguish whether the lung cancer patient or individual is a patient with primary lung cancer or metastatic lung cancer.

[0016] Furthermore, the detecting of the marker in the body fluid sample is to detect the presence or relative abundance or concentration of the biomarker in the body fluid sample of the individual.

[0017] In some embodiments, relative abundance is preferably used to represent the peak area of ​​the biomarker in the detection spectrum obtained by high performance liquid chromatography-tandem mass spectrometry. For example, if the average peak area of ​​a biomarker measured in a control sample (an individual without lung cancer) is 500 and the average peak area measured in a lung cancer sample is 3000, then the abundance of the biomarker in the lung cancer sample is considered to be 6 times that in the control sample.

[0018] On the other hand, the present invention provides a marker combination for preparing a reagent for distinguishing primary from metastatic lung cancer in an individual, wherein the biomarker is selected from a combination of any two or more of the following: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0019] Furthermore, four combinations are included: Cyfra21-1, CEA, CA125, and Pro-SFTPB.

[0020] In another aspect, the present invention provides a marker combination as a reagent for distinguishing whether an individual with lung cancer has primary lung cancer or metastatic lung cancer. The kit includes the biomarkers as described above, or a detection reagent for the biomarker combination as described above.

[0021] In some embodiments, the detection reagent is an antibody against the biomarker as described above, and the antibody is a monoclonal antibody.

[0022] On the other hand, the present invention provides a system for distinguishing whether a lung cancer individual has primary lung cancer or metastatic lung cancer, the system comprising a data analysis module, the data analysis module being used to analyze the detection value of a biomarker in a biological fluid sample, the biomarker being one or more selected from the following: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0023] In some embodiments, the biomarkers consist of the following markers: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0024] Furthermore, it also includes a data detection system and a data input and output interface; the data detection system is used to detect biomarkers in samples and obtain detection values; the input interface in the data input and output interface is used to input the detection value of the biomarker, and after the data analysis module analyzes the detection value, the output interface is used to output the analysis result of whether the lung cancer patient has primary lung cancer or metastatic lung cancer.

[0025] In some embodiments, the data analysis module includes a model equation, and the embodiment is:

[0026]

[0027] Wherein, Y is the predicted value, i represents the i-th biomarker, m represents the number of biomarkers (m=4), Xi represents the detection value of the i-th biomarker (μg / mL), Ki represents the coefficient of the i-th biomarker, and b is a constant of 4.4195.

[0028] Markers coefficient Cyfra21-1 0.098 CEA 0.061 CA125 0.2323 Pro-SFTPB 0.4178 .

[0029] In some embodiments, when the diagnostic model prediction value Y≤0.518, the lung cancer patient to be tested is considered not to be a primary lung cancer patient; when the model prediction value Y>0.518, the lung cancer patient to be tested is considered to be a metastatic lung cancer patient.

[0030] In another aspect, the present invention provides use of the system as described above for constructing a detection model for predicting the probability value of whether a lung cancer patient has primary lung cancer or metastatic lung cancer.

[0031] The "primary lung cancer" here means that cancer or cancer cells first occur in the lungs. Generally, there are no cancer cells in other parts of the human body, and the lungs are the first place where cancer cells are produced, thus forming malignant nodules in the lungs. The so-called "metastatic lung cancer" means that there are no cancer cells or malignant nodules in the lungs, but cancer cells or malignant nodules are produced in other organs of the human body. As the cancer develops, it metastasizes to the lungs, thus forming malignant nodules in the lungs. For example, thyroid cancer metastasizes to the lungs, or breast cancer metastasizes to the lungs, or liver cancer metastasizes to the lungs. These are all malignant nodules in the lungs caused by cancer cells from other tissues colonizing in the lungs and causing cancerous changes in the lung tissue.

[0032] Whether the primary lung cancer metastasizes refers to whether the lung tumor of the lung cancer patient will metastasize to other parts of the body, for example, whether the lung cancer cells will metastasize to liver cancer, pancreatic cancer or other organs and colonize, causing the colonized organs to also become cancerous.

[0033] Another aspect of the present invention provides a marker for distinguishing whether an individual with lung cancer has metastasized and its use, wherein the marker is selected from one or more of the following markers: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0034] In some embodiments, the markers include a combination of the following markers: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0035] Furthermore, the reagent is used to detect biomarkers in a body fluid sample, and the body fluid sample includes any one of blood, urine, saliva, and sweat.

