Tuberculosis risk assessment method and system based on multi-dimensional biomarkers

Abstract

The invention relates to the technical field of medical prediction models, and discloses a tuberculosis risk assessment method and system based on a multi-dimensional biomarker, and the method comprises the following steps: 1, collecting a peripheral blood sample of a patient, and carrying out the detection; step 2, carrying out data processing and integration on the biomarkers; 3, constructing a risk prediction model based on logistic regression, and calculating the probability that the patient suffers from tuberculosis; 4, carrying out risk layering according to a calculation result, and providing a personalized risk assessment report; according to the method, the model prediction efficiency is excellent, and the AUC reaches 0.891 (95% CI: 0.834-0.947); the model calibration degree is good, and a calibration curve is close to an ideal straight line; under the optimal cutoff value, the sensitivity is 85.9% (95% CI: 0.785 to 0.933), and the specificity is 79.6% (95% CI: 0.683 to 0.909); according to the method, biomarkers in multiple dimensions are integrated, the prediction accuracy is improved, a visual risk prediction tool is provided, and clinical application is facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of medical prediction models, and specifically to a tuberculosis risk assessment method and system based on multi-dimensional biomarkers. Background Technique

[0002] Tuberculosis is a chronic infectious disease that seriously threatens human health. Currently, the diagnosis of tuberculosis mainly relies on the following methods, but all have obvious limitations:

[0003] 1. Bacteriological examination: Although it has high specificity, it takes a long time and has a low positive rate, especially it is difficult to diagnose in patients with negative sputum bacteria;

[0004] 2. Imaging examination: It lacks specificity and is difficult to effectively distinguish from other lung diseases;

[0005] 3. IGRA detection: Although it has important value in the diagnosis of tuberculosis infection, its positive rate in the diagnosis of active tuberculosis is relatively low, especially lower in patients with negative sputum bacteria.

[0006] The existing technology has the following deficiencies:

[0007] 1. The prediction efficiency of a single biomarker is limited;

[0008] 2. There is a lack of a prediction model that integrates multi-dimensional indicators. Existing models often ignore the combination of new immune markers (such as CD274) and traditional clinical indicators;

[0009] 3. Existing prediction models are mostly limited to single-type indicators and fail to make full use of multi-dimensional information such as immunology, biochemistry, and clinical manifestations. Therefore, a tuberculosis risk assessment method and system based on multi-dimensional biomarkers are proposed. Summary of the Invention

[0010] The purpose of the present invention is to provide a tuberculosis risk assessment method and system based on multi-dimensional biomarkers to solve the problems raised in the above background technique.

[0011] To achieve the above purpose, the present invention provides the following technical solution: A tuberculosis risk assessment method based on multi-dimensional biomarkers, including the following steps:

[0012] Step 1, collect peripheral blood samples of patients and perform tests;

[0013] Step 2, perform data processing and integration on the biomarkers;

[0014] Step 3, construct a risk prediction model based on logistic regression and calculate the probability of a patient developing tuberculosis;

[0015] Step 4: Perform risk stratification based on the calculation results and provide a personalized risk assessment report.

[0016] Preferably, in step 2 above, the biomarkers include the CD274 expression level, the IGRA test result, lymphocyte count (Ly), lactate dehydrogenase (LDH), carcinoembryonic antigen (CEA), and dyspnea symptoms.

[0017] Preferably, for the detection of CD274 expression above, flow cytometry is used and the result is determined as positive or negative. For the IGRA test, a standardized kit is used and the result is determined as positive or negative.

[0018] Preferably, in step 3 above, the prediction model is constructed by logistic regression, and the risk prediction calculation formula is: logit(P) = β0 + 1.733×CD274 + 1.592×IGRA - 0.931×Ly - 1.623×LDH - 0.480×CEA - 1.794×dyspnea, where: P is the probability of developing tuberculosis, β0 is the intercept term, the values of CD274, IGRA, and dyspnea are binary classifications, and Ly, LDH, and CEA are continuous variables.

