A disease risk assessment system based on physiological data

Abstract

A disease risk assessment system based on physiological data can accurately, stably and explainably identify high-risk groups of diseases in complex population scenarios and output risk stratification. It comprises: a data acquisition module that acquires physiological data and constructs a physiological index-organ system correlation matrix; a disease risk calculation module that estimates the risk contribution of each physiological index to the future disease risk of the target disease and outputs the corresponding risk contribution parameters; a feature recognition module that obtains stable weights for subsequent organ system comprehensive representation construction; a weighted feature generation module that constructs organ system comprehensive physiological indicators; a collaborative risk calculation module that establishes a risk model between the comprehensive physiological indicators and the future disease risk of the target disease, and determines the stratification threshold according to the self-adaptive threshold determination rule when there is a nonlinear trend; a risk assessment module that outputs: relative importance ranking, multi-organ collaborative combination list and its synergy strength, and joint risk score for individual samples.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a disease risk assessment system based on physiological data, mainly used to assess disease risk and identify high-risk groups. Background Technology

[0002] With the accelerating aging of the population and the increasing burden of chronic diseases, the incidence of brain diseases and various chronic physical illnesses is on the rise. The functional status of various organ systems in the human body is closely related to the occurrence and development of diseases, and some organ system-related physiological indicators may show abnormal fluctuations before clinical diagnosis. Given that clinical health checkups can collect multidimensional physiological data reflecting the functional status of different organ systems at an early stage before the onset of disease, how to organize and quantify the physiological indicators of health checkups from the perspective of organ systems to achieve stable assessment of future disease risks and identify high-risk groups has significant clinical and public health application value.

[0003] In existing technologies, disease risk assessment schemes typically employ traditional regression models or machine learning models, using physical examination indicators as input variables to construct risk prediction models. However, existing schemes generally suffer from the following shortcomings:

[0004] (1) Insufficient information utilization. Existing studies or systems often use a few single physiological indicators or single modality data as risk factors for analysis and prediction, or directly splice multiple indicators into the model. Since the health status of the same organ system is usually characterized by multiple indicators, single indicator modeling is difficult to reflect the overall health level of the organ; while simple splicing easily introduces redundancy and noise, affecting the model performance and reliability.

[0005] (2) Lack of structured representation at the organ system level. When multiple indicators reflect the same organ system, existing technologies often lack a general mechanism for organizing, merging and comprehensively representing indicators by organ system; there is also a lack of consistent processing methods for indicators related to multiple organs, resulting in unclear model input structure, insufficient interpretability, and difficulty in outputting organ-level risk contribution information that can be used for risk stratification and management.

[0006] (3) Insufficient stability and transferability. The population undergoing physical examinations varies in terms of age, gender, race, education level, and lifestyle, and these factors may simultaneously affect physiological indicators and disease outcomes. Existing protocols have different strategies in terms of controlling confounding factors, sample imbalance, and handling outliers and missing data, which can easily lead to fluctuations in model results under different sampling or population conditions, affecting the repeatability and cross-scenario applicability of risk assessment and increasing the difficulty of practical implementation.

[0007] (4) Insufficient characterization of coexisting and interactive / synergistic risks among multiple organs. In real-world populations, abnormalities in multiple organ systems often coexist, and the degree of impact of different organ abnormalities on disease occurrence may vary. Existing technologies often focus on the independent effects of a single organ or a few indicators, or roughly summarize the number of abnormal organs / abnormal indicators, making it difficult to achieve a refined characterization of multiple organ states and stratified identification of high-risk populations. At the same time, when multiple organ abnormalities coexist, there may be interactive relationships between different organ systems, making the risk effect of joint exposure not a simple sum of individual effects. Existing systems mostly focus on independent effects, or only conduct interaction analysis on a multiplicative scale, making it difficult to conduct stable and interpretable quantitative assessment of excess risk on an additive scale; and lacking stability control and verification mechanisms for interaction assessment results, it is difficult to reliably identify organ combinations with significant interactive / synergistic risks and corresponding high-risk populations.

[0008] In summary, existing technologies still have shortcomings in areas such as the structured organization of multidimensional physical examination physiological data, comprehensive characterization at the organ system level, model stability control, and interactive / synergistic risk assessment under multi-organ coexistence conditions. These limitations make it difficult to meet the application requirements for accurate, stable, and interpretable identification of high-risk individuals in complex population scenarios. Therefore, there is an urgent need for a technical solution that can fully utilize multidimensional physical examination physiological data, form stable risk characterization at the organ system level, and further assess the risks of multi-organ joint exposure and its interactions / synergies, thereby achieving the identification of high-risk individuals and risk stratification output. Summary of the Invention

[0009] To overcome the shortcomings of existing technologies, the technical problem to be solved by the present invention is to provide a disease risk assessment system based on physiological data. This system can make full use of multidimensional physical examination physiological data to form a stable risk characterization at the organ system level, and further assess the risks of multi-organ joint exposure and its interaction / synergy. Thus, it can accurately, stably and interpretably identify high-risk groups of diseases in complex population scenarios and output risk stratification.

