Disease prediction method and system based on human immune age

Abstract

The invention discloses a disease prediction method and system based on human immune age. According to the method, a specific immune marker in blood of a subject is detected, the immune age of the subject is calculated by using a multiple linear regression model, and further based on the immune age, a quantitative probability risk value of the respiratory system disease of an individual is generated by adopting age group binary logistic regression analysis, so that early-stage and probabilistic early warning of the disease risk is realized. The invention further provides a model for quantitatively associating the immune age with the lung function, and the model can be used for comprehensive risk assessment of specific diseases such as chronic obstructive pulmonary diseases. The method has the advantages of model robustness, high interpretability and low cost, and is suitable for health screening, risk stratification and personalized medical intervention.

Description

Technical Field

[0001] This invention relates to a method for predicting disease probability based on human immune age, and also to a corresponding disease probability prediction system, belonging to the field of healthcare informatics technology. Background Technology

[0002] The functional decline of the immune system with age, known as "immunosenescence," is a significant factor contributing to increased susceptibility to infections, tumors, and chronic diseases. Accurately assessing an individual's "immune age" is crucial for measuring their overall immune status, predicting age-related disease risk, and developing personalized health intervention strategies.

[0003] Currently, the assessment of immune age primarily relies on the detection and analysis of various immune-related biomarkers. Common detection methods include: (1) Immune cell phenotype analysis: The proportion and absolute number of immune cells such as T cells, B cells, NK cells, and monocytes in peripheral blood are accurately measured by flow cytometry. However, the results of this method are easily affected by transient factors such as recent infection and stress, and the stability of detection at a single time point is insufficient, which affects the accuracy of the assessment.

[0004] (2) Serum inflammatory marker detection: The inflammatory state of the system is reflected by detecting the levels of C-reactive protein (CRP) and pro-inflammatory cytokines such as interleukins (IL-6, IL-1β, IL-8) and TNF-α. However, the inflammatory response itself is highly nonspecific, and its elevated levels may be due to various non-aging factors such as infection, autoimmune diseases, and metabolic abnormalities, resulting in low specificity of this method and significant interference with the determination of immune age.

[0005] In recent years, researchers have begun to attempt to integrate multi-dimensional data (such as cell phenotype, serum biomarkers, epigenetics, transcriptomics, etc.) and use complex machine learning algorithms to build comprehensive predictive models in order to improve the comprehensiveness and accuracy of immune age assessment. However, these methods have extremely high requirements for sample size, data quality, and computational resources. The model construction process is complex, and they generally suffer from poor medical interpretability and difficulty in being widely promoted and popularized in clinical practice.

[0006] Therefore, the current technological field still lacks an immune age assessment scheme that combines good accuracy, strong interpretability, and clinical applicability. Of particular importance is how to leverage such immune age assessments to achieve effective and low-cost probability prediction of the risk of specific diseases (such as respiratory diseases), which has become a key technical problem urgently needing to be solved in health management and early intervention. Summary of the Invention

[0007] In view of the shortcomings of the prior art, the primary technical problem to be solved by the present invention is to provide a method for predicting the probability of diseases based on human immune age.

[0008] Another technical problem to be solved by the present invention is to provide a disease probability prediction system based on human immune age.

[0009] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution: According to a first aspect of the present invention, a method for predicting disease probability based on human immune age is provided, comprising at least the following steps: S1: Preliminary screening and identification of candidate immune biomarkers; S2: Perform statistical analysis on the candidate immune biomarker data of the healthy group sample, and screen out multiple indicators that are significantly correlated with age as immune biomarker groups for modeling. S3: Based on the immune biomarker clusters in the healthy population sample data, an immune age prediction model is established through multiple linear regression analysis; S4: Based on the immune age prediction model, calculate the immune age of each sample in the healthy group and the diseased group; S5: Divide the diseased group samples and the healthy group samples into different age groups according to their actual age, use whether or not they have respiratory diseases as the dependent variable and the immune age as the independent variable, and use binary logistic regression analysis to fit and obtain a respiratory disease risk prediction model corresponding to different age groups.

[0010] Preferably, the respiratory disease risk prediction model is used to characterize: Across all age groups, there is a positive correlation between an individual's immune age and their risk of developing respiratory diseases.

[0011] Preferably, the respiratory disease risk prediction model is a binary logistic regression model, in which the regression coefficient of immune age is positive, and the specific coefficient value is set based on different age groups.

[0012] Preferably, the immune age prediction model includes: Immune age = 106.199 - 10.612 × red blood cell count + 0.966 × TNF-α - 0.035 × absolute number of cytotoxic / suppressive T cells CD8 + 0.011 × absolute number of helper / inducing T cells CD4.

