Model establishment method for identifying nontuberculous mycobacterium lung disease and pulmonary tuberculosis

Abstract

The invention discloses a method for establishing a model for discriminating a nontuberculous mycobacterium lung disease and a pulmonary tuberculosis, and belongs to the technical field of constructing a medical mathematical model by combining clinical data with iconography images, and the method comprises the following steps: acquiring CT images of a nontuberculous mycobacterium lung disease patient and a pulmonary tuberculosis patient, extracting and quantifying related iconography indexes, and establishing a model for discriminating the nontuberculous mycobacterium lung disease and the pulmonary tuberculosis patient; taking as a first data set; calling clinical case data of the two groups of patients, extracting related clinical indexes and quantifying the related clinical indexes as a second data set; based on the first data set and the second data set, performing single-factor Logistic regression analysis to obtain an influence factor; carrying out multi-factor Logistic regression analysis on the basis of the influence factors to obtain independent influence factors; and based on the independent influence factors, constructing a medical mathematical model for identifying the nontuberculous mycobacterium lung disease and the pulmonary tuberculosis, verifying the medical mathematical model, and further establishing a webpage scoring device suitable for clinical application. The method is used for identifying the nontuberculosis mycobacterium lung disease patient and the tuberculosis patient, and family and social burdens caused by misdiagnosis and treatment are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of building medical mathematical models by combining clinical data with imaging images, and in particular to a model building method for distinguishing non-tuberculous mycobacterium pulmonary disease from pulmonary tuberculosis. Background Art

[0002] Nontuberculous mycobacteria (NTM) refers to a large class of mycobacteria other than Mycobacterium tuberculosis complex and Mycobacterium leprae complex. Currently, there are nearly 200 species and 13 subspecies discovered, most of which are opportunistic pathogens. They can invade the human body through the respiratory tract, gastrointestinal tract, and skin, and invade multiple parts of the body, such as the lungs, skin soft tissue, lymph nodes, and bones. The lungs are one of the most common sites of disease, which is called nontuberculous mycobacterium lung disease (NTM lung disease, NTM-LD). In recent years, the incidence and prevalence of NTM-LD have continued to increase worldwide, and NTM-LD has now become one of the main causes of chronic lung infection.

[0003] NTM-LD often presents with symptoms such as cough and sputum, hemoptysis, chest tightness and shortness of breath, low-grade fever and night sweats in clinical practice, which are very similar to the clinical manifestations of pulmonary tuberculosis (PTB-LD) caused by infection with Mycobacterium tuberculosis. The most important examination method for diagnosing NTM-LD requires the culture and identification of bacterial species from the patient's alveolar wash fluid, which takes about 3-4 weeks, which often affects the early diagnosis of NTM-LD in clinical practice. On the other hand, most NTMs also have natural anti-tuberculosis drug resistance. The treatment methods of the two diseases are different, which leads to clinical misdiagnosis. Therefore, it is particularly important to accurately distinguish between the two. With the development of analytical methods, the application of multi-factor-based mathematical model construction in the medical field has gradually become widespread. At present, there are many reports on the differential diagnosis of NTM-LD and PTB-LD patients by electronic computed tomography (CT) at home and abroad, but there are no reports on in-depth exploration of the CT imaging of the two, quantitative index analysis and mathematical models combined with clinical data.

[0004] Therefore, how to provide an identification model that can distinguish non-tuberculous mycobacterial lung disease from pulmonary tuberculosis and avoid the family and social burden caused by misdiagnosis and mistreatment is a technical problem that needs to be solved urgently by technical personnel in this field. Summary of the invention

[0005] To this end, the present invention provides a method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, so as to solve the problem of misdiagnosis and mistreatment caused by the difficulty in distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis in the prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis comprises the following steps:

[0008] Step S1: obtaining CT images of patients with non-tuberculous mycobacterial lung disease and patients with pulmonary tuberculosis, extracting and quantifying relevant imaging indicators as a first data set;