[0010] The technical solution of this invention is: a disease risk assessment system based on physiological data, comprising: The data acquisition module acquires physiological data representing the health level of various organ systems in the body extracted from clinical health checkups. Based on the physiological functions represented by each physiological data, it constructs a physiological indicator-organ system correlation matrix to represent the correlation strength between each physiological indicator and at least one organ system. The disease incidence risk calculation module extracts a set of physiological indicators belonging to each organ system based on the correlation matrix, establishes a regularized proportional risk model for predicting the future incidence risk of the target disease, estimates the risk contribution of each physiological indicator to the future incidence risk of the target disease while controlling for the influence of other physiological indicators within the same organ system, and outputs the corresponding risk contribution parameters. The feature recognition module divides the samples into candidate groups with different exposure levels for each physiological indicator, and adopts a confounding factor control strategy to achieve a balanced distribution of different groups across covariates such as age, gender, race, and education level. Subsequently, under multiple repeated sampling / resampling conditions, the risk contribution parameter of the physiological indicator is repeatedly estimated to obtain a stability measure of the risk contribution parameter. The stability measure is used as a screening constraint to retain only physiological indicators with consistent risk contribution direction and stability meeting a preset threshold under repeated sampling conditions, thus obtaining stable weights for subsequent comprehensive organ system characterization construction. The weighted feature generation module constructs a comprehensive physiological index for each organ system based on the selected physiological indicators, their weight coefficients, and stability within that organ system. This comprehensive physiological index is a weighted combination of the values ​​of each physiological indicator, their weight coefficients, and stability, to reduce the impact of sample perturbation on the comprehensive physiological index and improve cross-population applicability. It outputs comprehensive physiological indices for multiple organ systems as input for subsequent collaborative risk assessment and high-risk population identification. The collaborative risk calculation module establishes a risk model between the comprehensive physiological indicators of each organ system and the future risk of the target disease, and determines the stratification threshold according to the adaptive threshold determination rule when there is a nonlinear trend. The adaptive threshold determination rule determines the threshold based on the comprehensive optimal criterion of risk discrimination and stability measurement in the candidate threshold set, thereby dividing the samples into different risk levels. The risk assessment module, based on the comprehensive physiological indicators of the organ system and the synergistic combination of multiple organs, outputs: relative importance ranking, a list of synergistic combinations of multiple organs and their synergistic strength, and a joint risk score for individual samples.

[0011] This invention achieves full utilization of multidimensional physical examination physiological data to form a stable risk characterization at the organ system level through the collaboration of a data acquisition module, a disease risk calculation module, a feature recognition module, a weighted feature generation module, a collaborative risk calculation module, and a risk assessment module. It further assesses the risks of multi-organ joint exposure and its interaction / synergy, thereby enabling accurate, stable, and interpretable identification of high-risk groups for diseases in complex population scenarios and outputting risk stratification. Attached Figure Description

[0012] Figure 1This is a schematic diagram of the working principle of the disease risk assessment system based on physiological data according to the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] It should be noted that the term "comprising" and any variations thereof in the specification, claims and accompanying drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products or devices.