[0013] Preferably, the application of the disease probability prediction method in predicting the risk of respiratory diseases includes the following steps: Obtain immune marker detection data for the individual to be tested, wherein the immune markers include at least red blood cell count, TNF-α, absolute CD8+ T cell count, and absolute CD4+ T cell count; The immune biomarker detection data are input into the immune age prediction model to calculate the immune age of the individual to be tested. The age group to which the individual belongs is determined based on the individual's actual age, and the immune age is substituted into the respiratory disease risk prediction model based on the corresponding age group to generate the probability risk value of the individual's respiratory disease. The individual to be tested is risk-stratified based on the probability risk value, and when the probability risk value exceeds a preset threshold, an early warning message is output. The respiratory disease risk prediction model is a binary logistic regression model, in which the regression coefficient of immune age is positive, and the specific coefficient value is set based on different age groups.

[0014] Preferably, in the respiratory disease risk prediction model... For individuals aged 25-40, the risk of respiratory diseases increases by 6.3% for every additional unit of immune age. For individuals aged 41-60, the risk of respiratory diseases increases by 4.7% for every additional unit of immune age. For individuals aged 60 and above, the risk of respiratory diseases increases by 4.5% for every additional unit of immune age.

[0015] Preferably, the application of the disease probability prediction method in the probability prediction of chronic obstructive pulmonary disease includes the following steps: Obtain the red blood cell count, TNF-α level, absolute CD8+ T cell count, and absolute CD4+ T cell count of the individual being tested; The immune age of the individual under test was calculated based on the red blood cell count, TNF-α level, absolute number of CD8+ T cells, and absolute number of CD4+ T cells. The lung function of the individual being tested was calculated based on the formula: Lung function = 1.017 - 0.003 × Immune age. By combining the smoking history index and environmental exposure parameters of the individual under test as correction factors, the calculated value of lung function is weighted and corrected to generate a comprehensive COPD risk score; When the COPD comprehensive risk score exceeds the preset warning threshold, a high-risk warning is triggered and the individual to be tested is advised to undergo a lung function diagnostic test.

[0016] According to a second aspect of the present invention, a disease probability prediction system based on human immune age is provided, comprising a processor and a memory, wherein the processor and the memory are coupled together; wherein the memory is used to store a computer program; and the processor is used to run the computer program stored in the memory to perform the above-described disease probability prediction method based on human immune age.

[0017] Compared with existing technologies, this invention successfully quantifies complex immune states into an intuitive "immune age" indicator by constructing an immune age prediction model, effectively overcoming the high cost and strong data dependence problems of existing multidimensional models. Based on this, this invention pioneers a prediction paradigm that links immune age with organ function decline and the risk of specific diseases, capable of outputting objective probability risk values ​​for individuals of different age groups to develop respiratory diseases. This invention achieves a leap from traditional qualitative assessment to accurate probability prediction, providing a reliable quantitative basis for early, low-cost risk stratification and personalized intervention in health management. Attached Figure Description

[0018] Figure 1 This is an overall flowchart of the disease probability prediction method provided in the first embodiment of the present invention; Figure 2 The histogram of the linear regression model in the disease probability prediction method provided in the first embodiment of the present invention; Figure 3 The PP diagram of the linear regression model in the disease probability prediction method provided in the first embodiment of the present invention; Figure 4 A scatter plot of the linear regression model in the disease probability prediction method provided in the first embodiment of the present invention; Figure 5 The histogram of the linear regression model in the disease probability prediction method provided in the second embodiment of the present invention; Figure 6 The PP diagram of the linear regression model in the disease probability prediction method provided in the second embodiment of the present invention; Figure 7 A scatter plot of the linear regression model in the disease probability prediction method provided in the second embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of the disease probability prediction system provided in the fourth embodiment of the present invention. Detailed Implementation

[0019] The technical content of the present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0020] The technical concept of this invention lies in: quantifying "immune age" and comparing it with actual age to accurately reflect the aging degree of an individual's immune system; based on this, an immune age prediction model is constructed using multidimensional cell analysis, thereby achieving probabilistic risk prediction for infections, tumors, and other chronic diseases—for example, the constructed respiratory disease risk prediction model is a specific application capable of outputting quantitative probabilistic risk values. Those skilled in the art will understand that this probabilistic prediction framework based on immune age is universal and can be further extended to risk probability assessment for other types of diseases, such as liver diseases, thus enabling a leap from qualitative judgment to probabilistic early warning, and providing a reliable basis for personalized anti-aging and disease intervention.

[0021] First Embodiment like Figure 1 As shown, the disease probability prediction method based on human immune age provided in the first embodiment of the present invention includes at least the following steps: S1: Preliminary screening and identification of immunomarker groups This is the preparatory work for modeling. Based on preliminary research or scientific literature, a set of indicators theoretically related to immune age is pre-selected from numerous possible immune and blood indicators as a "candidate immune biomarker group," which specifically includes the following sub-steps: S11: Based on preliminary research, red blood cell count, TNF-α (tumor necrosis factor-α), absolute number of cytotoxic / suppressive T cells CD8, absolute number of helper / inducing T cells CD4, total number of T lymphocytes (total absolute number of T cells), NK cell count (absolute number of natural killer cells), monocyte CD14CD68CD11b, absolute number of neutrophils D15CD66bCD11b, serum albumin, and hemoglobin concentration HGB were initially screened as candidate immune markers.