[0009] Step S2: retrieve clinical case data of patients with non-tuberculous mycobacterium pulmonary disease and patients with pulmonary tuberculosis, extract relevant clinical indicators and quantify them as the second data set;

[0010] Step S3: Based on the first data set and the second data set, the influencing factors for distinguishing non-tuberculous mycobacterial lung disease and pulmonary tuberculosis are screened through univariate logistic regression analysis;

[0011] Step S4: Based on the influencing factors of non-tuberculous mycobacterial lung disease and pulmonary tuberculosis, multivariate logistic regression analysis is performed to screen out independent influencing factors for distinguishing non-tuberculous mycobacterial lung disease and pulmonary tuberculosis;

[0012] Step S5: constructing a medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis based on the independent influencing factors of nontuberculous mycobacterial lung disease and pulmonary tuberculosis.

[0013] Furthermore, the method further comprises step S6 of validating the model, specifically, obtaining other non-tuberculous mycobacterium lung disease patients and pulmonary tuberculosis patients as validation populations to validate the accuracy and discrimination of the model.

[0014] Furthermore, the method further includes step S7 of applying a model, specifically, making a webpage scorer based on the medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, obtaining a patient score, and estimating the probability of the patient suffering from nontuberculous mycobacterial lung disease or pulmonary tuberculosis according to the score result.

[0015] Furthermore, the first data set includes thin-walled cavities, centrilobular nodules, disseminated lesions, bronchiectasis, involvement of multiple lobes and segments, exudation, fibrous cords, and mediastinal lymphadenopathy.

[0016] Furthermore, the second data set includes age, gender, cough and sputum, fever, wheezing, chest pain, hemoptysis, diabetes, autoimmune system diseases, chronic obstructive pulmonary disease, malignant tumor diseases, interstitial lung disease, AIDS, blood system diseases, pleural thickening and hypoproteinemia.

[0017] Furthermore, the influencing factors of non-tuberculous mycobacterial lung disease and pulmonary tuberculosis are age, cough and sputum, hemoptysis, thin-walled cavities, centrilobular nodules, bronchiectasis, involvement of multiple lobes and segments, exudation, diabetes, autoimmune system diseases, chronic obstructive pulmonary disease and blood system diseases.

[0018] Furthermore, the independent influencing factors are cough and sputum, hemoptysis, thin-walled cavities, centrilobular nodules, bronchiectasis, diabetes and autoimmune system diseases.

[0019] Furthermore, the medical mathematical model for distinguishing non-tuberculous mycobacterial lung disease from pulmonary tuberculosis is specifically:

[0020] Obtain the independent influencing factors of the patients, and based on the degree of influence of the independent influencing factors on the differentiation of nontuberculous mycobacterial lung disease and pulmonary tuberculosis, use the "rms package" in R software to construct a nomogram and assign influence weights to different independent influencing factors;

[0021] According to the impact weights of different independent impact factors and based on the scoring scale, determine the scoring values ​​of different independent impact factors in the "yes" or "no" state;

[0022] Obtain the score values ​​of different independent influencing factors under clinical analysis and calculate the total score of the patient;

[0023] A total score scale is established based on the minimum and maximum score ranges of the patient's scores, and the total score is divided into different score intervals;

[0024] Determine the scoring interval that each patient's total score falls into, determine the probability that the patient suffers from non-tuberculous mycobacterium pulmonary disease, and then calculate the probability of the patient suffering from non-tuberculous mycobacterium pulmonary disease according to the formula P 肺结核 =1-P 非结核分枝杆菌肺病 , calculate the probability of the patient developing tuberculosis.

