Method for constructing model for performing treatment outcome grouping on AIDS (acquired immune deficiency syndrome) and pulmonary tuberculosis patients and application thereof

Abstract

The invention discloses a method for constructing a model for performing treatment outcome grouping on AIDS (acquired immune deficiency syndrome) and phthisis patients and application of the method. The invention provides a method for constructing a model for performing anti-tuberculosis treatment outcome grouping on AIDS (Acquired Immune Deficiency Syndrome) and pulmonary tuberculosis patients, which comprises the following steps of: taking five specific index data of known AIDS and pulmonary tuberculosis patient groups with different anti-tuberculosis treatment outcomes as training samples; training a model for performing anti-tuberculosis treatment outcome grouping on the AIDS and pulmonary tuberculosis patients; the five specific indexes are CAR, extrapulmonary tuberculosis, other pulmonary infections, lung cavities and CD4 + T lymphocyte count. According to the method, clinical doctors can quickly identify AIDS and tuberculosis patients with high treatment failure risk, personalized management is carried out on the AIDS and tuberculosis patients, the method is of great importance to reduction of the occurrence rate of retreatment and drug-resistant tuberculosis, and a new strategy is provided for precise treatment.

Description

Technical Field

[0001] The present invention relates to bioinformatics and healthcare informatics, and particularly to a method for constructing a model for clustering treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis and its application. Background Art

[0002] The combination of human immunodeficiency virus (HIV) and tuberculosis (TB) is a priority for tuberculosis outbreak prevention and control. Compared with people not infected with HIV, people living with HIV are approximately 14 times more likely to develop tuberculosis. According to the Global Tuberculosis Report 2024, there were 10.8 million new tuberculosis cases globally in 2023, of which 662,000 were co-infected with HIV. Tuberculosis caused 1.25 million deaths, including 161,000 HIV-positive individuals, accounting for 12.9% of all tuberculosis-related deaths. Therefore, timely and effective anti-tuberculosis treatment is crucial.

[0003] Due to the impaired immune status of HIV-infected patients and the additional side effects of concomitant medications, patients are more likely to experience anti-tuberculosis treatment failure. In 2023, the treatment success rate for patients co-infected with HIV / TB globally was 79%, lower than the 88% success rate of patients not infected with HIV. The risk of death for AIDS patients with drug-resistant tuberculosis is 4 times that of AIDS patients not infected with HIV. Therefore, identifying the key factors leading to treatment failure is crucial for improving the success rate of tuberculosis treatment and advancing the goals of the WHO End TB Strategy.

[0004] Risk factors associated with poor anti-tuberculosis treatment outcomes include age, gender, smoking, alcoholism, diabetes, malnutrition, and previous treatment history. Most previous studies have focused on tuberculosis patients, and the risk factors associated with anti-tuberculosis treatment failure in patients with AIDS complicated with pulmonary tuberculosis (PTB) are not fully understood. There is also no prediction model to help clinicians quickly identify PTB patients with a higher risk of treatment. Summary of the Invention

[0005] The technical problem to be solved by the present invention is how to cluster or assist in clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis.

[0006] In a first aspect, the present invention claims a method for constructing a model for clustering anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis.

[0007] The method for constructing a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis, which is claimed by the present invention, includes using the data of 5 specific indicators of known groups of AIDS patients complicated with pulmonary tuberculosis with different anti-tuberculosis treatment outcomes as training samples to train a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis.

[0008] The 5 specific indicators are CAR, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavities, and CD4 + T lymphocyte count.

[0009] The CAR is the ratio (value) of the content of C-reactive protein (mg / L) in the serum to the content of albumin (g / L) in the serum, and this ratio is the corresponding data. That is, the CAR is the ratio of the milligrams of C-reactive protein to the grams of albumin per unit volume of serum.

[0010] For extrapulmonary tuberculosis, other pulmonary infections, and pulmonary cavities, they are either present or absent. If "present", the corresponding data is assigned a value of 1, and if "absent", the corresponding data is assigned a value of 0.

[0011] The CD4 + T lymphocyte count is the CD4 + T lymphocyte count in peripheral blood, with the unit of cells / μL (i.e., the number of CD4 + T lymphocytes per 1 μL of peripheral blood), and the value before this unit is the corresponding data.

[0012] In an embodiment of the present invention, the known AIDS patients complicated with pulmonary tuberculosis with different anti-tuberculosis treatment outcomes are two types of AIDS patients complicated with pulmonary tuberculosis with known successful anti-tuberculosis treatment and poor anti-tuberculosis treatment outcomes. Among them, the successful anti-tuberculosis treatment includes cure and completion of treatment. Cure is defined as that for pulmonary tuberculosis patients with positive sputum smear or culture, after completing the specified anti-tuberculosis treatment course, at the end of the last month of treatment, and the last sputum smear or culture result is negative; completion of treatment includes the following two situations: ① For patients with negative etiology, they complete the specified course, and the sputum smear or culture result at the end of the course is negative or no sputum examination is performed; ② For patients with positive etiology, they complete the specified course, there is no sputum examination result at the end of the course, but the last sputum smear or culture result is negative. The poor anti-tuberculosis treatment outcomes include treatment failure, death, or loss to follow-up. Treatment failure means that the patient has no response to treatment, has drug adverse reactions, or obtains evidence of drug resistance and needs to stop treatment or permanently switch to a new treatment plan; death is defined as death due to any reason before or during anti-tuberculosis treatment; loss to follow-up includes abandonment of treatment, interruption of treatment for more than 2 months due to the patient's self-discontinuation of medication or other reasons, or the patient transferring to another hospital resulting in the inability to track the treatment outcome. The meaning of "known AIDS patients complicated with pulmonary tuberculosis with different anti-tuberculosis treatment outcomes" involved hereinafter is the same as this.

[0013] In the present invention, the known group of AIDS patients co-infected with pulmonary tuberculosis with different anti-tuberculosis treatment outcomes consists of at least 200 AIDS patients co-infected with pulmonary tuberculosis. The same applies hereinafter.