[0022] S12: Perform statistical analysis on the candidate immune biomarker data of multiple samples, and select multiple indicators with significant correlation (significance < 0.05) as immune biomarker groups.

[0023] We collected data on candidate immune biomarkers from 400 blood samples for statistical analysis. The data for these candidate immune biomarkers were obtained from tests ordered by doctors and provided by the hospital; details are omitted here.

[0024] For the candidate immune biomarker population data of the sample, correlation statistical analysis (e.g., linear regression analysis) was performed using SPSS version 26 (IBM SPSS Statistics 26) statistical analysis software to screen out indicators with significant correlation (significance < 0.05), as shown in the table below.

[0025] Table 1 The table above shows the standard results report automatically calculated and output by SPSS 26 software after performing linear regression analysis and using the Enter method to include all selected independent variables (candidate immune biomarkers) into the model.

[0026] As shown in Table 1, the significance of erythrocyte count, TNF-α, absolute number of cytotoxic / suppressive T cells (CD8), and absolute number of helper / inducing T cells (CD4) was <0.05. The significance of total absolute T cells, NK cells, absolute B cells, monocytes (CD14CD68CD11b), absolute number of neutrophils (CD15CD66bCD11b), serum albumin, and hemoglobin concentration (HGB) was >0.05. This means that erythrocyte count, TNF-α, absolute number of cytotoxic / suppressive T cells (CD8), and absolute number of helper / inducing T cells (CD4) can significantly affect age, while total absolute T cells, NK cells, absolute number of B cells, monocytes (CD14CD68CD11b), absolute number of neutrophils (CD15CD66bCD11b), serum albumin, and hemoglobin concentration (HGB) cannot be considered to affect age.

[0027] Therefore, indicators with significant correlation (significance < 0.05) can be selected from Table 1, which are the indicators used for modeling. These include: Red blood cell count; TNF-α (tumor necrosis factor-α); Absolute number of CD8+ T cells (cytotoxic / suppressive T cells); Absolute number of CD4+ T cells (helper / inducer T cells); These biomarkers constitute a "group of immune biomarkers" used for modeling.

[0028] S2: Collect blood samples from multiple (e.g., 165 healthy individuals aged 18-79) for comparative analysis.

[0029] The statistical results of the sample data obtained from 165 healthy individuals are shown in Table 2 below.

[0030] Table 2 As shown in Table 2, all 165 sample data points were valid, and the immune biomarkers exhibited sufficient variability across the population, thus supporting their inclusion in the immune age prediction model. Specifically, the four indicators (red blood cells, TNF-α, CD8+, and CD4+) presented in Table 2 all showed considerable variability (as evidenced by the standard deviation, minimum, and maximum values). The mean red blood cell count was 4.6146 (x 10^6). 12The mean red blood cell count (pg / mL) was 2.4685, with a small standard deviation (0.472), indicating relatively stable inter-individual differences in red blood cell count among healthy individuals. The mean TNF-α was 2.4685 pg / mL, but the standard deviation (2.986) was greater than the mean, ranging from 0.00 to 18.45. This suggests that TNF-α levels are highly unevenly distributed in the population, with some individuals exhibiting high inflammation levels while most have low levels. This high variability makes it a discriminative predictor. The mean absolute CD8+ T cell count was 545.51 cells / μL, ranging from 156.14 to 1023.66, exhibiting significant variability (standard deviation 193.03), reflecting substantial differences in CD8+ T cell counts across different individual immune states. The mean absolute CD4+ T cell count was 852.70 cells / μL, ranging from 327.06 to 1924.99, with a standard deviation of 282.70, exhibiting the greatest variability. This highlights the central role of CD4+ T cells in the immune system and the high degree of individualization in their numbers.

[0031] It is precisely these inter-individual differences that allow them to be used to capture and predict age. Combining Tables 1 and 2, TNF-α levels significantly increase with age (Beta = 0.220), consistent with the high variability observed in Table 2, reflecting age-related "inflammatory senescence." CD8+ T cells decrease with age (Beta = -0.294), while CD4+ T cells are positively correlated with age (Beta = 0.294). Red blood cell counts decrease slightly with age (Beta = -0.157), consistent with trends observed in some studies regarding changes in hematopoietic function or erythropoietin levels in older adults. In other words, in healthy individuals, with increasing chronological age (calculated based on birth date), individual inflammation levels (TNF-α) and CD4+ T cell counts tend to increase, while red blood cell and CD8+ T cell counts tend to decrease. These four indicators together constitute a "group of immune biomarkers" that can effectively predict an individual's immune age.