[0019] A tuberculosis risk assessment system based on multi-dimensional biomarkers, comprising a sample collection and detection unit, a data processing unit, a risk assessment unit, and a result output unit;

[0020] The sample collection and detection unit is used for collecting and detecting the biomarkers of patients;

[0021] The data processing unit is used for integrating multi-dimensional data and calculating the risk probability;

[0022] The risk assessment unit is used for generating the risk probability and performing individualized nomogram analysis;

[0023] The result output unit is used for visualizing the risk score and providing clinical decision support.

[0024] Preferably, the sample collection and detection unit is connected to the data processing unit, the data processing unit is connected to the risk assessment unit, and the risk assessment unit is connected to the result output unit.

[0025] Preferably, the sample collection and detection unit above includes a peripheral blood sample collection module, a tuberculosis-specific antigen stimulation module, a flow cytometry detection module, and a conventional laboratory detection module;

[0026] The peripheral blood sample collection module is used for collecting the peripheral venous blood sample of patients;

[0027] The tuberculosis-specific antigen stimulation module is used to perform a tuberculosis-specific antigen stimulation experiment on the collected blood sample;

[0028] The flow cytometry detection module is used to detect the expression level of CD274 on monocytes;

[0029] The conventional laboratory testing module is used to detect the conventional laboratory indicators of the patient.

[0030] Preferably, the data processing unit includes a multi-dimensional data integration module, a calculation module based on logistic regression, and a risk probability calculation module;

[0031] The multi-dimensional data integration module is used to integrate all the data obtained by the sample collection and detection unit;

[0032] The calculation module based on logistic regression is used to construct a risk prediction model based on logistic regression and construct the calculation formula of the prediction model;

[0033] The risk probability calculation module is used to calculate the probability of the patient developing tuberculosis according to the formula of the prediction model.

[0034] Preferably, the risk assessment unit includes an individualized nomogram analysis module and a risk stratification module;

[0035] The individualized nomogram analysis module is used to construct a nomogram of the prediction model for convenient clinical application;

[0036] The risk stratification module is used to evaluate the risk level of the patient according to the risk probability calculated by the prediction model, classify the patient into different risk levels, and provide a basis for clinical decision-making.

[0037] Preferably, the result output unit includes a risk score visualization module, a clinical decision support module, and a prediction report generation module;

[0038] The risk score visualization module is used to present the calculated risk probability in a graphical manner;

[0039] The clinical decision support module is used to classify the patient into a low-risk or high-risk group according to the calculated risk probability in combination with the optimal cut-off value;

[0040] The prediction report generation module is used to automatically generate a tuberculosis risk assessment report, the content of which includes various detection indicators, risk probability, risk stratification, and clinical suggestions of the patient.

[0041] Compared with the prior art, the present invention adopting the above technical solutions has the following technical effects: The prediction efficacy of the model of the present invention is excellent, with the AUC reaching 0.891 (95% CI: 0.834 - 0.947); the calibration of the model is good, and the calibration curve is close to the ideal straight line; at the optimal cut-off value, the sensitivity is 85.9% (95% CI: 0.785 - 0.933), and the specificity is 79.6% (95% CI: 0.683 - 0.909); the present invention integrates biomarkers in multiple dimensions, improves the prediction accuracy, and provides a visual risk prediction tool for convenient clinical application. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0043] Figure 1 It is the nomogram of the prediction model of the present invention;

[0044] Figure 2 It is the ROC curve graph of the model of the present invention;

[0045] Figure 3 It is the calibration curve graph of the model of the present invention;

[0046] Figure 4 It is the clinical decision curve of the model of the present invention;

[0047] Figure 5 It is the system schematic diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0049] It should be noted that the structures, ratios, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those who are familiar with this technology to understand and read, and are not used to limit the limiting conditions that the present application can be implemented. Therefore, they do not have technical essential significance. Any modification of the structure, change of the proportional relationship, or adjustment of the size, without affecting the effects that the present application can produce and the purposes that can be achieved, should still fall within the scope that the technical content disclosed in the present application can cover.