[0014] Further, the model can be a multi-factor regression model, such as a multi-factor Cox regression model.

[0015] Even further, the method further includes the step of visualizing the multi-factor regression model to obtain a visualized model. Among them, the visualization can be achieved through a nomogram, such as constructing a nomogram model based on Cox regression.

[0016] In a second aspect, the present invention claims protection for a device for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis.

[0017] The device for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis claimed by the present invention may include a model construction module and a clustering module;

[0018] The model construction module is configured to construct a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis according to the method described in the first aspect above;

[0019] Among them, the model can be a multi-factor regression model, such as a multi-factor Cox regression model; or a nomogram model constructed based on Cox regression.

[0020] The clustering module is configured to receive the data of the 5 specific indicators described in the first aspect above of the subject, input the data of the 5 specific indicators into the model for calculation, compare the calculated result value with a determination threshold, and cluster the anti-tuberculosis treatment outcomes of the subject according to the comparison result;

[0021] Among them, the subject is an AIDS patient co-infected with pulmonary tuberculosis to be clustered for anti-tuberculosis treatment outcomes; the determination threshold is the determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis obtained based on the model.

[0022] Further, the device for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis claimed by the present invention may also be as follows:

[0023] The device includes unit X and unit Y;

[0024] Unit X is used to construct a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis; Unit X includes a data receiving module and a data analysis and processing module;

[0025] The data receiving module is configured to receive the data of the five specific indicators described in the first aspect above for different groups of AIDS patients complicated with pulmonary tuberculosis with known anti-tuberculosis treatment outcomes;

[0026] The data analysis and processing module is configured to receive the data of the five specific indicators of different groups of AIDS patients complicated with pulmonary tuberculosis with known anti-tuberculosis treatment outcomes sent from the data receiving module, and use it as a training sample to train a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis, and obtain a determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis based on the model;

[0027] Unit Y is used to cluster the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis; the subject is an AIDS patient complicated with pulmonary tuberculosis to be clustered for anti-tuberculosis treatment outcomes; Unit Y includes a data input module, a data operation module and a conclusion output module;

[0028] The data input module is configured to input the data of the five specific indicators of the subject;

[0029] The data operation module is configured to receive the data of the five specific indicators of the subject sent from the data input module, retrieve the model stored in the data analysis and processing module in Unit X, perform operations based on the model according to the data of the five specific indicators of the subject, and compare the result value obtained from the operation with the determination threshold stored in the data analysis and processing module in Unit X;

[0030] The conclusion output module is configured to receive the comparison result sent from the data operation module, and determine and output the result of clustering the anti-tuberculosis treatment of the subject according to the comparison result.

[0031] Wherein, the determination threshold can be obtained by the following method: Based on the model, score the contribution degree of each of the five specific indicators to the anti-tuberculosis treatment outcome, add the five scores obtained to get the total score, and obtain the determination threshold based on the total score.

[0032] The anti-tuberculosis treatment outcome of a subject with a total score lower than the determination threshold is better than or candidates to be better than that of a subject with a total score higher than the determination threshold.

[0033] The total score (Total Point) can be obtained by visualizing the multi-factor regression model. The visualization can be achieved through a nomogram.

[0034] In a third aspect, the present invention claims to protect a data processing device.

[0035] The data processing device claimed in the present invention includes a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the following steps:

[0036] (A1) Model construction: Construct a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with tuberculosis according to the method described in the first aspect above.

[0037] Among them, the model can be a multi-factor regression model, such as a multi-factor Cox regression model; or a nomogram model constructed based on Cox regression.

[0038] (A2) Clustering: Receive the data of the 5 specific indicators described in the first aspect above of the subject, input the data of the 5 specific indicators into the model for calculation, compare the calculated result value with the determination threshold, and realize the clustering of the anti-tuberculosis treatment outcomes of the subject according to the comparison result.

[0039] Among them, the subject is an AIDS patient co-infected with tuberculosis whose anti-tuberculosis treatment outcome is to be clustered; the determination threshold is the determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with tuberculosis obtained based on the model.

[0040] Further, the processor can also execute the computer program to implement the following steps:

[0041] (a1) Data reception: Receive the data of the 5 specific indicators described in the first aspect above of different groups of AIDS patients co-infected with tuberculosis with known anti-tuberculosis treatment outcomes.

[0042] (a2) Data analysis and processing: Construct a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with tuberculosis with the data of the 5 specific indicators of different groups of AIDS patients co-infected with tuberculosis with known anti-tuberculosis treatment outcomes, and obtain the determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with tuberculosis based on the model.

[0043] (a3) Data input: Input the data of the 5 specific indicators of the subject; the subject is an AIDS patient co-infected with tuberculosis whose anti-tuberculosis treatment outcome is to be clustered.

[0044] (a4) Data calculation: Calculate based on the model with the data of the 5 specific indicators of the subject, and compare the calculated result value with the determination threshold.

[0045] (a5) Conclusion output: Determine and output the result of clustering the anti-tuberculosis treatment outcomes of the subject according to the obtained comparison result.

[0046] Further, the data processing device may be a computer device.

[0047] Among them, the determination threshold can be obtained by the following method: Based on the model, score the contribution degree of each of the 5 specific indicators to the anti-tuberculosis treatment outcome, add the obtained 5 scores to get the total score, and obtain the determination threshold based on the total score.

[0048] The anti-tuberculosis treatment outcome of the subjects with a total score lower than the determination threshold is better than or candidate better than that of the subjects with a total score higher than the determination threshold.

[0049] The total score (Total Point) can be obtained by visualizing the multi-factor regression model. The visualization can be achieved by a nomogram.

[0050] In a fourth aspect, the present invention claims to protect a method for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis.

[0051] The method for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis claimed by the present invention may include the following steps:

[0052] (A1) Construct a model: Construct a model for clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis according to the method in the first aspect above.

[0053] Among them, the model may be a multi-factor regression model, such as a multi-factor Cox regression model; or a nomogram model constructed based on Cox regression.