[0036] In some embodiments, the biomarkers of the present invention are obtained by screening blood samples, and are particularly suitable for developing blood test reagents or kits for lung cancer prediction, especially for distinguishing whether the primary lung cancer has metastasized in patients with primary lung cancer.

[0037] Furthermore, the detecting of the marker in the body fluid sample is to detect the presence or relative abundance or concentration of the biomarker in the body fluid sample of the individual.

[0038] On the other hand, the present invention provides a system for diagnosing whether primary lung cancer has metastasized, the system comprising a data analysis module, the data analysis module being used to analyze the detection values ​​of biomarkers in biological fluid samples, the biomarkers being one or more selected from the following: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0039] In some embodiments, the biomarkers consist of the following markers: Cyfra21-1, CEA, CA125, Pro-SFTPB.

[0040] Furthermore, it also includes a data detection system and a data input and output interface; the data detection system is used to detect biomarkers in samples and obtain detection values; the input interface in the data input and output interface is used to input the detection value of the biomarker, and after the data analysis module analyzes the detection value, the output interface is used to output the analysis result of whether the primary lung cancer has metastasized.

[0041] In some embodiments, the calculation module includes an equation of the model, wherein the equation is:

[0042]

[0043] Where Y is the predicted value, i represents the i-th biomarker, m represents the number of biomarkers (m=4), Xi represents the detection value of the i-th biomarker (μg / mL), Ki represents the coefficient of the i-th biomarker as shown in the following table, and b is a constant of 10.7;

[0044] Markers coefficient Cyfra21-1 -0.1973837 CEA -0.6102073 CA125 -0.4825551 Pro-SFTPB -0.5241662 .

[0045] In some aspects, when the diagnostic model prediction value Y≤0.525, it is considered that the primary lung cancer patient to be tested will not metastasize; when the model prediction value Y>0.525, it is considered that the primary lung cancer patient to be tested may metastasize.

[0046] Whether the primary lung cancer of the present invention has metastasized refers to whether the primary lung cancer has metastasized to other organs or tissues and caused other organs or tissues to carry cancer cells, thereby causing cancer.

[0047] The Pro-SFTPB is an amino acid sequence with a UniProt database number of P07988; CA125 is an amino acid sequence with a UniProt database number of Q8WXI7; CEA is an amino acid sequence with a UniProt database number of Q13984; and Cyfra21-1 is an amino acid sequence with a UniProt database number of P08727.

[0048] The beneficial effects of the present invention are:

[0049] Four new biomarkers were screened that can distinguish whether a lung cancer patient has primary lung cancer or metastatic lung cancer; or can distinguish whether primary lung cancer will metastasize. Although these four markers are known, they have been found to have new uses. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The AUC result diagram of the optimal combination model under different hyperparameter combinations of the glmnet algorithm in Example 1;

[0051] Figure 2 is the ROC curve of the lung cancer combined diagnosis model constructed in Example 1 in the model group;

[0052] Figure 3 is the ROC curve of the lung cancer joint diagnosis model constructed in Example 1 in the test group;

[0053] Figure 4 This is a performance evaluation result diagram of the lung cancer joint diagnosis model constructed in Example 1 in the test group;

[0054] Figure 5 : ROC curves of different lung cancer diagnosis models constructed in Example 1.

[0055] Figure 6 The AUC result diagram of the optimal combination model under different hyperparameter combinations of the glmnet algorithm in Example 2;

[0056] Figure 7 is the ROC curve of the lung cancer combined diagnosis model constructed in Example 2 in the model group;

[0057] Figure 8 is the ROC curve of the lung cancer joint diagnosis model constructed in Example 2 in the test group;

[0058] Fig. 9 This is a performance evaluation result diagram of the lung cancer joint diagnosis model constructed in Example 2 in the test group;

[0059] Fig.10 : ROC curves of different lung cancer diagnosis models constructed in Example 2.