[0032] S3: Based on the immune biomarker clusters in the sample data of healthy people, a multiple linear regression model is established using correlation statistical analysis software (such as SPSS software) to obtain an immune age prediction model for healthy people.

[0033] The data obtained in S2 were fitted using a multiple linear regression model, resulting in the following regression equation, which serves as a model for predicting the immune age of healthy individuals: Immune age = 106.199 - 10.612 × red blood cell count + 0.966 × TNF-α - 0.035 × absolute number of cytotoxic / suppressive T cells CD8 + 0.011 × absolute number of helper / inducing T cells CD4.

[0034] The model summary of the immune age prediction model for healthy individuals is shown in Table 3 below.

[0035] Table 3 Table 3 shows a correlation coefficient R of 0.495, indicating a moderate positive correlation between the independent and dependent variables (immune age) (the closer the R value is to 1, the stronger the linear relationship); the R-squared (coefficient of determination) is 0.245, demonstrating a significant and quantifiable link between the immune system state and the aging process. This means that in this model, approximately 24.5% of the differences in immune age among different individuals are determined by changes in four indicators: red blood cells, TNF-α, CD8, and CD4. These independent variables can explain 24.5% of the degree of age variability (more than 10% is considered good in medicine and is within an acceptable range). This confirms that the immune system can serve as a powerful biomarker for the biological age of aging. The standard estimation error was 16.609, indicating that individuals with the same immune index values ​​had a large difference in their chronological age (average ±16.6 years). This precisely reflects the asynchronous nature of "immune age" and "chronological age", indicating that the model has a certain discriminative ability and can identify those individuals whose immune status deviates significantly from their age norm. The extremely low DW value strongly suggests the presence of systematic biological signals in the residuals that were not captured by the model. Table 4 below is an ANOVA (analysis of variance) table of the immune age prediction model of this invention.

[0036] Table 4 The F-value in the table is the ratio of "between-group variation" to "within-group variation". The larger the F-value, the more the model explains the variation that is much greater than the unexplained variation (error), indicating that the model is more effective. The F-value of the model in this invention is 13.01, which is a fairly large value, and the significance p-value is extremely small. This shows that the linear regression model for predicting age (as a proxy for immune age) is statistically highly significant. In other words, this table proves that the linear relationship between these four immune indicators (red blood cells / RDW-SD, TNF-α, CD8, CD4) and immune age is not caused by randomness, but is real.

[0037] Table 5 below is used to further examine whether the independent contribution of each independent variable (immune indicator) is significant; Table 5 The significance levels of erythrocyte count, TNF-α count, absolute cytotoxic / suppressive T cell count (CD8), and absolute helper / inducer T cell count (CD4) were <0.001, <0.05, <0.001, and <0.05, respectively, indicating that these parameters significantly affect age. The unstandardized coefficient B for erythrocyte count was -10.612, indicating a significant negative impact of erythrocyte count on age; that is, lower erythrocyte counts correlate with older age, and vice versa. The unstandardized coefficient B for NF-α was 0.966, significantly positively affecting age; that is, the higher the TNF-α level, the older the age, and vice versa. The unstandardized coefficient B for the absolute number of cytotoxic / suppressive T cells (CD8) was -0.035, significantly negatively affecting age; that is, the lower the absolute number of cytotoxic / suppressive T cells (CD8), the older the age, and vice versa. The unstandardized coefficient B for the absolute number of helper / inducing T cells (CD4) was 0.966, significantly positively affecting age; that is, the higher the absolute number of helper / inducing T cells (CD4), the older the age, and vice versa.

[0038] Therefore, the immune age prediction model satisfies the variance inflation factor (VIF) < 5 (see Table 5), which means that there is no multicollinearity among the variables (in the medical field, VIF < 10 is considered to be the absence of multicollinearity among the variables).

[0039] To further illustrate the effectiveness of the immune age prediction model of the present invention, Figure 2 The histogram of the linear regression model shown demonstrates that the model error is random and there is no systematic bias. Figure 3 The points in the normal PP plot shown are very close to the middle diagonal, which is consistent with the histogram, further illustrating that the residuals satisfy the normality. Figure 4 The distribution of points in the scatter plot of the linear regression model shown does not have an obvious pattern, indicating that the model successfully captures the linear trend in the data and does not miss any obvious nonlinear relationships. The prediction error of the model is relatively stable across different age groups (young, middle-aged, and elderly), and there will be no situation where the prediction is very accurate for young people but very inaccurate for the elderly.