[0050] Example

[0051] Please refer to Figures 1 - 5 , the present invention provides a technical solution: a tuberculosis risk assessment method based on multi-dimensional biomarkers, comprising the following steps:

[0052] Step 1, collect peripheral blood samples of patients and conduct tests;

[0053] Step 2, perform data processing and integration on the biomarkers. The biomarkers include CD274 expression level, IGRA test result, lymphocyte count (Ly), lactate dehydrogenase (LDH), carcinoembryonic antigen (CEA), and dyspnea symptoms. The CD274 expression test is performed by flow cytometry, and the result is determined as positive or negative. The IGRA test is performed using a standardized kit, and the result is determined as positive or negative;

[0054] Step 3, construct a risk prediction model based on logistic regression to calculate the probability of a patient developing tuberculosis. The prediction model is constructed by logistic regression, and the risk prediction calculation formula: logit(P) = β0 + 1.733×CD274 + 1.592×IGRA - 0.931×Ly - 1.623×LDH - 0.480×CEA - 1.794×dyspnea, where: P is the probability of developing tuberculosis, β0 is the intercept term (16.324), the values of CD274, IGRA, and dyspnea are binary classifications, and Ly, LDH, and CEA are continuous variables; for CD274 and IGRA, if positive, the value is 1, if negative, the value is 0; for dyspnea, if "yes", the value is 1, if "no", the value is 0;

[0055] Ly, LDH, and CEA are all continuous variables, and the values of each variable in the regression equation are as follows:

[0056] Ly: Perform square root transformation on the specific value obtained from the test. For example, if the measured result is 25%, take the square root of 25, and the result is 5.

[0057] LDH: Perform log2 transformation on the specific value obtained from the test. For example, if the measured result is 100, take log2 of 100, and the result is approximately 6.64.

[0058] CEA: Perform box-cox transformation on the specific value obtained from the test, with λ = -0.2. For example, if the measured result is 2.62, perform box-cox transformation on 2.62, and the result is approximately 0.88;

[0059] Step 4, perform risk stratification based on the calculation results and provide a personalized risk assessment report.

[0060] A tuberculosis risk assessment system based on multi-dimensional biomarkers, comprising a sample collection and detection unit, a data processing unit, a risk assessment unit, and a result output unit;

[0061] The sample collection and detection unit is used for collecting and detecting the biomarkers of patients;

[0062] The data processing unit is used for integrating multi-dimensional data and calculating the risk probability;

[0063] The risk assessment unit is used for generating the risk probability and performing individual nomogram analysis;

[0064] The result output unit is used for visualizing the risk score and providing clinical decision support.

[0065] The sample collection and detection unit is connected to the data processing unit, the data processing unit is connected to the risk assessment unit, and the risk assessment unit is connected to the result output unit.

[0066] The sample collection and detection unit includes a peripheral blood sample collection module, a tuberculosis-specific antigen stimulation module, a flow cytometry detection module, and a conventional laboratory detection module;

[0067] The peripheral blood sample collection module is used for collecting the peripheral venous blood samples of patients;

[0068] The tuberculosis-specific antigen stimulation module is used for performing tuberculosis-specific antigen stimulation experiments on the collected blood samples;

[0069] The flow cytometry detection module is used for detecting the expression level of CD274 on monocytes;

[0070] The conventional laboratory detection module is used for detecting the conventional laboratory indexes of patients.

[0071] Selection of collection subjects: 134 research subjects were selected, including 85 tuberculosis patients diagnosed by laboratory tests and 49 patients with other lung diseases (including pneumonia, lung cancer, chronic obstructive pulmonary disease, etc.). All research subjects were clinically diagnosed and had complete clinical data and test results. The basic characteristics of the research subjects are shown in Table 1. The results showed that there were no significant differences in demographic characteristics such as age and gender between the two groups (P>0.05). Among the clinical indexes, the positive rates of CD274 (69.4% vs 26.5%, P<0.001) and IGRA (75.3% vs 28.6%, P<0.001) were significantly higher in the tuberculosis group than in the control group, suggesting that these indexes may have important diagnostic value.