[0054] (A2) Cluster: Receive the data of the 5 specific indicators described in the first aspect above of the subject, input the data of the 5 specific indicators into the model for calculation, compare the calculated result value with the determination threshold, and realize clustering of the anti-tuberculosis treatment outcomes of the subject according to the comparison result.

[0055] Among them, the subject is a patient with AIDS complicated with pulmonary tuberculosis to be clustered for anti-tuberculosis treatment outcomes; the determination threshold is the determination threshold for clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis obtained based on the model.

[0056] Further, the method for stratifying or assisting in stratifying the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis claimed by the present invention may also be as follows:

[0057] The method may include the following steps:

[0058] (b1) Data reception: Receive the data of the five specific indicators described in the first aspect above for different groups of AIDS patients with pulmonary tuberculosis with known anti-tuberculosis treatment outcomes;

[0059] (b2) Data analysis and processing: Construct a model for the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis using the data of the five specific indicators of different groups of AIDS patients with pulmonary tuberculosis with known anti-tuberculosis treatment outcomes, and obtain a determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis based on the model; Achieve clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis according to the model and the determination threshold.

[0060] Further, in (b2), the achieving clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis according to the model and the determination threshold can be performed by a method including the following steps (b3) and (b4):

[0061] (b3) Data input: Input the data of the five specific indicators of the subject; The subject is an AIDS patient with pulmonary tuberculosis to be clustered for anti-tuberculosis treatment outcomes;

[0062] (b4) Data operation: Perform an operation based on the model according to the data of the five specific indicators of the subject, and compare the resulting value obtained from the operation with the determination threshold, so as to achieve clustering or assisting in clustering the anti-tuberculosis treatment outcomes of the subject.

[0063] The method may not include the step of obtaining a biological sample from an animal body. The methods may not be directed to a living human body or animal body, but only to data. The method may be an information processing method in which all steps are implemented by a data processing device such as a computer.

[0064] Among them, the determination threshold can be obtained by the following method: Based on the model, score the contribution degree of each of the five specific indicators to the anti-tuberculosis treatment outcome, add the five scores obtained to get a total score, and obtain the determination threshold based on the total score.

[0065] The anti-tuberculosis treatment outcome of a subject with a total score lower than the determination threshold is better than or a candidate to be better than that of a subject with a total score higher than the determination threshold.

[0066] The total score (Total Point) can be obtained by visualizing the multi-factor regression model. The visualization can be achieved through a nomogram.

[0067] In the fifth aspect, the present invention claims protection for a computer program product.

[0068] When the computer program claimed by the present invention is executed by a processor, it implements the steps described in the fourth aspect above.

[0069] Among them, the computer program product can be a software product that mainly realizes its solution through a computer program.

[0070] In a sixth aspect, the present invention claims to protect a computer-readable storage medium.

[0071] A computer program is stored on the computer-readable storage medium claimed by the present invention, and it is characterized in that: when the computer program is executed by a processor, it implements the steps described in the fourth aspect above.

[0072] Among them, the computer-readable storage medium refers to a carrier for storing data, and can be a magnetic tape, a magnetic disk, a floppy disk, an optical disc, a magneto-optical disc, a ROM, a PROM, a VCD, a DVD, a hard disk, a flash memory, a USB flash drive, a CF card, an SD card, an MMC card, an SM card, a Memory Stick or an xD card, etc.

[0073] In the above-mentioned relevant aspects, during the process of determining the determination threshold, the xtile software or other software having the same function or based on the same principle as the xtile software can be used to determine the determination threshold based on the total score.

[0074] Exemplarily, there can be 2 determination thresholds, where the relatively smaller value is denoted as threshold 1, and the relatively larger value is denoted as threshold 2.

[0075] Correspondingly, in the above-mentioned relevant aspects, according to the comparison result, the subjects are grouped according to the anti-tuberculosis treatment outcome, or the result of determining and outputting the grouping of the subjects according to the anti-tuberculosis treatment outcome according to the comparison result can be: the anti-tuberculosis treatment outcome of the subjects whose result value (such as the total score) obtained by operating based on the model according to the 5 specific index data is less than the threshold 1 (corresponding to the low-risk group) is better than or candidate better than the subjects whose result value (such as the total score) obtained by operating based on the model according to the 5 specific index data is greater than or equal to the threshold 1 (corresponding to the medium-high risk group); or, the anti-tuberculosis treatment outcome of the subjects whose result value (such as the total score) obtained by operating based on the model according to the 5 specific index data is less than the threshold 1 (corresponding to the low-risk group) is better than or candidate better than the subjects whose result value (such as the total score) obtained by operating based on the model according to the 5 specific index data is greater than or equal to the threshold 2 (corresponding to the high-risk group).

[0076] Correspondingly, in the above - mentioned relevant aspects, achieving the clustering of the anti - tuberculosis treatment outcomes of the subject according to the comparison results, or achieving the clustering or auxiliary clustering of the anti - tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis based on the model and the determination threshold can be as follows: Subjects with a result value (such as the total score) obtained by calculating based on the model from the 5 specific index data less than the threshold 1 are in the low - risk group, subjects with a result value greater than or equal to the threshold 1 and less than the threshold 2 are in the medium - risk group, and subjects with a result value greater than or equal to the threshold 2 are in the high - risk group.

[0077] In an embodiment of the present invention, the model is a nomogram model constructed based on Cox regression, where the threshold 1 is 358 (total score), and the threshold 2 is 373 (total score). That is, subjects with a detected total score less than 358 are in the low - risk group, subjects with a total score greater than or equal to 358 and less than 373 are in the medium - risk group, and subjects with a total score greater than or equal to 373 are in the high - risk group.

[0078] In this application, the other pulmonary infections are pulmonary infections caused by pathogens other than Mycobacterium tuberculosis.

[0079] In this application, extrapulmonary tuberculosis is tuberculosis occurring in various parts outside the lungs. For example, scrofula, tuberculous meningitis, intestinal tuberculosis, etc.