[0060] Detailed description

[0061] (1) Diagnosis or testing

[0062] The diagnosis or detection here refers to the detection or assay of biomarkers in a sample, or the content of a target biomarker, such as an absolute content or a relative content, and then the presence or quantity of the target marker is used to indicate whether the individual providing the sample may have or suffer from a certain disease, or the possibility of having a certain disease. The meanings of diagnosis and detection here are interchangeable. The result of such a test or diagnosis cannot be directly used as a direct result of being ill, but is an intermediate result. If a direct result is obtained, other auxiliary means such as pathology or anatomy are required to confirm that a certain disease is present. For example, the present invention provides a variety of new biomarkers associated with lung cancer, and the changes in the content of these markers are directly associated with whether or not a person has lung cancer, or whether a lung cancer patient has primary lung cancer or metastatic lung cancer, or whether or not a primary lung cancer has metastasized.

[0063] (2) Association between markers or biomarkers and lung cancer

[0064] Markers and biomarkers have the same meaning in the present invention. The connection here means that the appearance or change of the content of a certain biomarker in a sample has a direct correlation with a specific disease, for example, a relative increase or decrease in the content indicates whether it is primary lung cancer or metastatic lung cancer in a lung cancer patient, or whether primary lung cancer has metastasized, which provides auxiliary means for treatment and can take early intervention treatment according to different situations.

[0065] For the numerous biomarkers in serum discovered by the present invention, these markers can be used to distinguish whether a lung cancer patient has primary lung cancer or metastatic lung cancer, or whether primary lung cancer has metastasized. The markers here can be used as a single marker for direct detection or diagnosis. The selection of such a marker indicates that the relative change in the content of the marker has a strong correlation with the type of occurrence of lung cancer patients or whether it has metastasized. Of course, it is understandable that one or more markers with strong correlation with lung cancer can be selected for simultaneous detection. It is normal to understand that in some ways, selecting biomarkers with strong correlation for detection or diagnosis can achieve a certain standard of accuracy, such as 60%, 65%, 70%, 80%, 85%, 90% or 95% accuracy, which can indicate that these markers can obtain the intermediate value of diagnosing a certain disease, but it does not mean that a certain disease can be directly confirmed. .

[0066] Of course, the differential protein with the larger ROC value can also be selected as a diagnostic marker. The so-called strong and weak are generally calculated and confirmed by some algorithms, such as the contribution rate of markers to lung cancer or weight analysis. Such calculation methods can be significance analysis (p value or FDR value) and fold change (Fold change). Multivariate statistical analysis mainly includes principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and orthogonal partial least squares discriminant analysis (OPLS-DA), and of course other methods, such as ROC analysis, etc. Of course, other model prediction methods are also possible. When specifically selecting biomarkers, the differential proteins disclosed in the present invention can be selected, or other existing known marker combinations can be selected or combined for prediction through model methods. DETAILED DESCRIPTION

[0067] The present invention is further described in detail below in conjunction with the accompanying drawings and examples. It should be noted that the following examples are intended to facilitate understanding of the present invention and do not have any limiting effect on it. The reagents used in this example are all known products and were obtained by purchasing commercially available products.

[0068] The present invention team further studied the contents disclosed in the patent application, application number: 202211486610.8, and found that in lung cancer patients, the amount of certain biomarkers present showed significant differences in distribution. After in-depth research, testing and analysis, it was found that the concentrations of certain biomasses in primary lung cancer patients and metastatic lung cancer were significantly different. At the same time, it was also found that certain biomarkers were significantly different in samples with or without metastasis of primary lung cancer. We believe that these markers can not only distinguish between healthy people and lung cancer patients, but also may be able to distinguish between primary and metastatic lung cancer in lung cancer patients, or can predict or diagnose whether primary lung cancer has metastasized.

[0069] Therefore, based on the 8 markers screened out on the basis of the patent application, 4 markers were further screened out: Cyfra21-1, CEA, CA125, Pro-SFTPB. The above 4 markers have the potential to distinguish primary lung cancer from metastatic lung cancer, or whether primary lung cancer has metastasized, in lung cancer patients. Their concentrations in blood samples of certain primary lung cancer and metastatic lung cancer are significantly different, and there are also significant differences in blood samples of primary lung cancer that has metastasized and that has not metastasized. In order to verify this judgment, we conduct the following two examples to verify our ideas.