[0072] Table 1. Basic characteristics table of research subjects:

[0073]

[0074]

[0075]

[0076] Sample collection and processing method: Peripheral blood sample collection and pretreatment;

[0077] Collect 3 mL of peripheral venous blood from the subject, collect it using a heparin sodium anticoagulant tube, and divide the blood sample evenly into 3 culture tubes, 1 mL per tube;

[0078] Add respectively: 10 μL of 1×PBS (negative control, medium group), 10 μL of Mtb-specific antigen peptide (antigen stimulation, TB group), and 10 μL of PHA (positive control, P group);

[0079] Cultivate at 37 °C for 16 - 20 hours, centrifuge at 3000 rpm for 10 minutes, and collect cells and supernatant respectively for detection.

[0080] Index detection method: 1. CD274 expression detection: Detect the expression of CD274 on CD14+ monocytes by flow cytometry. Result determination: Positive: The percentage of CD274+ cells in the sample after stimulation in CD14+ monocytes ≥ 3.75% and after stimulation / before stimulation ≥ 2 times; Negative: The percentage of CD274+ cells in the sample after stimulation in CD14+ monocytes < 3.75% or after stimulation / before stimulation < 2 times; 2. IGRA detection: Use a standardized detection kit. Result determination: Positive: IFN-γ(T) - IFN-γ(N) ≥ 14 pg / mL; Negative: IFN-γ(T) - IFN-γ(N) < 14 pg / mL; 3. Lymphocyte count / ratio detection: Main method: Flow cytometry. Specific steps: Sample collection: EDTA anticoagulated whole blood; Detection method: Automatic hematology analyzer method; Reference value: Absolute lymphocyte count: (0.8 - 4.0) × 10^9 / L, lymphocyte ratio: 20 - 40%; 4. Carcinoembryonic antigen (CEA) detection: Main method: Chemiluminescence immunoassay. Detection process: Sample requirement: Fasting venous serum; Specific steps: Incubate the serum sample with the labeled anti-CEA antibody; Add the luminescent substrate; Measure using a chemiluminescence immunoassay analyzer. Normal reference value: Non-smokers: < 5 ng / mL; Smokers: < 10 ng / mL; 5. Lactate dehydrogenase (LDH) detection: Main method: Rate method. Detection process: Sample type: Serum or heparin anticoagulated plasma. Detection principle: Lactate + NAD+ ←(LDH)→ Pyruvate + NADH + H+; Measure the generation rate of NADH; Specific steps: Mix the sample with the substrate reagent; Measure at a wavelength of 340 nm using an automatic biochemical analyzer; Reference interval: 120 - 250 U / L (37 °C).

[0081] The data processing unit includes a multi-dimensional data integration module, a calculation module based on logistic regression, and a risk probability calculation module;

[0082] The multi-dimensional data integration module is used to integrate all the data obtained by the sample collection and detection unit;

[0083] The calculation module based on logistic regression is used to construct a risk prediction model based on logistic regression and establish the calculation formula of the prediction model;

[0084] The risk probability calculation module is used to calculate the probability of a patient developing tuberculosis according to the formula of the prediction model.

[0085] Variable screening

[0086] Through univariate and multivariate logistic regression analyses, a total of 6 predictive variables were finally included: CD274 expression (binary classification), IGRA result (binary classification), lymphocyte count (Ly, continuous variable), lactate dehydrogenase (LDH, continuous variable), carcinoembryonic antigen (CEA, continuous variable), and dyspnea symptom (binary classification);

[0087] Table 2. Results of multivariate logistic regression analysis

[0088]

[0089]

[0090] Construction of the prediction model

[0091] Risk prediction formula: logit(P) = β0 + 1.733×CD274 + 1.592×IGRA - 0.931×Ly - 1.623×LDH - 0.480×CEA - 1.794×dyspnea, where: P is the probability of developing tuberculosis, β0 is the intercept term (16.324), the values of CD274, IGRA, and dyspnea are binary classifications, and Ly, LDH, and CEA are continuous variables.

[0092] Nomogram prediction tool: For the convenience of clinical application, a nomogram of the prediction model was constructed, such as Figure 1 .