[0080] In this application, the pulmonary cavity is an air - containing cavity with a complete wall, and the wall is generally more than 1 mm thick. Whether a patient has a pulmonary cavity is determined based on whether there is chest imaging evidence showing the presence of an air - containing cavity.

[0081] In this application, the anti - tuberculosis treatment outcome is the outcome of anti - tuberculosis treatment based on rifampicin or rifabutin. Among them, the anti - tuberculosis treatment based on rifampicin is a quadruple therapy based on rifampicin (R), combined with isoniazid (H), pyrazinamide (Z), and ethambutol (E); the anti - tuberculosis treatment based on rifabutin is a quadruple therapy based on rifabutin (Rfb), combined with isoniazid (H), pyrazinamide (Z), and ethambutol (E).

[0082] In this application, the anti - tuberculosis treatment outcome can be divided into two types: successful anti - tuberculosis treatment and poor anti - tuberculosis treatment outcome. The specific meanings of the successful anti - tuberculosis treatment and the poor anti - tuberculosis treatment outcome are as described in the relevant description above.

[0083] Those skilled in the art can clearly understand that the model construction method claimed in this application and a series of technical solutions derived from this model (covering data processing devices, computer program products, computer-readable storage media, and the method for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis) have significant cross-field universality and potential for scenario expansion. Its application scope not only covers the core scenarios of medical technology (such as disease prognosis prediction), but also can deeply penetrate into the cross-field of artificial intelligence (AI) and medical technology. In addition, the relevant technical solutions can be further expanded to emerging fields such as public health management (such as regional drug-resistant tuberculosis transmission prediction) and intelligent upgrading of medical devices, reflecting its multi-dimensional technology radiation ability. Those skilled in the art can understand that the protection scope of the present invention covers all product forms and implementation carriers directly or indirectly derived from the foregoing application scenarios, such as hardware device types, software platform types, system integration types, etc. The implementation manners of the above product forms include but are not limited to: independent devices, modular components, and cloud services, etc.

[0084] The present invention constructs a nomogram for clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis by collecting comprehensive data of individual AIDS patients with pulmonary tuberculosis, including laboratory indicators and clinical visit records, and using the Least Absolute Shrinkage and Selection Operator (LASSO)-Cox model to determine the risk factors leading to adverse outcomes, so as to help clinicians quickly identify patients with higher treatment risks. In addition, according to the total score of the nomogram, the present invention stratifies the risk of AIDS patients with pulmonary tuberculosis into a low-risk group, a medium-risk group, and a high-risk group, which can help clinicians better identify patients who may have adverse treatment outcomes.

[0085] The present invention preliminarily determines the risk factors affecting the anti-tuberculosis treatment outcomes of AIDS patients, and the developed clustering prediction model for the anti-tuberculosis treatment outcomes of AIDS patients helps clinicians quickly identify AIDS patients with pulmonary tuberculosis who have a higher risk of treatment failure, and conduct personalized management for them, which is crucial for reducing the incidence of retreatment and drug-resistant tuberculosis, and also provides a new strategy for precision treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is a schematic flow chart of the method for clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis according to the present invention.

[0087] Figure 2 It is the research method flow of the present invention.

[0088] Figure 3It is a variable selection graph based on the LASSO-Cox regression model. Among them, A is the association between the logarithm (λ) of the variables included in the LASSO analysis and the regression coefficients. B is the process of screening the optimal λ value in the LASSO regression model by the 10-fold cross-validation method.

[0089] Figure 4 It is a nomogram for predicting the poor treatment outcomes of AIDS patients complicated with pulmonary tuberculosis. The prediction nomogram includes CAR, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavities, and CD4 + T lymphocyte count. Based on these 5 indicators, the total score can be calculated.

[0090] Figure 5 It is the validation of the nomogram model. The ROC curve is used to evaluate the discrimination of the constructed nomogram model, and the calibration curve is used to evaluate the consistency between the predicted value and the actual observed value of the model. Among them, A is the ROC curve and AUC of the training set; B is the ROC curve and AUC of the validation set; C is the calibration curve of the training set; D is the calibration curve of the validation set.

[0091] Figure 6 It is the decision curve (decision curve analysis, DCA) graph of the nomogram, which is used to evaluate the clinical applicability of the model. Among them, A is the DCA of the training set; B is the DCA of the validation set.

[0092] Figure 7 It is the Kaplan-Meier curve (K-M curve) of patients with different risk stratifications, which is used to evaluate the incidence of poor treatment outcomes in patients in different risk clusters. The ordinate is the probability of the occurrence of poor treatment outcomes, and the abscissa is the time after the start of treatment. Detailed implementation manners

[0093] In the present invention, the meaning of the technical term "clustering" is the same as that of "stratification". Clustering or stratification means distinguishing the subjects by "clustering features" or "stratification features".

[0094] In the present invention, the technical term "nomogram" is involved. A nomogram (Alignment Diagram), also known as a nomograph (Nomogram), is a graphical tool based on multiple factor regression analysis. It integrates multiple predictive indicators and represents them with scaled line segments, thereby showing the relationships between various variables in a predictive model. Simply put, by constructing a multiple factor regression model (such as Cox regression, Logistic regression, etc.), according to the contribution degree (the magnitude of the regression coefficient) of each influencing factor in the model to the outcome variable, scores are given to each value level of each influencing factor, and then the scores are added together to obtain the total score. Finally, through the functional conversion relationship between the total score and the occurrence probability of the outcome event, the predicted value of the outcome event of this individual is calculated.

[0095] In the present invention, the technical terms "risk stratification" and "treatment outcome clustering / stratification" are involved. They have the same meaning, both comprehensively evaluating the possible outcomes after a patient's treatment, dividing the risk levels based on multiple factors, and can thus guide clinical decisions, evaluate treatment effects, and predict a patient's prognosis.

[0096] In the present invention, the technical term "Cox regression model" is involved.

[0097] The core objective of the COX regression model: The core of Cox regression is to estimate the impact of covariates on the risk of event occurrence, while allowing the baseline risk function h0(t) of the survival time to remain unknown. This makes the Cox model more flexible than a fully parametric model.