[0070] Example 1: A classification model for identifying whether lung cancer remission is primary lung cancer or metastatic lung cancer by combining four differentially expressed proteins and its establishment

[0071] 1. Get the data

[0072] Study population:

[0073] From 2020 to 2022, 1,250 lung cancer patients were collected, including 650 with metastatic lung cancer and 600 with primary lung cancer. All enrolled patients signed informed consent. The results of all lung cancer patients were confirmed by pathology through biopsy, and primary lung adenocarcinoma and metastatic lung adenocarcinoma were identified by immunohistochemistry. The participants were divided into a model group (primary lung cancer n=585, metastatic lung cancer n=540) and a test group (primary lung cancer n=65, metastatic lung cancer n=60) in a ratio of 9:1. The data information is shown in Table 1:

[0074] Table 1: Modeling sample information

[0075] Model Group Test Group Primary lung cancer 585 65 Metastatic lung cancer 540 60

[0076] The inclusion criteria for lung cancer patients were as follows: (a) primary lung cancer means no history of other malignant tumors, and metastatic lung cancer means a history of other malignant tumors, but no malignant tumors in the lungs, and then developed lung cancer; (b) surgery within one month after blood collection, and postoperative pathological confirmation of primary or metastatic lung cancer. After informed consent, all serum samples collected were stored in a serum bank at -80°C.

[0077] In this example, the collected serum samples were subjected to enzyme-linked immunosorbent assay (ELISA) to obtain the concentrations of Cyfra21-1, CEA, CA125 and Pro-SFTPB in the serum.

[0078] 2. Statistical Analysis of Experimental Data

[0079] The Shapiro Wilk's test was used to assess normal distribution, and the nonparametric test Wilcoxon test was used to analyze the differences in blood marker concentrations between primary lung cancer patients and metastatic lung cancer in the model group and the test group, respectively. In the model group, a combined diagnostic model of 4 lung cancer markers was constructed using a combination of multiple machine learning methods. The area under the receiver operator characteristic (ROC) curve (AUC) was estimated using the predicted probability value with a 95% confidence interval (CI) to evaluate the discriminatory ability of the multivariate diagnostic model. Using the test group, the Youden index (YI) was calculated to determine the cut-off value for distinguishing primary lung cancer patients from metastatic predicted probability. In addition, ROCs for individual markers and different subgroups were constructed and compared. Standard descriptive statistics such as frequency, mean, median, positive predictive value (PPV), negative predictive value (NPV), and standard deviation (SD) were calculated to describe the experimental results of the study population. Statistical analysis was performed using R3.6.1, and p values ​​less than 0.05 were considered statistically significant.

[0080] 3. Steps to construct a joint diagnosis model for distinguishing primary from metastatic lung cancer

[0081] S101, four protein markers including Cyfra21-1, CEA, CA125 and Pro-SFTPB of the samples in the model group were used as the original training data set.

[0082] S102, select the generalized linear model (glmnet) algorithm for building a prediction model, and the grid search range in the process of optimizing the hyperparameters of the algorithm. In this step, the grid search range for the hyperparameter optimization of the model is set for each algorithm as shown in Table 2.

[0083] Table 2: Parameter grid search ranges for the glmnet algorithm

[0084]

[0085] S103, according to the algorithm and hyperparameter setting range set in step S102, one of the hyperparameter combinations is selected as a parameter for constructing a prediction model.

[0086] S104, the original data set is divided into K subsets according to the K-fold cross validation mechanism. To ensure that the ratio of majority class samples to minority class samples in each fold is the same as that of the original data set, a Stratified K-Folds cross validation mechanism is used to perform data segmentation.

[0087] S105, according to the K training data subsets obtained by segmentation in step S104, one of the subsets is selected as a verification set Ddev.

[0088] S106, merging the training data subsets not selected in step S105 to form a training data pool Dtrain1.

[0089] S107, building a prediction model based on the selected supervised classification algorithm and hyperparameters according to the training data set Dtrain obtained in step S106.

[0090] S108, according to the prediction model obtained in step S107, the validation set Ddev is evaluated to obtain the AUC value, and the current prognosis prediction model and the corresponding AUC value are stored in the prediction model pool Pool. Step S108 is to evaluate the prediction model obtained in step S107 on the validation set determined in the current iteration, and store both the model and the evaluation results in the prediction model pool for future base prediction model selection. The evaluation mentioned in this step can be an AUC value or other reasonable indicators for evaluating model performance.

[0091] S109, determine whether each subset has been used as a validation set. Step S109 is to determine whether the K subsets obtained in step S104 have all been used as validation sets and have been trained on the model. If all subsets have been used as validation sets and have completed training, execute step S110; if there are subsets that have not been used as validation sets, execute step S105. This step ensures that every sample in the original data set has been used as a validation set, improves model stability, and prevents the model from overfitting to a certain subset.