[0025] The present invention has the following advantages:

[0026] The present application obtains CT images of patients with nontuberculous mycobacterium lung disease and patients with pulmonary tuberculosis, extracts relevant imaging indicators and quantifies them as a first data set; retrieves clinical case data of patients with nontuberculous mycobacterium lung disease and patients with pulmonary tuberculosis, extracts relevant clinical indicators and quantifies them as a second data set; based on the first data set and the second data set, a univariate Logistic regression analysis is performed to screen out influencing factors for distinguishing nontuberculous mycobacterium lung disease and pulmonary tuberculosis; based on the influencing factors of nontuberculous mycobacterium lung disease and pulmonary tuberculosis, a multivariate Logistic regression analysis is performed to screen out independent influencing factors for distinguishing nontuberculous mycobacterium lung disease and pulmonary tuberculosis; based on the independent influencing factors of nontuberculous mycobacterium lung disease and pulmonary tuberculosis, a medical mathematical model for distinguishing nontuberculous mycobacterium lung disease from pulmonary tuberculosis is constructed.

[0027] This application aims to evaluate the basic clinical data, symptoms and signs, CT imaging quantitative indicators and concurrent underlying diseases of NTM-LD and PTB-LD patients admitted during the same period, to construct a disease prediction model and verify its evaluation effect, in order to provide a predictive tool for the early non-invasive differential diagnosis of NTM-LD and PTB-LD, and to avoid the family and social burden caused by misdiagnosis and mistreatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0029] The structures, proportions, sizes, etc. illustrated in this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with the technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantial technical significance. Any structural modification, change in proportion or adjustment of size shall still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0030] Figure 1 The present invention provides a flow chart of a method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis.

[0031] Figure 2 This is a nomogram of the NTM-LD and PTB-LD identification model provided by the present invention.

[0032] Figure 3 This is a validation diagram of the NTM-LD and PTB-LD identification models provided by the present invention.

[0033] Figure 4 A web page scorer for the NTM-LD and PTB-LD identification models provided by the present invention. DETAILED DESCRIPTION

[0034] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] NTM-LD and PTB-LD are two completely different diseases with completely different treatment plans, but they have similar clinical and imaging manifestations, which increases the difficulty of clinical diagnosis and treatment. Studies have shown that delaying the diagnosis and treatment of NTM-LD will lead to a worse prognosis. It can be seen that accurate diagnosis of NTM-LD or PTB-LD is particularly important for appropriate and timely treatment. At present, the gold standard for distinguishing NTM-LD from PTB-LD is still invasive biopsy, which is invasive and has certain risks. Therefore, it is very necessary to try to create an effective and non-invasive method to distinguish NTM-LD from PTB-LD. Some researchers have studied the differential diagnosis of NTM-LD and PTB-LD imaging images. The CT signs of the two overlap and lack specificity. Doctors cannot make judgments based on the presence or absence of a single lesion. Therefore, this study comprehensively considers the occurrence, location, volume, shape of multiple lesions and combines clinical data. A multi-factor mathematical model is used to distinguish NTM-LD from PTB-LD, and a web scorer is further produced, which is conducive to the application of clinicians.

[0036] A method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, such as Figure 1 As shown, the following steps are included:

[0037] Step S1: obtaining CT images of patients with non-tuberculous mycobacterial lung disease and patients with pulmonary tuberculosis, extracting and quantifying relevant imaging indicators as a first data set;

[0038] Step S2: retrieve clinical case data of patients with non-tuberculous mycobacterium pulmonary disease and patients with pulmonary tuberculosis, extract relevant clinical indicators and quantify them as the second data set;

[0039] Step S3: Based on the first data set and the second data set, the influencing factors for distinguishing nontuberculous mycobacterial lung disease and pulmonary tuberculosis are screened through univariate logistic regression analysis;

[0040] Step S4: Based on the influencing factors of nontuberculous mycobacterial lung disease and pulmonary tuberculosis, multivariate logistic regression analysis was performed to screen for independent influencing factors for distinguishing nontuberculous mycobacterial lung disease and pulmonary tuberculosis;

[0041] Step S5: Based on the independent influencing factors of nontuberculous mycobacterial lung disease and pulmonary tuberculosis, a medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis is constructed.