[0098] The mathematical form of the COX regression model

[0099] The risk function form of the Cox model is as follows:

[0100]

[0101] h(t): The conditional risk function at time t given the covariates; h0(t): The baseline risk function, independent of the covariates; X1……X n are covariates, that is, each factor in our multiple factor analysis; β1……βn are the coefficients before the variables, called regression coefficients. Here, the regression coefficients are estimated by the maximum likelihood method. The formula of Cox regression is similar to a generalized linear regression. Finally, a function h(t) can be obtained, which reflects the probability of a patient's death at time t under the influence of several factors.

[0102] In the present invention, the implementation of the Cox model:

[0103] Construct the Cox regression model based on the "survival" package in R software. The input data includes the outcome event Outcome, the time of the occurrence of the outcome event time, and covariates such as CAR, CD4, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavity, ESR, PCT, etc. Bring the data into R software and construct a proportional hazards model based on the "survival" package. The output data includes: 1) HR value: The Cox model is used to measure the degree of influence of a certain factor on the occurrence of adverse outcome events, and HR is the quantification of the degree of influence. Coef is β in the formula, and exp(coef) is our HR value. For a certain covariate X (X is binary), if HR = 1, then whether X is 0 or 1 has no influence on the occurrence of adverse outcome events. If HR > 1, then when X is 1, it will increase the risk of the occurrence of adverse outcome events. If HR < 1, then when X is 1, it will reduce the risk of the occurrence of adverse outcome events; 2) 95% CI and p: CI here means the confidence interval. Two other values that need to be noted are the 95% confidence interval and the significance level p. Generally speaking, if the 95% CI of a certain factor crosses 1, then this factor has no significant influence on survival. At this time, p > 0.05 has no statistical significance. When p < 0.05, it indicates that this factor significantly affects the survival situation.

[0104] The present invention will be further described in detail below in conjunction with specific embodiments. The provided embodiments are only for clarifying the present invention and not for limiting the scope of the present invention. The following provided embodiments can be used as a guide for those of ordinary skill in the art to make further improvements and do not constitute any limitation to the present invention in any way.

[0105] The experimental methods in the following embodiments are all conventional methods unless otherwise specified, and are carried out according to the techniques or conditions described in the literature in this field or according to the product instructions. The materials, reagents, etc. used in the following embodiments can be obtained from commercial channels unless otherwise specified.

[0106] Figure 1 It is a flow chart of the method for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis according to the present invention. The execution subject of the method for clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis provided by the present invention can be any applicable terminal device or network device.

[0107] In step S1, receive the data of 5 specific indicators of groups of AIDS patients complicated with pulmonary tuberculosis with different known anti-tuberculosis treatment outcomes;

[0108] In step S2, a model for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis is constructed using the data of the five specific indicators of AIDS patients co-infected with pulmonary tuberculosis with different known anti-tuberculosis treatment outcomes, and a determination threshold for clustering the anti-tuberculosis treatment outcomes of AIDS patients co-infected with pulmonary tuberculosis is obtained based on the model;

[0109] In step S3, the data of the five specific indicators of the subject are input; the subject is an AIDS patient co-infected with pulmonary tuberculosis to be clustered for anti-tuberculosis treatment outcomes;

[0110] In step S4, based on the data of the five specific indicators of the subject, calculations are performed based on the model, and the result value obtained from the calculations is compared with the determination threshold;

[0111] In step S5, the result of clustering the anti-tuberculosis treatment outcomes of the subject is determined and output according to the obtained comparison result;

[0112] Among them, the five specific indicators are CAR, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavities, and CD4 + T lymphocyte count;

[0113] The CAR is the ratio (value) of the content of C-reactive protein (mg / L) in the serum to the content of albumin (g / L) in the serum, and this ratio is the corresponding data; that is, the CAR is the ratio between the milligrams of C-reactive protein and the grams of albumin per unit volume of serum.

[0114] The extrapulmonary tuberculosis, the other pulmonary infections, and the pulmonary cavities are all either present or absent. If "present", the corresponding data is assigned a value of 1, and if "absent", the corresponding data is assigned a value of 0;

[0115] The CD4 + T lymphocyte count is the CD4 + T lymphocyte count in peripheral blood, with the unit of cells / μL (that is, the number of CD4 + T lymphocytes per 1 μL of peripheral blood), and the value before this unit is the corresponding data.

[0116] The following examples involve biological and medical detection methods:

[0117] 1. CD4 + T lymphocyte count detection

[0118] Take a freshly collected peripheral blood sample containing an appropriate amount of sample cells, add an appropriate amount of reference microspheres and fluorescently labeled antibodies (CD45-FITC / CD4-PE / CD8-ECD / CD3-PC5 Antibody Cocktail, Beckman Coulter International Trading (Shanghai) Co., Ltd.), incubate in the dark at room temperature for 30 min, and after lysing red blood cells, perform CD4 + T lymphocyte, CD8 + The number of T lymphocytes was detected to obtain CD4 + T lymphocyte, CD8 + Absolute number of T lymphocytes

[0119] 2. Detection of C-reactive protein (CRP) content in serum

[0120] (1) Take an appropriate amount of serum sample from peripheral blood, make a mark on the tube and set it aside for later use;

[0121] (2) Specimen numbering and scanning entry: Log in to the laboratory information management system → specimen arrangement system → select the instrument → enter the sample number → scan the barcode number → mark the sample number on the test tube → store it.

[0122] (3) Place the detection system in a clean place.

[0123] (4) Confirm that the ID chip matches the batch number of the reagent kit and insert the ID chip into the instrument.

[0124] (5) Take out the detection buffer and let it return to room temperature.

[0125] (6) Pierce the tin foil on the detection buffer tube; use a sampler to suck up the sample (10 μl of whole blood, serum, plasma or quality control product); wipe off the sample on the outer periphery of the sampler capillary and insert it into the detection buffer tube, tighten the cap, and mix well 5 times.

[0126] (7) Remove the lid of the sampler, drip off two drops and then add two drops of the sample mixture into the sample addition hole of the reaction plate, and let it stand for 3 minutes.