[0040] Second Embodiment In the second embodiment of the present invention, a lung function prediction model for healthy individuals is constructed based on the immune age prediction model in the first embodiment. Here, the forced expiratory volume in one second (FEV1) is used as the gold standard for evaluating lung function, and hospital staff can collect FEV1 data using a pulmonary function analyzer. Again, based on the lung function (FEV1) data of 165 healthy individuals from the first embodiment, the data is input into SPSS software; the immune ages of these 165 healthy individuals, calculated in the first embodiment, are also input into SPSS software; a linear regression model is used for fitting, resulting in the lung function prediction model for healthy individuals: Lung function = 1.017 - 0.003 × Immune age.

[0041] Table 6 shows the statistical results obtained by inputting immune age and FEV1 into SPSS software. The mean immune age was 49.90 years, very close to the mean chronological age (49.92 years, from the previous table), but the standard deviation (9.300) was much smaller than the standard deviation of chronological age (18.886). This indicates that the immune age model successfully "categorized" and "smoothed" biologically similar individuals, capturing the rate of biological aging rather than simply physiological (time) age. The mean lung function (FEV1) was 0.8759 L, ranging from 0.81 L to 0.97 L, with a very small standard deviation (0.03466). This indicates that the lung function of the study sample was within a relatively normal and narrow range, meeting the screening criteria for "healthy individuals".

[0042] Table 6 Table 7 Table 7 is a model summary of the lung function prediction model for healthy individuals (generated by SPSS software). The R-squared value is 0.573, which means that the independent variable, immune age, can explain 57.3% of the variability in lung function (medically, more than 10% is considered good and within an acceptable range).

[0043] Table 8 Table 9 Table 8 shows that the significance of immunological age is less than 0.001, meaning that immunological age can significantly affect lung function. Table 9 shows that the unstandardized coefficient B of immunological age is -0.003, so immunological age significantly and negatively affects lung function; that is, the lower the immunological age, the higher the lung function, and the higher the immunological age, the lower the lung function. Furthermore, the VIF in Table 9 is 1, indicating that there is no multicollinearity among the variables (in the medical field, VIF < 10 is considered to indicate that there is no multicollinearity among the variables).

[0044] Third Embodiment The third embodiment of the present invention provides a respiratory disease risk prediction model based on immune age, including the following steps.

[0045] refer to Figure 1 From the aforementioned 165 healthy individuals, 138 were selected as the healthy group based on their actual age (25–76 years), and another 123 patients with respiratory diseases were collected as the disease group. Sample data were obtained from both groups. The sample data included at least immune markers (red blood cell count, TNF-α, absolute CD8+ T cell count, and absolute CD4+ T cell count). Based on the immune age prediction model in the first embodiment, the immune age corresponding to each sample is calculated. The healthy group and the disease group were divided into three age groups according to their actual age (detailed below): 25-40 years old, 41-60 years old, and 60 years old and above. (1) For the sample aged 25 to 40 (a total of 61 people, including 42 healthy people and 19 people with diseases), the immune age and whether or not the sample had respiratory diseases (the value of the diseased group was 1; the value of the healthy group was 0) of each sample data were entered into SPSS software, and a binary logistic regression model was used for regression analysis to fit the respiratory disease risk prediction model based on immune age for this age group.

[0046] Table 10 Exp(B), or OR value, is the effect size of a binary logistic regression. The significance of immune age is 0.039 < 0.05, indicating that immune age is an independent influencing factor for respiratory diseases and can significantly affect respiratory diseases. The Exp(B) value, or OR value, is 1.063, which means that the probability of developing respiratory diseases is 1.063 times that of developing respiratory diseases when immune age is one unit older than that when it is one unit younger. In other words, every unit increase in immune age leads to a 6.3% increase in the risk of developing respiratory diseases.

[0047] Table 11 For the immunological age-based respiratory disease risk prediction model, the null hypothesis of the Hosmer-Lemersho test is: the observed data and the regression model fit well; the alternative hypothesis is: the observed data and the regression model do not fit well. The table above shows that the significance of the Hosmer-Lemersho test is P = 0.750 > 0.05, therefore the null hypothesis is accepted. This indicates that the binary logistic regression model established for this data fits the actual data well. That is, the results given by the binary logistic regression model can reliably reflect the true relationship between the original variables. Tables 10 and 11 together demonstrate that the immunological age-based respiratory disease risk prediction model for the 25–40 age group is effective and reliable.

[0048] (2) 41-60 years old (80 people in total, including 35 healthy people and 45 people with diseases): For the samples in the 41-60 age group, the immune age and whether or not they have respiratory diseases (the value of the disease group is 1; the value of the healthy group is 0) of each sample data were entered into SPSS software, and a binary logistic regression model was used for regression analysis to fit the respiratory disease risk prediction model based on immune age for this age group.