[0093] The risk assessment unit includes an individualized nomogram analysis module and a risk stratification module;

[0094] The individualized nomogram analysis module is used to construct a nomogram of the prediction model for convenient clinical application;

[0095] A risk stratification module, which is used to evaluate the risk level of patients according to the risk probability calculated by the prediction model, classify patients into different risk levels, and provide a basis for clinical decision-making.

[0096] Model diagnostic efficacy evaluation

[0097] ROC curve analysis, such as Figure 2 : The ROC curve graph of the model;

[0098] Calibration analysis, such as Figure 3 : The calibration curve graph of the model, the intercept of the calibration curve is 6.984754e-09≈0, and the slope is 9.999999e-01≈1; Hosmer-Lemeshow test: χ 2 = 2.825, P = 0.971.

[0099] Summary of model performance evaluation:

[0100] Table 3. Main evaluation indicators of the model and their 95% confidence intervals

[0101] Index Value 95% Confidence Interval Optimal Cut-off Value 0.555 - AUC 0.891 0.834-0.947 Accuracy (ACC) 0.836 0.834-0.838 Sensitivity (SEN) 0.859 0.785-0.933 Specificity (SPE) 0.796 0.683-0.909 Kappa Coefficient 0.649 0.516-0.782

[0102] Internal validation is carried out by the Bootstrap method for internal validation (1000 repeated samplings):

[0103] C statistic: 0.891 (95% CI: 0.834 - 0.947);

[0104] Adjusted AUC: 0.891 (95% CI: 0.816 - 0.943);

[0105] Optimization index: 0.

[0106] Clinical decision curve analysis: In the risk threshold range of 0.05 - 0.6, the model shows a positive net benefit, the maximum net benefit appears at the risk threshold of 0.1, and the maximum net benefit value is 0.631.

[0107] The clinical decision curve of the model, such as Figure 4 .

[0108] The result output unit includes a risk score visualization module, a clinical decision support module, and a prediction report generation module;

[0109] The risk score visualization module is used to present the calculated risk probability in a graphical way;

[0110] The clinical decision support module is used to classify patients into low-risk or high-risk groups according to the calculated risk probability and in combination with the optimal cut-off value;

[0111] A prediction report generation module for automatically generating a tuberculosis risk assessment report, including various test indexes, risk probabilities, risk stratifications and clinical suggestions of patients.

[0112] System application examples

[0113] 1. Risk assessment process:

[0114] Data collection, entry of patients' basic information, entry of laboratory test results, CD274 expression level, IGRA test results, conventional laboratory indexes and clinical symptom assessment;

[0115] Risk calculation, data preprocessing and standardization, calculation of risk probability using a prediction model and visual assessment through a nomogram;

[0116] Risk stratification is based on the optimal cut-off value of 0.555: low risk: predicted probability < 0.555, high risk: predicted probability ≥ 0.555.

[0117] 2. Clinical application suggestions

[0118] Low-risk patients (predicted probability < 0.555): It is recommended to conduct regular follow-up, reexamine every 3 - 6 months and pay attention to observing changes in symptoms;

[0119] High-risk patients (predicted probability ≥ 0.555): It is recommended to conduct confirmatory tests, consider starting preventive treatment and shortening the follow-up interval.

[0120] 3. Quality control requirements

[0121] Test quality control: Sample collection and processing should strictly follow standard operating procedures, regularly calibrate instruments and use quality control products for quality control;

[0122] Data quality control: Double-entry verification, regular data review and outlier verification.

[0123] Those skilled in the art can understand that the features recited in the various embodiments and / or claims of the present invention can be combined or combined in various ways, even if such combinations or combinations are not explicitly recited in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features recited in the various embodiments and / or claims of the present invention can be combined and combined in various ways. All such combinations and / or combinations fall within the scope of the present invention.

Claims

1. A tuberculosis risk assessment method based on multi-dimensional biomarkers, characterized in that: It includes the following steps: Step 1: Collect the peripheral blood samples of the patient and conduct tests; Step 2: Conduct data processing and integration on the biomarkers; Step 3: Construct a risk prediction model based on logistic regression and calculate the probability of the patient developing tuberculosis; Step 4: Conduct risk stratification according to the calculation results and provide a personalized risk assessment report.