[0127] (8) Insert the reaction plate into the carrier of the i-CHROMAReader immunoassay fluorescence analyzer and press the "Select" key. The instrument will automatically detect the reaction plate. Note: Ensure that the reaction plate is in the correct orientation and is fully inserted to the bottom.

[0128] (9) Read / print the detection results from the display screen of the i-CHROMAReader immunoassay fluorescence analyzer.

[0129] 3. Detection of albumin content in serum

[0130] Reagent and equipment preparation: reaction reagents (pH 4.15 succinic acid buffer solution, 2.4 g / L polyoxyethylene lauryl ether, bromocresol green), albumin standard solution, spectrophotometer, water bath, and serum samples from the peripheral blood of patients.

[0131] (1) Label 3 test tubes and operate according to Table 1:

[0132] Table 1

[0133] Additive B S U Serum - - 21 μl Albumin standard solution - 21 μl - Distilled water 21 μl - - Reaction reagent 3ml 3ml 3ml

[0134] (2) Mix well. After incubating at 37 °C for 1 minute, at a wavelength of 630 nm, zero the spectrophotometer with tube B and read the absorbance A of each tube.

[0135] Data processing and calculation:

[0136] 4. Imaging determination of pulmonary cavities

[0137] A pulmonary cavity refers to an air-containing cavity with a complete wall, and the wall is generally more than 1 mm thick. Whether a patient has a pulmonary cavity is determined based on whether there is evidence of an air-containing cavity in chest imaging. The pulmonary cavities in the present invention mainly include: (1) cavities in infiltrative caseous foci: cavities formed after caseous necrosis occurs in infiltrative lesions. The cavity wall is relatively thin and is mainly composed of proliferated tuberculous granulation tissue, with a thin layer of caseous material on the inner wall. (2) Fibro-caseous cavities and caseous cavities: cavities occurring in lesions. The cavity wall has a thick layer of caseous material, a thin layer of tuberculous granulation tissue, and a fibrous capsule. The fibrous capsule of a tuberculoma is complete. (3) Fibrous cavities: having a typical three-layer structure of caseous necrosis, tuberculous granulation tissue, and fibrous tissue. Fibrous tissue is the main component of the cavity wall. Due to the contraction and traction of fibrous tissue, the cavity has an irregular shape, etc.

[0138] In the present invention, the determination results of pulmonary cavities are divided into two types: present and absent.

[0139] 5. Determination of other pulmonary infectious diseases

[0140] In the present invention, "other pulmonary infections" refer to pulmonary infections caused by pathogens other than Mycobacterium tuberculosis, such as Pneumocystis pneumonia, Klebsiella pneumonia, etc. Specifically, it is necessary to combine the symptoms of the patient such as coughing, expectoration, fever, etc., as well as the results of etiological detection and pulmonary imaging symptoms for judgment.

[0141] In the present invention, the determination results of other pulmonary infections are divided into two types: present and absent.

[0142] 6. Determination of extrapulmonary tuberculosis

[0143] In the present invention, "extrapulmonary tuberculosis" refers to tuberculosis occurring in various parts outside the lungs, including scrofula, tuberculous meningitis, intestinal tuberculosis, etc. The presence of Mycobacterium tuberculosis is determined based on bacteriological or etiological evidence of tissue or body fluid components outside the lungs, and if such evidence exists, it is determined as extrapulmonary tuberculosis; if there are no relevant manifestations and evidence, extrapulmonary tuberculosis cannot be determined.

[0144] In the present invention, the determination results of extrapulmonary tuberculosis are divided into two types: present and absent.

[0145] Example 1: Method for constructing a model for clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis and its application

[0146] Ethical statement: Each participant signed an informed consent form, and the relevant research of the present invention was approved by the Medical Ethics Committee of Beijing You'an Hospital, Capital Medical University.

[0147] 1. Study population

[0148] The study subjects were Chinese adults. Inpatients with AIDS complicated with pulmonary tuberculosis (PTB) who were drug-sensitive and received initial treatment in the Department of Infectious Diseases and Immunology of Beijing You'an Hospital from January 2020 to January 2024 were included. All enrolled patients received a standard anti-tuberculosis treatment regimen based on rifampicin / rifabutin, which can represent the most common tuberculosis management regimen in China. A total of 203 inpatients with AIDS complicated with PTB were included in the present invention as a research cohort for constructing a model for clustering the anti-tuberculosis treatment outcomes of patients with AIDS complicated with pulmonary tuberculosis, including 141 (70%) patients with successful treatment outcomes and 62 (30%) patients with unfavorable treatment outcomes. There were significant differences in the incidences of extrapulmonary tuberculosis, other pulmonary infections, and pulmonary cavities between the two groups of patients (Table 2).

[0149] Among them, the rifampicin-based anti-tuberculosis treatment is a quadruple therapy based on rifampicin (R), combined with isoniazid (H), pyrazinamide (Z), and ethambutol (E). The rifabutin-based anti-tuberculosis treatment is a quadruple therapy based on rifabutin (Rfb), combined with isoniazid (H), pyrazinamide (Z), and ethambutol (E). The relevant reference is "AIDS Group of the Chinese Society of Infectious Diseases, AIDS Group of the Chinese Society of Tropical Diseases and Parasitology. Expert consensus on the diagnosis and treatment of HIV co-infected with Mycobacterium tuberculosis [J]. Chinese Journal of Clinical Infectious Diseases, 2017, 10(2): 81-90."

[0150] The successful anti-tuberculosis treatment includes cure and completion of treatment. Cure is defined as that for pulmonary tuberculosis patients with positive sputum smear or culture, after completing the prescribed anti-tuberculosis treatment course, at the end of the last month of treatment, and the last sputum smear or culture result is negative; Completion of treatment includes the following two situations: ① For patients with negative etiology, they complete the prescribed treatment course, and the sputum smear or culture result at the end of the course is negative or no sputum examination is conducted; ② For patients with positive etiology, they complete the prescribed treatment course, there is no sputum examination result at the end of the treatment course, but the last sputum smear or culture result is negative.