[0049] Table 12 As shown in Table 12, the significance of immune age is 0.043 < 0.05, indicating that immune age is an independent influencing factor for respiratory diseases and can significantly affect respiratory diseases. The Exp(B) or OR value is 1.047, which means that the probability of developing respiratory diseases is 1.047 times that of developing respiratory diseases when the immune age is one unit older than that when it is one unit younger. In other words, every unit increase in immune age leads to a 4.7% increase in the risk of developing respiratory diseases.

[0050] As shown in the table below, the reliability of the aforementioned model was verified using the Hosmer-Lemersho test. In the table, the significance P = 0.569 > 0.05, therefore the null hypothesis (good fit between the observed data and the regression model) is accepted. This indicates that the binary logistic regression model established for this data fits the actual data well, meaning that the results given by the binary logistic regression model can reliably reflect the true relationship between the original variables.

[0051] Table 13 (3) 60 years and older (120 people in total, including 61 healthy people and 59 people with diseases): For the sample of 60 years and older, the immune age (as independent variable) and whether or not the sample has respiratory diseases (the value of the diseased group is 1; the value of the healthy group is 0) (as dependent variable) of each sample data were entered into SPSS software, and a binary logistic regression model was used for regression analysis to fit the respiratory disease risk prediction model based on immune age for this age group.

[0052] Table 14 In Table 14, the significance of immune age is 0.017 < 0.05, indicating that immune age is an independent influencing factor for respiratory diseases and can significantly affect respiratory diseases. The Exp(B) or OR value is 1.045, which means that the probability of developing respiratory diseases is 1.045 times that of developing respiratory diseases when the immune age is one unit older than that when it is one unit younger. In other words, every unit increase in immune age leads to a 4.5% increase in the risk of developing respiratory diseases.

[0053] As shown in the table below, the reliability of the aforementioned model was verified using the Hosmer-Lemersho test. In the table, the significance P = 0.101 > 0.05, therefore the null hypothesis (good fit between the observed data and the regression model) is accepted. This indicates that the binary logistic regression model established for this data fits the actual data well, meaning that the results given by the binary logistic regression model can reliably reflect the true relationship between the original variables.

[0054] Table 15 Therefore, in this embodiment, immune age (independent variable) is used to predict the risk of disease (dependent variable). This model based on immune age to predict the risk of respiratory diseases has multiple practical applications in medicine and health management: 1. Accurately identify high-risk groups and optimize screening strategies. For example, for people over 60 years old, priority can be given to strengthening targeted screenings such as lung function and chest CT for individuals with advanced immune age (even if they are asymptomatic) to improve the early detection rate and reduce medical costs. 2. Early warning and intervention timing assessment, for example, as a window of opportunity, when an elderly person's immune age is found to be significantly beyond their actual age, it can serve as an early warning signal, indicating that their risk of respiratory diseases is accumulating, and it is recommended to initiate lifestyle interventions (nutrition, exercise, vaccination) or immunomodulatory therapy, rather than waiting for the onset of the disease; 3. Personalized health management and efficacy monitoring: Using immune age as a baseline indicator for health management, develop individualized plans and assess whether interventions (such as influenza vaccines and immune-enhancing therapies) are effective in delaying immune aging and reducing actual disease risk.

[0055] Finally, to verify the effectiveness of the present invention, the inventors, based on the processing method in prior patent document CN110957010B, used the data (total T cell percentage, cytotoxic / suppressive T cell percentage, helper / inducer T cell percentage, natural killer (NK) cell percentage, B cell percentage, and monocyte CD14CD68CD11b) and actual age of the aforementioned 165 samples, and used the same SPSS software to obtain the model-related statistical data in CN110957010B (see Tables 16 to 18), as comparative example 1 of the model effect.

[0056] The statistical results of the model in Comparative Example 1 (compared with CN110957010B) are shown in the following three tables.

[0057] Table 16 This table displays the specific parameters of the model, listing the magnitude, direction, and statistical significance of the influence of each independent variable on the dependent variable. In the table above, the VIF (Variable Interest Scale) for the total T cell percentage, natural killer (NK) cell percentage, and B cell percentage is greater than 10, indicating severe multicollinearity and insufficient model robustness.

[0058] Table 17 The table above is used to test whether the entire regression model is statistically significant. The R-squared value is 0.082, meaning that the independent variables—monocyte CD14CD68CD11b, percentage of cytotoxic / suppressive T cells, percentage of B cells, percentage of helper / inducer T cells, percentage of total T cells, and percentage of natural killer (NK) cells—can explain 8.2% of the age variability (medically, above 10% is better, within an acceptable range, but with weak explanatory power). However, 8.2% in Comparative Example 1 is < 24.5% (as shown in Table 3 of the first embodiment of this invention), therefore, this table indicates that Comparative Example 1 is less effective than the immune age prediction model in the first embodiment of this invention.