[0042] Step S6 verifies the model, specifically: obtaining other non-tuberculous mycobacterium lung disease patients and pulmonary tuberculosis patients as a verification population to verify the accuracy and discrimination of the model.

[0043] Step S7 applies the model, specifically: based on the medical mathematical model for distinguishing non-tuberculous mycobacterial lung disease from pulmonary tuberculosis, a webpage scorer is created to obtain the patient score, and the probability of the patient suffering from non-tuberculous mycobacterial lung disease is estimated according to the score result, and then the probability of the patient suffering from non-tuberculous mycobacterial lung disease is estimated according to the formula P 肺结核 =1-P 非结核分枝杆菌肺病 , calculate the probability of the patient developing tuberculosis.

[0044] The first dataset included thin-walled cavities, centrilobular nodules, disseminated lesions, bronchiectasis, multilobar and multisegmental involvement, exudates, fibrous cords, and mediastinal lymphadenopathy. The second dataset included age, sex, cough and sputum, fever, wheezing, chest pain, hemoptysis, diabetes, autoimmune diseases, chronic obstructive pulmonary disease, malignant neoplastic diseases, interstitial lung disease, AIDS, hematologic diseases, pleural thickening, and hypoproteinemia.

[0045] The influencing factors for differentiating nontuberculous mycobacterial lung disease from pulmonary tuberculosis were age, cough and sputum, hemoptysis, thin-walled cavitation, centrilobular nodules, bronchiectasis, multi-lobular and multi-segment involvement, exudation, diabetes, autoimmune system diseases, chronic obstructive pulmonary disease, and blood system diseases. The independent influencing factors were cough and sputum, hemoptysis, thin-walled cavitation, centrilobular nodules, bronchiectasis, diabetes, and autoimmune system diseases.

[0046] Specifically, this study included 300 NTM-LD patients and 300 PTB-LD patients. They were randomly divided into a training group and a validation group at a ratio of 2:1. There was no statistical difference in gender between the two groups (P>0.05), but there was a statistical difference in age distribution (P<0.05). The demographic characteristics of the training group and the validation group are shown in Table 1.

[0047] Table 1 Demographic characteristics of the training group and internal validation group

[0048]

[0049]

[0050] The data indicators in the first and second data sets were subjected to univariate logistic regression analysis to obtain the influencing factors for distinguishing nontuberculous mycobacterial lung disease and pulmonary tuberculosis. The screening results are shown in Table 2. There were significant differences in the distribution of cough and sputum, hemoptysis; thin-walled cavities, central nodules in lobule, disseminated foci, bronchiectasis, multi-lobular and multi-segment involvement, exudation, and fibrous cords between the two groups (all P < 0.05). There were no significant differences in the distribution of fever, wheezing, chest pain, malignant tumor diseases, interstitial lung disease, AIDS, pleural thickening, hypoproteinemia, and mediastinal lymphadenopathy between the two groups (all P > 0.05).

[0051] Table 2 Univariate Logistic regression analysis of nontuberculous mycobacterial lung disease and pulmonary tuberculosis

[0052]

[0053]

[0054] Note: "*" in the table represents P < 0.05, indicating that the index has a statistically significant difference in the distribution between nontuberculous mycobacterial lung disease and pulmonary tuberculosis, that is, it is an influencing factor for distinguishing nontuberculous mycobacterial lung disease and pulmonary tuberculosis.

[0055] The indicators with statistically significant differences in univariate logistic regression analysis were used as independent variables for multivariate logistic regression analysis. The final univariate analysis showed that age, cough and sputum, hemoptysis, thin-walled cavitation, central lobular nodules, bronchiectasis, multi-lobular and multi-segment involvement, exudation, diabetes, autoimmune system diseases, chronic obstructive pulmonary disease, and blood system diseases were included in the multivariate regression analysis. The results of multivariate analysis showed that cough and sputum, hemoptysis, thin-walled cavitation, central lobular nodules, bronchiectasis, diabetes, and autoimmune system diseases were independent risk factors for the differentiation of NTM-LD and PTB-LD, and finally entered the model (see Table 3).