[0151] The poor anti-tuberculosis treatment outcomes include treatment failure, death or loss to follow-up. Treatment failure means that the patient shows no response to treatment, has adverse drug reactions or evidence of drug resistance, and thus needs to stop treatment or permanently switch to a new treatment regimen; Death is defined as death due to any reason before or during anti-tuberculosis treatment; Loss to follow-up includes abandonment of treatment, interruption of treatment for more than 2 months due to the patient's self-discontinuation of medication or other reasons, or the patient transferring to another hospital, resulting in the inability to track the treatment outcome.

[0152] Table 2. Clinical characteristics of AIDS patients complicated with pulmonary tuberculosis with different treatment outcomes

[0153]

[0154]

[0155] 2. Data collection

[0156] Collect the baseline data (data before anti-tuberculosis treatment regimen) and treatment outcomes of all enrolled patients, including basic demographic information of treatment (such as gender, age), the situation of extrapulmonary disseminated infection, the situation of comorbidities, the chest imaging examination results at admission, laboratory test data, etc. (Table 2).

[0157] 3. Construction of clinical prediction model

[0158] The research method process of the present invention is as Figure 2 shown.

[0159] All 203 subjects were randomly divided into a training set (135 cases) and a validation set (68 cases) at a ratio of 2:1. The LASSO regression model was constructed using the data of the training set to determine the potential risk factors related to the occurrence of adverse outcomes. Subsequently, the non-zero coefficients screened by LASSO regression were incorporated into the Cox model to determine the independent risk factors affecting adverse outcomes. The selected prognostic factors were used to establish the final Cox model, and the model was visualized using Nomogram. The ROC, C-index, and calibration curve of the validation set and the test set were used to evaluate the discrimination and calibration of the model. Decision curve analysis (DCA) was used to evaluate the clinical applicability of the prognostic model.

[0160] Among them, the construction of the nomogram using the COX model was completed based on the "regplot" package in R software, and the code is as follows:

[0161]

[0162] The results of the LASSO-Cox model showed that CAR, extrapulmonary tuberculosis (Disseminated infection), other pulmonary infections (Other infectious diseases of the lungs), and cavities in the lungs were independent risk factors for poor outcomes in AIDS patients with PTB, while CD4 + T cell count was a protective factor affecting the outcomes of AIDS patients with PTB (Table 3).

[0163] Table 3. Multivariate COX regression results of AIDS patients with pulmonary tuberculosis

[0164]

[0165]

[0166] Note: * indicates statistically significant differences.

[0167] In the process of constructing the above model, for the 5 factors screened, the CAR refers to the ratio (value) of the content of C-reactive protein (mg / L) in the serum to the content of albumin (g / L) in the serum; that is, the CAR is the ratio of the milligrams of C-reactive protein to the grams of albumin per unit volume of serum. The extrapulmonary tuberculosis, the other pulmonary infections, and the cavities in the lungs are all divided into having or not having. If "having", it is assigned a value of 1, and if "not having", it is assigned a value of 0. The CD4 + T lymphocyte count is the CD4 + T lymphocyte count in peripheral blood, with the unit of cells / μL (that is, the number of CD4 + T lymphocytes per 1 μL of peripheral blood), and the value before this unit is the corresponding data.

[0168] Figure 3 is a variable selection plot based on the LASSO-Cox regression model. Among them, A is the association between the logarithm (λ) of the variables included in the LASSO analysis and the regression coefficients. B is the process of screening the optimal λ value in the LASSO regression model through 10-fold cross-validation.

[0169] Figure 4It is a nomogram constructed using 5 factors in the Cox final model for clustering the anti-tuberculosis treatment outcomes of AIDS patients with pulmonary tuberculosis. As can be seen from the figure, the prediction nomogram includes CAR, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavities, and CD4 + T lymphocyte count. According to the values of these five characteristics, the total score can be calculated, and the treatment failure risk of patients can be evaluated through the treatment failure probability corresponding to the total score.

[0170] Figure 5 For the validation of the nomogram model, the ROC curve is used to evaluate the discrimination of the constructed nomogram model, and the calibration curve is used to evaluate the consistency between the model predicted value and the actual observed value. Among them, A is the ROC curve and AUC of the training set; B is the ROC curve and AUC of the validation set; C is the calibration curve of the training set; D is the calibration curve of the validation set. As can be seen from the figure, the model constructed by the present invention has good discrimination and calibration.

[0171] Figure 6 It is the decision curve (decision curve analysis, DCA) diagram of the nomogram, which is used to evaluate the clinical applicability of the model. Among them, A is the DCA of the training set; B is the DCA of the validation set. As can be seen from the figure, the clinical applicability constructed by the present invention is high.

[0172] The above results show that: the present invention constructs a nomogram prediction nomogram using 5 factors in the Cox final model, and the results of ROC, C-index, calibration curve and decision curve analysis (DCA) of the validation set and the test set all show that the model constructed by the present invention has good accuracy, calibration and clinical applicability.

[0173] X-tile software refers to a tool for determining the optimal cut-off value in survival analysis, which realizes this function based on the exhaustive method. Specifically, the X-tile software traverses all possible cut-off points, calculates the difference in survival curves under each cut-off point, and selects the cut-off value that can maximize the difference in survival curves as the optimal cut-off value. This method helps to improve the accuracy and practicality of survival analysis. The specific operation process is as follows:

[0174] (1) Open the X-tile software and click Analyze;

[0175] (2) Click File-open - select data import, Censor corresponds to OS, that is, the survival status, Survivaltime corresponds to OS.time, the survival time, and marker1 is the variable to be studied (the 5 variables in the present invention);

[0176] (3) Click Do-kaplan-Maier-maker1 in the upper left corner, and select the optimal threshold according to the results given by the software.