[0059] Table 18 Based on the significance level of 0.046 (less than 0.05) in the table above, the percentages of cytotoxic / suppressive T cells, helper / inducer T cells, natural killer (NK) cells, B cells, and monocytes CD14, CD68, and CD11b could not significantly affect age.

[0060] Comparative conclusion: In Comparative Example 1, the significance, collinearity (VIF), R-squared, and F-value are all lower than the values ​​in the first embodiment of the present invention, indicating that the immune age prediction model of the first embodiment of the present invention is statistically superior.

[0061] To compare with the technical solution in another prior patent document CN113488173B, the inventors also provide Comparative Example 2. Since the immune age prediction models for men and women in CN113488173B are different, Comparative Example 2 includes Comparative Example 2-1 (male) and Comparative Example 2-2 (female).

[0062] Comparative Example 2-1: Sample data of 165 males from the first embodiment of the present invention were obtained: hematocrit, natural killer (NK) cell percentage, HGB, total protein, albumin / globulin ratio, red blood cell count, serum albumin, and actual age. These sample data were input into SPSS software and processed using the same method as CN113488173B. The statistical results of the model in Comparative Example 2-1 are shown in the following three tables.

[0063] Table 19 Table 20 Table 21 Based on the above three tables, the significance of serum albumin, total protein, albumin / globulin ratio, erythrocytes, HGB, and hematocrit is all >0.05; the VIF of serum albumin, total protein, and albumin / globulin ratio is >10, indicating multicollinearity. Therefore, the model in Comparative Example 2-1 is not as effective as the model in the first embodiment of this invention.

[0064] Comparative Example 2-2: Sample data of 165 women from the first embodiment of the present invention were obtained: erythrocyte distribution width - SD standard deviation, globulin, natural killer (NK) cell percentage, hemoglobin concentration (HGB), serum albumin, erythrocyte count, and chronological age. These sample data were input into SPSS software and processed using the same method as CN113488173B. The statistical results of the model in Comparative Example 2-2 are shown in the following three tables.

[0065] Table 22 Table 23 Table 24 Based on the three tables above, it can be seen that the significance of globulin, erythrocyte, hemoglobin concentration (HGB), and erythrocyte distribution width - SD standard deviation in the female sample data of Comparative Example 2-2 is greater than 0.05, indicating that they do not significantly affect immune age. The F-value of Comparative Example 2-2 (9.836) is lower than the F-value (13.010) in the embodiments of the present invention. Therefore, Comparative Example 2-2 is not as good as the model in the embodiments of the present invention. Fourth embodiment Building upon the disease (e.g., respiratory disease) prediction method based on immune age described in the third embodiment above, the fourth embodiment of this invention further provides a disease probability prediction system based on immune age. This system includes one or more processors and a memory. The memory is coupled to the processor and is used to store one or more programs. When the programs are executed by the processor, the processor enables the processor to implement the disease probability prediction method based on human immune age as described in the above embodiments. The processor controls the overall operation of the system to complete all or part of the steps of the disease probability prediction method based on human immune age. The memory stores various types of data to support the operation of the system.

[0066] Fifth embodiment The fifth embodiment of the present invention provides an application of a disease probability prediction method based on human immune age in the prediction of the risk of respiratory diseases, which includes at least the following steps: Obtain immune marker detection data for the individual to be tested, wherein the immune markers include at least red blood cell count, TNF-α, absolute CD8+ T cell count, and absolute CD4+ T cell count; The immune biomarker detection data are input into the immune age prediction model to calculate the immune age of the individual to be tested. The age group to which the individual belongs is determined based on the individual's actual age, and the immune age is substituted into the respiratory disease risk prediction model based on the corresponding age group to generate the probability risk value of the individual's respiratory disease. The individual to be tested is risk-stratified based on the probability risk value, and when the probability risk value exceeds a preset threshold, an early warning message is output.

[0067] Sixth Embodiment The sixth embodiment of the present invention provides an application of a disease probability prediction method based on human immune age in the probability prediction of chronic obstructive pulmonary disease, comprising at least the following steps: Obtain the red blood cell count, TNF-α level, absolute CD8+ T cell count, and absolute CD4+ T cell count of the individual being tested; The immune age of the individual under test was calculated based on the red blood cell count, TNF-α level, absolute number of CD8+ T cells, and absolute number of CD4+ T cells. The lung function of the individual being tested was calculated based on the formula: Lung function = 1.017 - 0.003 × Immune age. By combining the smoking history index and environmental exposure parameters of the individual under test as correction factors, the calculated value of lung function is weighted and corrected to generate a comprehensive COPD risk score; When the COPD comprehensive risk score exceeds the preset warning threshold, a high-risk warning is triggered and the individual to be tested is advised to undergo a lung function diagnostic test.

[0068] It should be noted that the above embodiments are merely illustrative examples. The technical solutions of the various embodiments can be combined, and the order of the steps can be changed, all of which are within the protection scope of this invention.

