A combined marker, a kit for detecting lung cancer and use thereof
By using a combination of eight protein biomarkers, including CDON, and the LCscore model, the problems of high misdiagnosis rate and radiation risk in existing lung cancer screening methods have been solved, achieving highly sensitive and specific early detection of lung cancer, which is suitable for screening high-risk populations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- THE FIRST AFFILIATED HOSPITAL OF ZHENGZHOU UNIV
- Filing Date
- 2023-09-19
- Publication Date
- 2026-08-04
AI Technical Summary
Existing lung cancer screening methods, such as LDCT and biomarkers, suffer from high misdiagnosis rates, radiation risks, significant differences in diagnostic criteria, and insufficient sensitivity and specificity, necessitating more precise early screening methods.
Eight protein biomarkers—CDON, CLEC4A, DGKZ, FLT3LG, IL6, KRT19, MLN, MVK, PSIP1, REG4, SIT1, and TGFA—were combined with an LCscore calculation model and thresholds to detect lung cancer using plasma serum samples.
It achieves highly sensitive and specific lung cancer detection, reduces the misdiagnosis rate, avoids radiation damage, and has strong consistency in test results, making it suitable for early screening of high-risk groups.
Smart Images
Figure CN117741150B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of lung cancer biomarkers, specifically to a combination biomarker, reagent kit, and its use for detecting lung cancer. Background Technology
[0002] Lung cancer is one of the most common cancers worldwide and a leading cause of death. According to the World Health Organization (WHO), more than 1.5 million people die from lung cancer globally each year. Lung cancer is a malignant tumor that often presents no obvious symptoms in its early stages. This means that in most cases, lung cancer is discovered at an advanced stage, making treatment more difficult and resulting in poorer outcomes and prognoses. Therefore, early lung cancer screening is crucial for early diagnosis and treatment, improving treatment success rates and survival rates while reducing patient suffering and financial burden. Early lung cancer screening primarily targets high-risk groups. Currently, common methods for early lung cancer screening include low-dose CT scans (LDCT) and serum biomarker testing.
[0003] LDCT (Lung-Lower CT) is currently the most widely used lung cancer screening technology. It boasts advantages such as high accuracy, non-invasiveness, simplicity, and speed, providing a high early detection rate for lung cancer. However, LDCT technology has the following drawbacks: 1. High sensitivity but low specificity, resulting in a certain rate of misdiagnosis in practical applications; 2. Requires high-performance CT equipment to obtain high-quality lung images; 3. Although the radiation dose of LDCT is low, it still carries a certain risk of radiation damage; 4. LDCT cannot directly diagnose lung cancer. Doctors need to make further judgments and diagnoses based on the imaging results. The diagnostic standards and procedures vary significantly between different hospitals and doctors, leading to potential discrepancies in diagnostic results.
[0004] Biomarkers used for lung cancer screening include carcinoembryonic antigen (CEA), cytokeratin 19 fragment (CYFRA 21-1), neuron-specific enolase (NSE), and progastrin-releasing peptide (ProGRP), or those with good detection performance for a single lung cancer subtype, often requiring combined use. CEA is a glycoprotein expressed in various types of cancer, including lung cancer, and some studies have reported elevated serum CEA levels in lung cancer patients. CYFRA 21-1 is another biomarker used for lung cancer screening; it is a marker for non-small cell lung cancer (NSCLC) and is released into the bloodstream during tumor growth. CYFRA 21-1 has higher detection sensitivity than CEA, but lower specificity, and its elevation can also be caused by some non-malignant diseases. Therefore, CYFRA 21-1 needs to be used in conjunction with other biomarkers to improve diagnostic accuracy. NSE is a glycolytic enzyme expressed in various types of cancer, including lung cancer, and is widely used for screening patients with small cell lung cancer (SCLC) and assessing treatment response to lung cancer. ProGRP is a tumor marker for small cell lung cancer (SCLC). The sensitivity and specificity of the aforementioned biomarkers vary significantly, and their sensitivity is low. Therefore, the screening and diagnosis of lung cancer still require the combined use of multiple biomarkers to improve accuracy, necessitating further research into more precise early lung cancer screening methods.
[0005] In the context of complex genetics, the application of proteomics technology can improve the early detection rate of lung cancer. Proteins, as essential components of cell structure and function, reflect the state and biological characteristics of cells. Proteomics can identify novel protein biomarkers through methods such as protein typing, protein quantification, and protein functional analysis, thereby improving the accuracy and sensitivity of early lung cancer diagnosis. Summary of the Invention
[0006] To address one of the aforementioned technical deficiencies, this application provides a combination biomarker, a reagent kit, and its uses for detecting lung cancer.
[0007] According to a first aspect of the embodiments of this application, a combined biomarker for detecting lung cancer is provided, comprising the proteins: CDON, CLEC4A, DGKZ, FLT3LG, IL6, KRT19, MLN, MVK, PSIP1, REG4, SIT1, and TGFA.