2. The tuberculosis risk assessment method based on multi-dimensional biomarkers according to claim 1, wherein: In Step 2, the biomarkers include the CD274 expression level, the IGRA test result, lymphocyte count (Ly), lactate dehydrogenase (LDH), carcinoembryonic antigen (CEA), and dyspnea symptoms.

3. The tuberculosis risk assessment method based on multi-dimensional biomarkers according to claim 2, wherein: The CD274 expression test is performed by flow cytometry, and the result is determined as positive or negative. The IGRA test is performed using a standardized kit, and the result is determined as positive or negative.

4. The tuberculosis risk assessment method based on multi-dimensional biomarkers according to claim 1, wherein: In Step 3, the prediction model is constructed by logistic regression. The risk prediction calculation formula: logit(P)=β0 + 1.733×CD274 + 1.592×IGRA - 0.931×Ly - 1.623×LDH - 0.480×CEA - 1.794×dyspnea, where: P is the probability of developing tuberculosis, β0 is the intercept term, the values of CD274, IGRA, and dyspnea are binary classifications, and Ly, LDH, and CEA are continuous variables.

5. A tuberculosis risk assessment system based on multi-dimensional biomarkers, characterized in that: It includes a sample collection and detection unit, a data processing unit, a risk assessment unit, and a result output unit; The sample collection and detection unit is used to collect and detect the biomarkers of the patient; The data processing unit is used to integrate multi-dimensional data and calculate the risk probability; The risk assessment unit is used to generate the risk probability and conduct individualized nomogram analysis; The result output unit is used to visualize the risk score and provide clinical decision support.

6. The tuberculosis risk assessment system based on multi-dimensional biomarkers according to claim 5, characterized in that: The sample collection and detection unit is connected to the data processing unit, the data processing unit is connected to the risk assessment unit, and the risk assessment unit is connected to the result output unit.

7. The tuberculosis risk assessment system based on multi-dimensional biomarkers according to claim 5, characterized in that: The sample collection and detection unit includes a peripheral blood sample collection module, a tuberculosis-specific antigen stimulation module, a flow cytometry detection module, and a conventional laboratory detection module; The peripheral blood sample collection module is used to collect the peripheral venous blood samples of the patient; The tuberculosis-specific antigen stimulation module is used to conduct a tuberculosis-specific antigen stimulation experiment on the collected blood samples; The flow cytometry detection module is used to detect the expression level of CD274 on monocytes; The conventional laboratory detection module is used to detect the conventional laboratory indicators of the patient.

8. A tuberculosis risk assessment system based on multi-dimensional biomarkers according to claim 1, characterized in that: The data processing unit includes a multi-dimensional data integration module, a calculation module based on logistic regression, and a risk probability calculation module; The multi-dimensional data integration module is used to integrate all the data obtained by the sample collection and detection unit; The calculation module based on logistic regression is used to construct a risk prediction model based on logistic regression and construct the calculation formula of the prediction model; The risk probability calculation module is used to calculate the probability of the patient developing tuberculosis according to the formula of the prediction model.

9. The tuberculosis risk assessment system based on multi-dimensional biomarkers according to claim 1, characterized in that: The risk assessment unit includes an individualized nomogram analysis module and a risk stratification module; The individualized nomogram analysis module is used to construct a nomogram of the prediction model for convenient clinical application; The risk stratification module is used to evaluate the risk level of patients according to the risk probability calculated by the prediction model, and classify patients into different risk levels to provide a basis for clinical decision-making.

10. The tuberculosis risk assessment system based on multi-dimensional biomarkers according to claim 1, wherein: The result output unit includes a risk score visualization module, a clinical decision support module, and a prediction report generation module; The risk score visualization module is used to present the calculated risk probability in a graphical manner; The clinical decision support module is used to classify patients into low-risk or high-risk groups according to the calculated risk probability in combination with the optimal cut-off value; The prediction report generation module is used to automatically generate a tuberculosis risk assessment report, the content of which includes various test indicators, risk probability, risk stratification, and clinical recommendations of the patient.

Images