[0092] S110, the AUC average value of all models in the prediction model pool Pool is obtained as the final performance evaluation value of the combined model, and the model parameters and the final performance evaluation AUC value are stored in the optimal model pool Poolbest.

[0093] S111, determine whether all hyperparameter combinations have constructed prediction models. Step S111 is to determine whether all algorithms and corresponding hyperparameter combinations obtained in step S102 have constructed prediction models. If all combinations have completed the construction of the model, then execute step S112; if there are combinations that have not completed the construction of the model, then execute step S103.

[0094] S113, selecting the model with the largest AUC value from the model set Poolbest obtained in step S112 as the final prediction model under this marker combination.

[0095] 4.4MP model parameter optimization results

[0096] Through the above model building steps, we obtained the models constructed under 9 different combinations of glmnet algorithm hyperparameters ( Figure 3 ), and the model performance was evaluated by AUC value. As shown in Table 3 and Figure 1 As shown: when the glmnet algorithm hyperparameter combination is alpha = 0.1, lambda = 0.0547, the AUC reaches a maximum value of 0.8602 (the 10-fold cross-validation method is used to calculate the AUC during the modeling process).

[0097] Table 3: AUC of the model constructed under different hyperparameter combinations of the glmnet algorithm

[0098] ALPHA LAMBDA AUC 0.1 0.0547 0.8602 0.55 0.0005 0.8600 0.55 0.0055 0.8600 0.1 0.0055 0.8599 0.1 0.0005 0.8599 1 0.0005 0.8599 1 0.0055 0.8596 0.55 0.0547 0.8576 1 0.0547 0.8521

[0099] The equation for building a model based on the optimal hyperparameter combination is:

[0100]

[0101] Wherein, Y is the predicted value, i represents the i-th biomarker, m represents the number of biomarkers (m=4), Xi represents the detection value of the i-th biomarker (μg / mL), Ki represents the coefficient of the i-th biomarker (Table 4), and b is a constant of 4.4195.

[0102] Table 4: Coefficients of the four biomarkers in the model

[0103]

[0104]

[0105] 5. Determination of diagnostic threshold of lung cancer combined diagnosis model (4MP) (differentiation between primary and metastatic)

[0106] The ROC curve was drawn with the predicted value in the model group, and the optimal diagnostic cutoff value was set to 0.518 according to the Youden index value. That is, when the predicted value of the diagnostic model is ≤0.518, the lung cancer patient to be tested is considered to be a primary lung cancer patient; when the predicted value of the model is >0.518, the lung cancer patient to be tested is considered to be a metastatic lung cancer patient. The results are as follows Figure 2 As shown: The model had an AUC of 0.861, a sensitivity of 78.5%, and a specificity of 78.1% in the model group.

[0107] 6. Validation of the combined lung cancer diagnosis model (4MP) (differentiation between primary and metastatic)

[0108] The ROC curve is plotted with the predicted values ​​in the test group, such as Figure 3 As shown in the figure, the AUC is 0.907. The optimal diagnostic cutoff value is set to 0.518 according to the Youden index value. That is, when the diagnostic model prediction value is ≤0.518, the subject is considered to be a primary lung cancer patient; when the model prediction value is >0.518, the subject is considered to be a metastatic lung cancer patient. The results are shown in the figure. Figure 4 As shown: The accuracy of the model in the test group was 84.8%, the Kappa value was 0.697, the sensitivity was 78.5%, the specificity was 91.7%, the positive predictive value was 91.1%, and the negative predictive value was 79.7%.

[0109] 7. Comparison of diagnostic value of different lung cancer diagnostic models

[0110] Table 5: Comparison of the area under the ROC curve of different diagnostic models

[0111] PANEL AUC 95%Cl DIFF P VALUE 4MP 0.907 0.855~0.959 / / Pro-SFTPB 0.849 0.783~0.915 -0.057948718 0.017394425 CYFRA21-1 0.723 0.635~0.811 -0.184102564 7.19E-06 CA125 0.617 0.513~0.721 -0.290512821 1.04E-07 CEA 0.633 0.535~0.731 -0.274102564 8.33E-07

[0112] like Figure 5 As shown in Table 5, the AUC of our model (4MP) is 0.06, 0.18, 0.29 and 0.27 higher than that of the single markers, respectively. The results of DeLong's test using the AUC difference significance test method showed that the diagnostic value of our model (4MP) was significantly (p<0.05) higher than that of the single marker. Although the combination method can distinguish primary lung cancer from metastatic lung cancer in lung cancer patients, the above-mentioned single markers also have diagnostic value and can be used alone to distinguish between primary and metastatic lung cancer in lung cancer patients. In particular, the AUC value of Pro-SFTPB is 0.849, while the AUC value of CYFRA21-1 is 0.723.