[0008] According to a second aspect of the embodiments of this application, a kit for detecting lung cancer is provided, comprising a combination biomarker for detecting lung cancer as described above.
[0009] Preferably, the test kit is designed for plasma serum protein samples.
[0010] Preferably, the kit includes an LCscore calculation model and a threshold; the LCscore calculation model is obtained based on the NPX values of each protein in the combined biomarker.
[0011] LCscore = -3.82 - 1.36 * NPX CDON -1.73*NPX CLEC4A -1.18*NPX DGKZ -0.23*NPX FLT3
[0012] LG +0.01*NPX IL6 +0.38*NPX KRT19 -0.89*NPX MLN -0.78*NPX MVK +1.23*NPX PSIP1 -0.80*NPX REG4 -0.36*NPX SIT1 +0.92*NPX TGFA The LCscore value of the tested sample can be calculated using the LCscore calculation model.
[0013] Preferably, the method for determining the threshold includes the following steps:
[0014] Step 1: Create the working characteristic curves for the training set:
[0015] Plasma serum protein was collected from each subject in the training set, which included n lung cancer patients and m non-lung cancer patients;
[0016] The LCscore values for all subjects were calculated based on the LCscore calculation model.
[0017] The operating characteristic curve of the training set was generated based on the LC score of all subjects: the true positive rate (sensitivity) was used as the vertical axis and the false positive rate (1-specificity) was used as the horizontal axis.
[0018] Step 2: Determine the threshold based on the working characteristic curve of the training set:
[0019] The threshold is set by using the value on the working characteristic curve of the training set where 1 - the smaller the specificity and the greater the sensitivity.
[0020] According to a third aspect of the embodiments of this application, a kit for detecting lung cancer is provided for use in the preparation of kits for predicting and diagnosing lung cancer.
[0021] The combined biomarkers for detecting lung cancer provided in this application exhibit high sensitivity and specificity. The training set shows a sensitivity of 83.3%, a 1-specificity of 88.2%, a positive predictive value (PPV) of 87.3%, a negative predictive value (NPV) of 84.5%, an AUC of 0.907, and a P-value of 0.831. The validation set shows a sensitivity of 87.5%, a 1-specificity of 83.33%, a positive predictive value (PPV) of 82.35%, a negative predictive value (NPV) of 88.24%, an AUC of 0.84, and a P-value of 0.676. This demonstrates the high accuracy of the combined biomarkers provided in this application. Attached Figure Description
[0022] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0023] Figure 1 Volcano diagram of proteins with significant differences in the screening process for combined biomarkers for detecting lung cancer provided in the embodiments of this application;
[0024] Figure 2 Intergroup test plots of protein features provided in the embodiments of this application;
[0025] Figure 3 A schematic diagram of the working characteristic curves of the training set provided in the embodiments of this application. Detailed Implementation
[0026] To make the technical solutions and advantages of the embodiments of this application clearer, the exemplary embodiments of this application will be described in further detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not an exhaustive list of all embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.
[0027] Example 1: Screening of combined biomarkers for detecting lung cancer:
[0028] (1) A positive group and a control group were set up; the positive group included 84 lung cancer patients and the control group included 84 non-lung cancer patients.
[0029] (2) Serum samples were extracted and sequenced from the positive group and the control group, and protein detection values were obtained by Olink384Panel detection.
[0030] (3) Screen for proteins that show significant differences between the positive and control groups; the volcano plot of proteins with significant differences is shown below. Figure 1 As shown.
[0031] (4) The model was obtained by performing 100 LASSO regressions on proteins with significant differences. Half or more of the proteins that appeared in the model were obtained and used as protein features. The intergroup test plot of protein features is shown in Figure 2.
[0032] The combined biomarkers for detecting lung cancer selected using the above method in this application are characterized by the following proteins: CDON, CLEC4A, DGKZ, FLT3LG, IL6, KRT19, MLN, MVK, PSIP1, REG4, SIT1, and TGFA.
[0033] (5) Input the acquired protein features into the model, use the predicted probability of lung cancer as the model's output value, perform 10-fold cross-validation on the model, and select the best-performing 1-fold model as the lung cancer prediction model. The resulting lung cancer prediction model is the Lcscore calculation model, where the LCscore calculation model is:
[0034] LCscore = -3.82 - 1.36 * NPX CDON -1.73*NPX CLEC4A -1.18*NPX DGKZ -0.23*NPX FLT3
[0035] LG +0.01*NPX IL6 +0.38*NPX KRT19 -0.89*NPX MLN -0.78*NPX MVK +1.23*NPX PSIP1 -0.80*NPX REG4 -0.36*NPX SIT1 +0.92*NPX TGFA .