[0177] According to the nomogram, two total score thresholds were obtained based on the xtile software, which were 358 and 373 respectively. Furthermore, the risk stratification of the anti-tuberculosis treatment outcomes of the AIDS patients complicated with pulmonary tuberculosis in the present invention was carried out, and the patients were divided into three groups: low risk (total score < 358), medium risk (358 ≤ total score < 373), and high risk (total score ≥ 373). The K-M curve showed significant differences among the three groups of patients. Clinicians should pay attention to these patients with a total score exceeding 358 points (corresponding to medium and high risks) in clinical practice, especially those with a total score exceeding 373 points (corresponding to high risks).

[0178] Figure 7 The K-M curves for patients with different risk stratifications. As can be seen from the figure, the thresholds set in the present invention can well achieve the clustering of patient treatment outcomes.

[0179] The relevant data of 5 indicators of 68 AIDS patients complicated with pulmonary tuberculosis in the validation set of the present invention, as well as the prediction results of anti-tuberculosis treatment outcomes obtained after detection by the model and nomogram of the present invention and the details of the actual anti-tuberculosis treatment outcomes are statistically shown in Table 4. Using R software, with successful anti-tuberculosis treatment as true negative and poor anti-tuberculosis treatment outcome as true positive, ROC curve analysis was performed, and the area under the curve AUC of the validation set was 0.81( Figure 5 in B). It can be seen that using the method of the present invention to predict the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis, the prediction results are accurate and have high specificity and sensitivity.

[0180] Table 4. Data of the validation set (68 cases)

[0181]

[0182]

[0183]

[0184] Note: In "Actual treatment outcome", 0 indicates successful anti-tuberculosis treatment (see the relevant description above for specific meaning), and 1 indicates poor anti-tuberculosis treatment outcome (see the relevant description above for specific meaning). "Prediction result" is divided into low risk (total score < 358), medium risk (358 ≤ total score < 373), and high risk (total score ≥ 373).

[0185] The present invention has been described in detail above. For those skilled in the art, without departing from the spirit and scope of the present invention and without the need for unnecessary experiments, the present invention can be implemented within a relatively wide range under equivalent parameters, concentrations, and conditions. Although specific embodiments of the present invention are given, it should be understood that the present invention can be further improved. In short, according to the principle of the present invention, this application intends to cover any modifications, uses, or improvements to the present invention, including changes made using conventional techniques known in the art that depart from the scope disclosed in this application.

Claims

1. A method for constructing a model for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients, including using the data of 5 specific indicators of known groups of AIDS - associated pulmonary tuberculosis patients with different anti - tuberculosis treatment outcomes as training samples to train a model for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients; The five specific indicators are CAR, extrapulmonary tuberculosis, other pulmonary infections, pulmonary cavities, and CD4 + T lymphocyte count; The CAR is the ratio of the content of C - reactive protein in serum to the content of albumin in serum; Both extrapulmonary tuberculosis, other pulmonary infections, and pulmonary cavities can be either present or absent; The CD4 + T lymphocyte count is the CD4 + T lymphocyte count in peripheral blood.

2. The method according to claim 1, wherein: The model is a multi - factor regression model; The method further includes the step of visualizing the multi - factor regression model.

3. A device for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis, characterized in that: The device includes a model construction module and a clustering module; The model construction module is configured to construct a model for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients according to the method described in claim 1 or 2; The clustering module is configured to receive the data of the 5 specific indicators described in claim 1 or 2 of the subject, input the data of the 5 specific indicators into the model for calculation, compare the calculated result value with a determination threshold, and cluster the anti - tuberculosis treatment outcomes of the subject according to the comparison result; The subject is an AIDS - associated pulmonary tuberculosis patient to be clustered for anti - tuberculosis treatment outcomes; The determination threshold is the determination threshold for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients obtained based on the model.

4. The device according to claim 3, characterized in that: The determination threshold is obtained by the following method: Based on the model, score the contribution degree of each of the 5 specific indicators to the anti - tuberculosis treatment outcome, add the 5 obtained scores to get a total score, and obtain the determination threshold based on the total score.

5. A data processing device, characterized in that: The data processing device includes a memory, a processor, and a computer program stored on the memory, characterized in that: the processor executes the computer program to implement the following steps: (A1) Construct a model: Construct a model for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients according to the method described in claim 1 or 2; (A2) Cluster: Receive the data of the 5 specific indicators described in claim 1 or 2 of the subject, input the data of the 5 specific indicators into the model for calculation, compare the calculated result value with a determination threshold, and cluster the anti - tuberculosis treatment outcomes of the subject according to the comparison result; The subject is an AIDS - associated pulmonary tuberculosis patient to be clustered for anti - tuberculosis treatment outcomes; the determination threshold is the determination threshold for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients obtained based on the model.

6. The data processing device according to claim 5, characterized in that: The determination threshold is obtained by the following method: Based on the model, score the contribution degree of each of the 5 specific indicators to the anti - tuberculosis treatment outcome, add the 5 obtained scores to get a total score, and obtain the determination threshold based on the total score.

7. A method for clustering or assisting in clustering the anti-tuberculosis treatment outcomes of AIDS patients complicated with pulmonary tuberculosis, characterized in that: The method includes the following steps: (A1) Construct a model: Construct a model for clustering the anti - tuberculosis treatment outcomes of AIDS - associated pulmonary tuberculosis patients according to the method described in claim 1 or 2; (A2) Subgrouping: Receive the five specific index data described in claim 1 or 2 of the subject, input the five specific index data into the model for calculation, compare the calculated result value with the determination threshold, and achieve subgrouping of the anti-tuberculosis treatment outcome of the subject according to the comparison result; The subject is a patient with AIDS complicated with pulmonary tuberculosis to be subgrouped for the anti-tuberculosis treatment outcome; the determination threshold is the determination threshold for subgrouping the anti-tuberculosis treatment outcome of patients with AIDS complicated with pulmonary tuberculosis obtained based on the model.

8. The method according to claim 7, characterized in that: The determination threshold is obtained by the following method: Based on the model, score the contribution degree of each of the five specific indexes to the anti-tuberculosis treatment outcome, add the five obtained scores to get the total score, and obtain the determination threshold based on the total score.

9. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method described in claim 7 or 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method described in claim 7 or 8.

Images