[0113] Example 2: Four differentially expressed proteins are combined to identify whether primary lung cancer has metastasized

[0114] 1. Sample collection

[0115] Medical records of 740 patients with lung cancer who underwent surgical treatment in Zhejiang Cancer Hospital from May 2021 to May 2022 were collected. Among them, 474 patients with complete clinical and follow-up data were included in the model group. Univariate and multivariate analyses and ROC curves of actual metastasis rates were performed, and the area under the curve, specificity, and sensitivity were recorded. The Youden index was used as the cutoff value for judging the presence or absence of metastasis. The data of 266 lung cancer patients served as the test group for the predictive ability of the model. Sample inclusion criteria: (1) postoperative pathological stage I-IIIA; (2) primary lung cancer with a single lesion in the first treatment; (3) complete medical records and follow-up results. Exclusion criteria: (1) Patients with multiple lung lesions or distant metastases before surgery; (2) Patients with a history of other malignant tumors; (3) Patients with non-small cell carcinoma after surgery; (4) Patients with lymph nodes not removed due to various reasons and no accurate pathological staging after surgery; (5) Patients with non-tumor causes of death; (6) Patients with incomplete medical records and follow-up results; (7) Patients with residual cancer cells at the resection margin after surgery. The data information of patients with metastasis and lung metastasis in the model group and the test group are shown in Table 6:

[0116] Table 6: Modeling sample information

[0117] Model Group Test Group Primary lung cancer without metastasis 203 143 Primary lung cancer metastasis 271 123

[0118] In this example, the collected serum samples were subjected to enzyme-linked immunosorbent assay (ELISA) to obtain the concentrations of Cyfra21-1, CEA, CA125 and Pro-SFTPB in the serum.

[0119] 2. Statistical Analysis of Experimental Data

[0120] The Shapiro Wilk's test was used to assess normal distribution, and the nonparametric test Wilcoxon test was used to analyze the differences in blood marker concentrations between patients with lung cancer metastasis and non-metastasis in the model group and the test group, respectively. In the model group, a combined diagnostic model of four lung cancer markers was constructed using a combination of multiple machine learning methods. The predicted probability values ​​were used to estimate the area under the receiver operator characteristic (ROC) curve (AUC) with a 95% confidence interval (CI) to evaluate the discriminatory ability of the multivariate diagnostic model. Using the test group, the Youden index (YI) was calculated to determine the predicted probability cut-off value for distinguishing primary lung cancer patients with metastasis or non-metastasis. In addition, ROCs for individual markers and different subgroups were constructed and compared. Standard descriptive statistics such as frequency, mean, median, positive predictive value (PPV), negative predictive value (NPV), and standard deviation (SD) were calculated to describe the experimental results of the study population. Statistical analysis was performed using R3.6.1, and p values ​​less than 0.05 were considered statistically significant.

[0121] 3. Steps to construct the combined diagnostic model (4MP) for distinguishing primary lung cancer metastasis from non-metastasis

[0122] S101, the concentration matrix of four protein markers, Cyfra21-1, CEA, CA125 and Pro-SFTPB, of the samples in the model group was used as the original training data set.

[0123] S102, select the generalized linear model (glmnet) algorithm for building the prediction model, and the grid search range in the process of optimizing the hyperparameters of the algorithm. In this step, the grid search range for the hyperparameter optimization of the model is set for each algorithm as shown in Table 7.

[0124] Table 7: Parameter grid search ranges for the glmnet algorithm

[0125]

[0126] S103, according to the algorithm and hyperparameter setting range set in step S102, one of the hyperparameter combinations is selected as a parameter for constructing a prediction model.

[0127] S104, the original data set is divided into K subsets according to the K-fold cross validation mechanism. To ensure that the ratio of majority class samples to minority class samples in each fold is the same as that of the original data set, a Stratified K-Folds cross validation mechanism is used to perform data segmentation.