[0069] The above provides a detailed description of the disease prediction method and system based on human immune age provided by this invention. Any obvious modifications made by those skilled in the art without departing from the essence of this invention will constitute an infringement of the patent rights of this invention and will incur corresponding legal liability.

Claims

1. A method for predicting disease probability based on human immune age, characterized in that... Includes the following steps: S1: Preliminary screening and identification of candidate immune biomarkers; S2: Perform statistical analysis on the candidate immune biomarker data of the healthy group sample, and screen out multiple indicators that are significantly correlated with age as immune biomarker groups for modeling. S3: Based on the immune biomarker clusters in the healthy population sample data, an immune age prediction model is established through multiple linear regression analysis; S4: Based on the immune age prediction model, calculate the immune age of each sample in the healthy group and the diseased group; S5: Divide the diseased group samples and the healthy group samples into different age groups according to their actual age, use whether or not they have respiratory diseases as the dependent variable and the immune age as the independent variable, and use binary logistic regression analysis to fit and obtain a respiratory disease risk prediction model corresponding to different age groups.

2. The disease probability prediction method based on human immune age as described in claim 1, characterized in that... The respiratory disease risk prediction model is used to characterize: Across all age groups, there is a positive correlation between an individual's immune age and their risk of developing respiratory diseases.

3. The disease probability prediction method based on human immune age as described in claim 2, characterized in that: The respiratory disease risk prediction model is a binary logistic regression model, in which the regression coefficient of immune age is positive, and the specific coefficient value is set based on different age groups.

4. The disease probability prediction method based on human immune age as described in claim 1, characterized in that... The immune age prediction model includes: Immune age = 106.199 - 10.612 × red blood cell count + 0.966 × TNF-α - 0.035 × absolute number of cytotoxic / suppressive T cells CD8 + 0.011 × absolute number of helper / inducing T cells CD4.

5. A disease probability prediction system based on human immune age, characterized in that... The device includes a processor and a memory, the processor and the memory being coupled together; wherein the memory is used to store a computer program; the processor is used to run the computer program stored in the memory, performing the disease probability prediction method based on human immune age as described in any one of claims 1 to 4.

6. A method for predicting lung function based on human immune age, characterized in that... Includes the following steps: Obtain the immune age and corresponding lung function test data of healthy group samples; Using the immune age as the independent variable and the lung function data as the dependent variable, a lung function prediction model was obtained through linear regression fitting: Lung function = 1.017 - 0.003 × immune age.

7. The lung function prediction method based on human immune age as described in claim 6, characterized in that: The lung function mentioned refers to the forced expiratory volume in one second.

8. The application of the disease probability prediction method based on human immune age as described in any one of claims 1 to 4 in the prediction of the risk of respiratory diseases, characterized in that... Includes the following steps: Obtain immune marker detection data for the individual to be tested, wherein the immune markers include at least red blood cell count, TNF-α, absolute CD8+ T cell count, and absolute CD4+ T cell count; The immune biomarker detection data are input into the immune age prediction model to calculate the immune age of the individual to be tested. The age group to which the individual belongs is determined based on the individual's actual age, and the immune age is substituted into the respiratory disease risk prediction model based on the corresponding age group to generate the probability risk value of the individual's respiratory disease. The individual to be tested is risk-stratified based on the probability risk value, and when the probability risk value exceeds a preset threshold, an early warning message is output. The respiratory disease risk prediction model is a binary logistic regression model, in which the regression coefficient of immune age is positive, and the specific coefficient value is set based on different age groups.

9. The lung function prediction method based on human immune age as described in claim 8, characterized in that... In the respiratory disease risk prediction model, For individuals aged 25-40, the risk of respiratory diseases increases by 6.3% for every additional unit of immune age. For individuals aged 41-60, the risk of respiratory diseases increases by 4.7% for every additional unit of immune age. For individuals aged 60 and above, the risk of respiratory diseases increases by 4.5% for every additional unit of immune age.

10. The application of the disease probability prediction method based on human immune age as described in any one of claims 1 to 4 in the probability prediction of chronic obstructive pulmonary disease, characterized in that... Includes the following steps: Obtain the red blood cell count, TNF-α level, absolute CD8+ T cell count, and absolute CD4+ T cell count of the individual being tested; The immune age of the individual under test was calculated based on the red blood cell count, TNF-α level, absolute number of CD8+ T cells, and absolute number of CD4+ T cells. The lung function of the individual being tested was calculated based on the formula: Lung function = 1.017 - 0.003 × Immune age. By combining the smoking history index and environmental exposure parameters of the individual under test as correction factors, the calculated value of lung function is weighted and corrected to generate a comprehensive COPD risk score; When the COPD comprehensive risk score exceeds the preset warning threshold, a high-risk warning is triggered and the individual to be tested is advised to undergo a lung function diagnostic test.

Images