[0056] Table 3 Multivariate Logistic regression analysis of nontuberculous mycobacterial lung disease and pulmonary tuberculosis

[0057]

[0058] Note: “*” in the table represents P < 0.05, indicating that when the indicator is included in the multivariate model, the effect is still statistically significant and cannot be simply attributed to random errors, that is, it is an independent influencing factor for identifying nontuberculous mycobacterial lung disease and pulmonary tuberculosis.

[0059] The specific medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis is:

[0060] The independent influencing factors for distinguishing the two groups of patients were obtained. Based on the degree of influence of the independent influencing factors on distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, a nomogram was constructed using the "rms package" in Rsoftware to assign influence weights to different independent influencing factors.

[0061] According to the impact weights of different independent impact factors and based on the scoring scale, determine the scoring values ​​of different independent impact factors in the "yes" or "no" state;

[0062] Obtain the score values ​​of different independent influencing factors under clinical analysis and calculate the total score of the patient;

[0063] A total score scale is established based on the minimum and maximum score ranges of the patient's scores, and the total score is divided into different score intervals;

[0064] Determine the scoring interval into which the patient's total score falls, determine the probability that the patient suffers from nontuberculous mycobacterium pulmonary disease, and then calculate the probability of the patient suffering from nontuberculous mycobacterium pulmonary disease according to the formula P 肺结核 =1-P 非结核分枝杆菌肺病 , calculate the probability of the patient developing tuberculosis.

[0065] like Figure 2 As shown in the figure, the nomogram prediction model for differential diagnosis of NTM-LD and PTB-LD was constructed by using the "rms package" in Rsoftware based on the independent influencing factors screened out by multivariate logistic regression analysis. The scores of each independent influencing factor are as follows: if the patient has symptoms of cough and sputum, the score is 50, and if there is no cough and sputum, the score is 0; if the patient has symptoms of hemoptysis, the score is 70, and if there is no hemoptysis, the score is 0; if the patient has symptoms of thin-walled cavitation, the score is 38, and if there is no thin-walled cavitation, the score is 0; if the patient has symptoms of central lobular nodules, the score is 100, and if there is no central lobular nodule, the score is 0; if the patient has symptoms of bronchiectasis, the score is 88, and if there is no bronchiectasis, the score is 0; if the patient has symptoms of diabetes, the score is 80, and if there is diabetes, the score is 0; if the patient has autoimmune system disease, the score is 88, and if there is no autoimmune system disease, the score is 0.

[0066] According to the clinical manifestations of cough and sputum, hemoptysis, thin-walled cavities, central lobular nodules, bronchiectasis, diabetes and autoimmune diseases, vertical lines are drawn on the nomogram to find the scores corresponding to the score lines, and the total score is added up. According to the total score line, the probability of the patient developing NTM-LD is found on the risk line of NTM-LD, and then the probability of the patient developing PTB-LD is calculated. For example, if the probability of diagnosing a patient with nontuberculous mycobacterium lung disease is 20%, then the probability of diagnosing the patient with pulmonary tuberculosis is 80%, and the patient is more likely to be diagnosed with pulmonary tuberculosis clinically.