[0128] S105, according to the K training data subsets obtained by segmentation in step S104, one of the subsets is selected as a verification set Ddev.

[0129] S106, merging the training data subsets not selected in step S105 to form a training data pool Dtrain1.

[0130] S107, building a prediction model based on the selected supervised classification algorithm and hyperparameters according to the training data set Dtrain obtained in step S106.

[0131] S108, according to the prediction model obtained in step S107, the validation set Ddev is evaluated to obtain the AUC value, and the current prognosis prediction model and the corresponding AUC value are stored in the prediction model pool Pool. Step S108 is to evaluate the prediction model obtained in step S107 on the validation set determined in the current iteration, and store both the model and the evaluation results in the prediction model pool for future base prediction model selection. The evaluation mentioned in this step can be an AUC value or other reasonable indicators for evaluating model performance.

[0132] S109, determine whether each subset has been used as a validation set. Step S109 is to determine whether the K subsets obtained in step S104 have all been used as validation sets and have been trained on the model. If all subsets have been used as validation sets and have completed training, execute step S110; if there are subsets that have not been used as validation sets, execute step S105. This step ensures that every sample in the original data set has been used as a validation set, improves model stability, and prevents the model from overfitting to a certain subset.

[0133] S110, the AUC average value of all models in the prediction model pool Pool is obtained as the final performance evaluation value of the combined model, and the model parameters and the final performance evaluation AUC value are stored in the optimal model pool Poolbest.

[0134] S111, determine whether all hyperparameter combinations have constructed prediction models. Step S111 is to determine whether all algorithms and corresponding hyperparameter combinations obtained in step S102 have constructed prediction models. If all combinations have completed the construction of the model, then execute step S112; if there are combinations that have not completed the construction of the model, then execute step S103.

[0135] S113, selecting the model with the largest AUC value from the model set Poolbest obtained in step S112 as the final prediction model for lung cancer diagnosis.

[0136] 4. Parameter optimization results of the combined diagnosis model of primary lung cancer metastasis and non-metastasis (4MP)

[0137] Through the above model building steps, we obtained the models constructed under 9 different combinations of glmnet algorithm hyperparameters ( Figure 3 ), and the model performance was evaluated by AUC value. As shown in Table 8 and Figure 6 As shown: when the glmnet algorithm hyperparameter combination is alpha = 0.55, lambda = 0.040473495, the AUC reaches a maximum value of 0.8413 (the 10-fold cross-validation method is used to calculate the AUC during the modeling process).

[0138] Table 8: AUC of the model constructed under different hyperparameter combinations of the glmnet algorithm

[0139] ALPHA LAMBDA AUC 0.1 0.000404735 0.841204784 0.1 0.004047349 0.841190579 0.1 0.040473495 0.841049875 0.55 0.000404735 0.841211794 0.55 0.004047349 0.841268111 0.55 0.040473495 0.841300659 1 0.000404735 0.841267973 1 0.004047349 0.841121087 1 0.040473495 0.840228489

[0140] The equation for building a model based on the optimal hyperparameter combination is:

[0141]

[0142] Wherein, Y is the predicted value, i represents the i-th biomarker, m represents the number of biomarkers (m=4), Xi represents the detection value of the i-th biomarker (μg / mL), Ki represents the coefficient of the i-th biomarker (Table 9), and b is a constant of 10.7.

[0143] Table 9: Coefficients of the four biomarkers in the model

[0144] Markers coefficient Cyfra21-1 -0.1973837 CEA -0.6102073 CA125 -0.4825551 Pro-SFTPB -0.5241662 .

[0145] 5. Determination of diagnostic threshold of lung cancer combined diagnosis model (4MP) (primary metastasis and non-metastasis)

[0146] The ROC curve was drawn with the predicted value in the model group, and the optimal diagnostic cutoff value was set to 0.525 according to the Youden index value. That is, when the predicted value of the diagnostic model is ≤0.525, it is considered that the primary lung cancer patient to be tested will not metastasize; when the predicted value of the model is >0.525, it is considered that the lung cancer patient to be tested may metastasize. The results are as follows Figure 7 As shown: The model had an AUC of 0.841, a sensitivity of 76%, and a specificity of 76.5% in the model group.