[0036] In this application, protein detection was performed on all serum samples in the positive and control groups. Differential protein analysis between cancer patients and non-cancer patients was conducted using sequencing results, and a model was constructed based on this. The resulting combined biomarkers for detecting lung cancer have high sensitivity and specificity.
[0037] Example 2
[0038] This application also provides a kit for detecting lung cancer, including a combination biomarker for detecting lung cancer as described above.
[0039] Furthermore, the test kit is designed for testing plasma serum protein samples.
[0040] The lung cancer detection kit provided in this application requires only a small amount of blood for testing, making collection convenient and effectively avoiding repeated invasive tissue biopsies. It does not require complex collection equipment and poses no radiation risk. Furthermore, it can provide real-time, dynamic feedback on lung cancer screening levels.
[0041] Furthermore, the kit also includes an LCscore calculation model and a threshold. The LCscore calculation model can calculate the LCscore value of the test sample. The LCscore value of the test sample is compared with the threshold. If the LCscore value of the test sample is greater than or equal to the threshold, the patient is a lung cancer patient or a suspected lung cancer patient; otherwise, the patient is a non-lung cancer patient or a suspected non-lung cancer patient.
[0042] The kit provided in this application calculates the LCscore value of the test sample using an LCscore calculation model. It determines whether the test sample has lung cancer by comparing the combined biomarker protein levels with a threshold, avoiding the problem of inconsistent diagnostic results caused by significant differences in diagnostic standards and procedures between different hospitals and doctors in existing technologies. The kit provided in this application can be used to assess the cancer risk of asymptomatic high-risk patients and can be used as a lung cancer screening tool for the general population.
[0043] Furthermore, the method for determining the threshold includes the following steps:
[0044] Step 1: Create the operating characteristic curve (ROC curve of the training set):
[0045] Plasma serum protein was collected from each subject in the training set, which included n lung cancer patients and m non-lung cancer patients; specifically, n and m are natural numbers, n is 66 and m is 68.
[0046] The LCscore values for all subjects were calculated based on the LCscore calculation model.
[0047] The ROC curve for the training set was generated based on the LC score values of all subjects: the ROC curve for the training set uses the true positive rate (sensitivity) as the ordinate and the false positive rate (1-specificity) as the abscissa.
[0048] Step 2: Determine the threshold based on the ROC curve of the training set:
[0049] The threshold is set by using the value on the ROC curve of the training set where 1 represents a smaller specificity and a higher sensitivity; for example... Figure 3As shown, the threshold in this application is 0.57, the sensitivity (Sens) of the training set ROC curve is 83.3%, the 1-specificity (Spec) is 88.2%, the positive predictive value (PPV) is 87.3%, the negative predictive value (NPV) is 84.5%, the AUC is 0.907, and the P is 0.831.
[0050] Step 3: Calculate the sensitivity and 1-specificity of the validation set using thresholds, and plot the ROC curve of the validation set. The validation set includes 18 lung cancer patients and 16 non-lung cancer patients. The characteristics of the ROC curve of the validation set are shown in Table 1 below.
[0051] Table 1. ROC curve characteristics of the validation set
[0052] Sens 0.8750 Spec 0.8333 PPV 0.8235 NPV 0.8824 AUC 0.84(0.685-0.996) P 0.676
[0053] Table 1 demonstrates that the combined markers and LCscore calculation model provided in this application have high accuracy.
[0054] In this application, the threshold is determined using a working characteristic curve, which yields a more accurate threshold and ensures that the test sample is accurately assessed for lung cancer.
[0055] Example 3
[0056] This application also provides the application of a kit for detecting lung cancer in the preparation of kits for predicting and diagnosing lung cancer.
[0057] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0058] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.
[0059] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.
Claims
1. The application of a reagent for detecting a combination of biomarkers for lung cancer in the preparation of a kit for predicting and diagnosing lung cancer, characterized in that, The combined biomarkers include the following proteins: CDON, CLEC4A, DGKZ, FLT3LG, IL6, KRT19, MLN, MVK, PSIP1, REG4, SIT1, and TGFA.
2. The application of the reagent for detecting combined biomarkers of lung cancer according to claim 1 in the preparation of kits for predicting and diagnosing lung cancer, characterized in that, The kit includes an LCscore calculation model and a threshold; The LCscore calculation model is derived from the NPX values of each protein in the combined biomarker: LCscore=-3.82-1.36*NPX CDON -1.73*NPX CLEC4A -1.18*NPX DGKZ -0.23*NPX FLT3LG +0.01*NPX IL6 +0.38*NPX KRT19 -0.89*NPX MLN -0.78*NPX MVK +1.23*NPX PSIP1 -0.80*NPX REG4 -0.36*NPX SIT1 +0.92*NPX TGFA ; The LCscore value of the tested sample can be calculated using the LCscore calculation model.