[0067] like Figure 3 As shown, the nomogram was constructed using the "rms package" in Rsoftware, and Bootstrap internal validation was used to draw the calibration curve and decision curve to evaluate and verify the prediction model. The area under the ROC curve (AUC) of the training group was 0.872 (95% CI: 0.838-0.906), with a specificity of 83% and a sensitivity of 75%, indicating that the nomogram model had good discrimination ( Figure 3 A), the calibration curve of the training group shows that the predicted probability is consistent with the actual probability ( Figure 3 B). In addition, 200 patients were used as the validation group to verify the model, of which 85 out of 100 NTM-LD patients were correctly diagnosed, and 72 out of 100 PTB-LD patients were correctly diagnosed, with a diagnostic accuracy of 78.5%. The AUC was 0.846 (95% CI: 0.792-0.900), the specificity was 86%, and the sensitivity was 71% ( Figure 3 C), indicating that the nomogram model still showed good differential diagnostic ability. The calibration curve of the validation group showed that the probability of this model distinguishing NTM-LD from PTB-LD was basically consistent with the actual probability, and it had good calibration performance ( Figure 3 D). In addition, the clinical decision curve (DCA) was used to evaluate the clinical utility of the nomogram ( Figure 3 E-3F). The DCA of the decision curves of the training group and the validation group shows that within a large threshold range, the net benefit of the model is higher than the net benefit in extreme cases. The range of thresholds that can be selected is relatively large, indicating that the model has good clinical applicability.

[0068] NTM-LD and PTB-LD are similar in clinical symptoms. The results of this study showed that there was no significant statistical difference in the clinical symptoms of fever, wheezing, and chest pain between the two groups (P>0.05), while the incidence of cough, sputum, and hemoptysis in the NTM-LD group was significantly higher than that in the PTB-LD group (P<0.05). NTM-LD and PTB-LD also have similarities in imaging. The results of this study showed that there was no significant difference in imaging between the two groups in terms of dissemination, involvement of multiple lobes and segments, exudation, and fibrous cords (P>0.05), while the incidence of lobular centrifugal nodules, bronchiectasis, and thin-walled cavities in the NTM-LD group was significantly higher than that in the PTB-LD group (P<0.05), which is consistent with previous research results. The reason why NTM-LD patients develop such imaging features may be related to the following factors: NTM is weaker in virulence than PTE, and NTM-LD pathology is mostly granulomatous lesions, mainly composed of epithelial-like cells and lymphocytes, with less caseous necrosis. The diseased tissue can involve the bronchioles and large airways, causing damage to the bronchial muscle layer, causing airway stenosis, increased intrabronchial pressure, and further formation of bronchiectasis and thin-walled cavities. The necrotic material in the cavity can spread along the bronchial wall to the distal end to form centrilobular nodules.

[0069] The results of this study showed that there was no statistically significant difference between the two groups in terms of common underlying diseases such as hypoproteinemia, chronic obstructive pulmonary disease, malignant tumor diseases, interstitial lung disease, AIDS, and blood system diseases (P>0.05), but the probability of NTM-LD combined with autoimmune system diseases was significantly higher than that of the PTB-LD group (P<0.05), and the probability of PTB-LD combined with diabetes was significantly higher than that of NTM-LD (P<0.05). This may be because NTM is a conditional pathogen, and people with autoimmune system diseases have weak immunity and reduced levels of various body functions, making it easier for pathogens to invade immunodeficient patients.

[0070] In this application, we selected the most significant indicators (cough and sputum, hemoptysis, lobular central nodules, disseminated lesions, bronchiectasis, fibrous cords, combined diabetes and autoimmune system diseases) to construct a prediction model based on the results of multivariate logistic regression analysis. This prediction model showed good diagnostic performance in both the training set and the validation set. The training set results had a sensitivity of 75%, a specificity of 83%, an AUC value of 0.872, and a 95% CI of 0.838-0.906. The validation set results had an accuracy of 78.5%, a sensitivity of 71%, a specificity of 86%, an AUC value of 0.846, and a 95% CI of 0.792-0.900. Based on the sensitivity and specificity obtained from the test, it was verified that the model has high accuracy and discrimination.

[0071] like Figure 4As shown, the present application further combines R language software to produce a web scorer, which can be promoted and applied in the clinical frontline. The patient score can be obtained by simply inputting the "yes" or "no" status of the independent influencing factor into the web scorer, and the probability of the patient suffering from non-tuberculous mycobacterial lung disease or pulmonary tuberculosis can be predicted. For patients with a high risk of NTM-LD, PTB-LD can be excluded as early as possible, and the "gold standard" of etiology can be used for timely diagnosis, and early intervention treatment can be carried out at the same time, thereby improving the quality of diagnosis and treatment and improving patient prognosis.