[0147] 6. Validation of the lung cancer combined diagnosis model (4MP) (differentiation between primary metastasis and non-metastasis)

[0148] The ROC curve is plotted with the predicted values ​​in the test group, such as Figure 8 As shown in the figure, the AUC is 0.852. The optimal diagnostic cutoff value is set to 0.525 according to the Youden index value. That is, when the diagnostic model prediction value is ≤0.525, it is considered that the patient with lung cancer to be tested will not metastasize; when the model prediction value is >0.525, it is considered that the patient with lung cancer to be tested may metastasize. The results are shown in the figure. Fig. 9 As shown: The accuracy of the model in the test group was 77.4%, the Kappa value was 0.547, the sensitivity was 78.3%, the specificity was 76.4%, the positive predictive value was 79.4%, and the negative predictive value was 75.2%.

[0149] 7. Comparison of diagnostic value of different lung cancer diagnostic models

[0150] Table 10: Comparison of the area under the ROC curve of different diagnostic models

[0151] PANEL AUC 95%Cl DIFF P VALUE 4MP 0.852 0.807-0.897 / / Pro-SFTPB 0.746 0.687-0.805 -0.106543863 6.55E-05 CYFRA21-1 0.59 0.520-0.659 -0.262664165 1.23E-11 CA125 0.68 0.616-0.744 -0.172664734 8.56E-09 CEA 0.719 0.658-0.780 -0.133321963 2.54E-06

[0152] like Fig.10As shown in Table 10, the AUC of our model (4MP) is 0.1065, 0.263, 0.173 and 0.133 higher than that of traditional single markers. The results of DeLong's test using the AUC difference significance test method show that the diagnostic value of our model (4MP) is significantly (p<0.05) higher than that of traditional markers or traditional marker combination models. However, for a single marker, it also has diagnostic value and can be used to diagnose or predict the possibility of metastasis of primary lung cancer.

[0153] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. Use of a biomarker in the preparation of a reagent for diagnosing whether a primary lung cancer patient has metastasized to other tissues or organs, wherein the biomarker is selected from one or more of the following biomarkers: CA125 and Pro-SFTPB.

2. The use according to claim 1, wherein the biomarkers consist of Cyfra21-1, CEA, CA125 and Pro-SFTPB.

3. The use according to claim 2, wherein the reagent is used to detect biomarkers in a body fluid sample, and the body fluid sample comprises any one of blood, urine, saliva, and sweat.

4. The use according to claim 3, wherein detecting the marker in the body fluid sample comprises detecting the presence or relative abundance or concentration of the biomarker in the body fluid sample of the individual.

5. A system for diagnosing or predicting whether lung cancer metastasizes to other tissues or organs, the system comprising a calculation module, the calculation module comprising a model equation, the equation being: in, Y is the predicted value, i represents the i-th biomarker, m represents the number of biomarkers, m=4, Xi represents the detection value of the i-th biomarker (μg / mL), Ki represents the coefficient of the i-th biomarker as shown in the following table, and b is a constant of 10.7; 。 6. According to the system of claim 5, when the diagnostic model prediction value Y≤0.525, it is considered that the primary lung cancer patient to be tested will not metastasize to other tissues or organs; when the model prediction value Y>0.525, it is considered that the lung cancer patient to be tested may metastasize to other tissues or organs.

7. The system according to claim 6, further comprising a data analysis module, wherein the data analysis module is used to analyze the detection value of the biomarker.

8. The system according to claim 6, further comprising a display, wherein the result of the calculation is directly displayed on the display to show whether the primary lung cancer has metastasized to other tissues or organs.

9. The system according to claim 14, wherein the biological sample is a blood sample.

10. Use of a biomarker in preparing a reagent for diagnosing whether a lung cancer patient has primary lung cancer or metastatic lung cancer, wherein the biomarker is selected from one or more of the following: Cyfra21-1, CEA, CA125, Pro-SFTPB.

11. The use according to claim 10, wherein the biomarkers include Cyfra21-1, CEA, CA125, and Pro-SFTPB.

12. The use according to claim 11, wherein the reagent is used to detect biomarkers in a body fluid sample, and the body fluid sample comprises any one of blood, urine, saliva, and sweat.

13. The use according to claim 12, wherein the detecting of the marker in the body fluid sample is detecting the presence or relative abundance or concentration of the biomarker in the body fluid sample of the individual.

14. The use according to claim 12, wherein the liquid sample is a serum sample.

Citation Information

Patent Citations

  • Biomarker for lung cancer detection and system thereof

    CN115575636A