[0072] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, characterized in that: The following steps are involved: Step S1: obtaining CT images of patients with non-tuberculous mycobacterial lung disease and patients with pulmonary tuberculosis, extracting and quantifying relevant imaging indicators as a first data set; Step S2: retrieve clinical case data of patients with non-tuberculous mycobacterium pulmonary disease and patients with pulmonary tuberculosis, extract relevant clinical indicators and quantify them as the second data set; Step S3: Based on the first data set and the second data set, the influencing factors for distinguishing non-tuberculous mycobacterial lung disease and pulmonary tuberculosis are screened through univariate logistic regression analysis; Step S4: Based on the influencing factors of non-tuberculous mycobacterial lung disease and pulmonary tuberculosis, multivariate logistic regression analysis is performed to screen out independent influencing factors for distinguishing non-tuberculous mycobacterial lung disease and pulmonary tuberculosis; Step S5: constructing a medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis based on the independent influencing factors of nontuberculous mycobacterial lung disease and pulmonary tuberculosis.

2. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The method further includes step S6 of validating the model, specifically, obtaining other non-tuberculous mycobacterium lung disease patients and pulmonary tuberculosis patients as validation populations to validate the accuracy and discrimination of the model.

3. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The method further includes step S7 of applying a model, specifically, making a webpage scorer based on the medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis, obtaining a patient score, and estimating the probability of the patient suffering from nontuberculous mycobacterial lung disease or pulmonary tuberculosis according to the score result.

4. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The first dataset included thin-walled cavities, centrilobular nodules, disseminated lesions, bronchiectasis, multilobar and multisegmental involvement, infiltration, fibrous cords, and mediastinal lymphadenopathy.

5. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The second data set includes age, gender, cough and sputum, fever, wheezing, chest pain, hemoptysis, diabetes, autoimmune diseases, chronic obstructive pulmonary disease, malignant tumor diseases, interstitial lung disease, AIDS, blood system diseases, pleural thickening and hypoproteinemia.

6. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The influencing factors of non-tuberculous mycobacterial lung disease and pulmonary tuberculosis are age, cough and sputum, hemoptysis, thin-walled cavities, centrilobular nodules, bronchiectasis, involvement of multiple lobes and segments, exudation, diabetes, autoimmune system diseases, chronic obstructive pulmonary disease and blood system diseases.

7. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The independent influencing factors are cough and sputum, hemoptysis, thin-walled cavities, centrilobular nodules, bronchiectasis, diabetes and autoimmune system diseases.

8. The method for establishing a model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis according to claim 1, characterized in that: The medical mathematical model for distinguishing nontuberculous mycobacterial lung disease from pulmonary tuberculosis is specifically: Obtain the independent influencing factors of the patients, and based on the degree of influence of the independent influencing factors on the differentiation of non-tuberculous mycobacterial lung disease and pulmonary tuberculosis, use the "rms package" in Rsoftware to construct a nomogram and assign influence weights to different independent influencing factors; According to the impact weights of different independent impact factors and based on the scoring scale, determine the score values ​​of different independent impact factors in the "yes" or "no" state; Obtain the score values ​​of different independent influencing factors under clinical analysis and calculate the total score of the patient; A total score scale is established based on the minimum and maximum score ranges of the patient's scores, and the total score is divided into different score intervals; Determine the scoring interval that each patient's total score falls into, determine the probability that the patient suffers from non-tuberculous mycobacterium pulmonary disease, and then calculate the probability of the patient suffering from non-tuberculous mycobacterium pulmonary disease according to the formula P 肺结核 =1-P 非结核分枝杆菌肺病 , calculate the probability of the patient developing tuberculosis.

Images