[0015] like Figure 1 As shown, this disease risk assessment system based on physiological data includes: The data acquisition module acquires physiological data representing the health level of various organ systems in the body extracted from clinical health checkups. Based on the physiological functions represented by each physiological data, it constructs a physiological indicator-organ system correlation matrix to represent the correlation strength between each physiological indicator and at least one organ system. The disease incidence risk calculation module extracts a set of physiological indicators belonging to each organ system based on the correlation matrix, establishes a regularized proportional risk model for predicting the future incidence risk of the target disease, estimates the risk contribution of each physiological indicator to the future incidence risk of the target disease while controlling for the influence of other physiological indicators within the same organ system, and outputs the corresponding risk contribution parameters. The feature recognition module divides the samples into candidate groups with different exposure levels for each physiological indicator, and adopts a confounding factor control strategy to achieve a balanced distribution of different groups across covariates such as age, gender, race, and education level. Subsequently, under multiple repeated sampling / resampling conditions, the risk contribution parameter of the physiological indicator is repeatedly estimated to obtain a stability measure of the risk contribution parameter. The stability measure is used as a screening constraint to retain only physiological indicators with consistent risk contribution direction and stability meeting a preset threshold under repeated sampling conditions, thus obtaining stable weights for subsequent comprehensive organ system characterization construction. The weighted feature generation module constructs a comprehensive physiological index for each organ system based on the selected physiological indicators, their weight coefficients, and stability within that organ system. This comprehensive physiological index is a weighted combination of the values ​​of each physiological indicator, their weight coefficients, and stability, to reduce the impact of sample perturbation on the comprehensive physiological index and improve cross-population applicability. It outputs comprehensive physiological indices for multiple organ systems as input for subsequent collaborative risk assessment and high-risk population identification. The collaborative risk calculation module establishes a risk model between the comprehensive physiological indicators of each organ system and the future risk of the target disease, and determines the stratification threshold according to the adaptive threshold determination rule when there is a nonlinear trend. The adaptive threshold determination rule determines the threshold based on the comprehensive optimal criterion of risk discrimination and stability measurement in the candidate threshold set, thereby dividing the samples into different risk levels. The risk assessment module, based on the comprehensive physiological indicators of the organ system and the synergistic combination of multiple organs, outputs: relative importance ranking, a list of synergistic combinations of multiple organs and their synergistic strength, and a joint risk score for individual samples.

[0016] This invention achieves full utilization of multidimensional physical examination physiological data to form a stable risk characterization at the organ system level through the collaboration of a data acquisition module, a disease risk calculation module, a feature recognition module, a weighted feature generation module, a collaborative risk calculation module, and a risk assessment module. It further assesses the risks of multi-organ joint exposure and its interaction / synergy, thereby enabling accurate, stable, and interpretable identification of high-risk groups for diseases in complex population scenarios and outputting risk stratification.

[0017] Preferably, in the data acquisition module, the physiological data includes: behavior, brain imaging, blood, urine, etc., and the features extracted from them include: physical examination data represented by blood pressure, grip strength, and vital capacity; blood biochemical data represented by alanine aminotransferase, bilirubin, and vitamin D; urine data represented by urea level and urinary creatinine; metabolomics data represented by amino acids and lipids; and metabolomics data represented by BCAN and GDF15.

[0018] Preferably, in the data acquisition module, the same physiological indicator in the correlation matrix is ​​associated with multiple organ systems, and each association is assigned a corresponding association weight; the correlation matrix is ​​updated according to the stability of the contribution of physiological indicators to the organ system representation in the training data.

[0019] Preferably, in the data acquisition module, physiological indicators are resampled in the training data. Based on the correlation structure or common variation patterns among physiological indicators within each organ system, a membership consistency score between each physiological indicator and each organ system is calculated. Furthermore, the stability measure of the membership consistency score under resampling conditions is calculated. The association weights in the association matrix are iteratively updated based on the stability measure, and the association weights of each physiological indicator on different organ systems are subjected to non-negativity constraints and normalization, causing the association weights to cluster towards organ systems with higher membership consistency and better stability. These organ systems include: liver, cardiovascular system, lungs (respiratory system), musculoskeletal system, kidneys, pancreas, brain, immune system, and metabolic system. For example, grip strength belongs to the musculoskeletal system, urea and creatinine belong to the kidney system, and albumin belongs to both the kidney and liver systems.

[0020] Preferably, in the disease risk calculation module, for all physiological indicators belonging to the same organ or system, an elastic net cox proportional hazards model is established to calculate the impact of each physiological feature on the future disease risk when other physiological features belonging to the organ are controlled out, and the corresponding weight coefficient beta is output from the model.

[0021] Preferably, in the feature recognition module, each physiological indicator is divided into two groups based on the median as a threshold. Using a propensity score matching algorithm, two groups with no inter-group differences in age, gender, race, and education level (variables of interest) are identified from the two groups at a 1:1 ratio. The influence of this on the future incidence of the disease is established using a Cox proportional hazards regression model, and the corresponding weight coefficients are output from the model. Depending on the amount of data, the matching ratio between the two groups is varied (e.g., 1:2, 1:3, 2:3, etc.), and this sampling process is repeated a total of 1000 times at the same matching ratio. The coefficient of variation (cv) of the weight coefficients from these 1000 iterations is then used as a measure of the feature's stability. For each physiological feature, features with more than 950 weight coefficients greater than or less than 0 are retained for subsequent analysis.

[0022] Preferably, in the weighted feature generation module, for each organ or system, based on the weight and stability measure of the physiological features selected for a certain organ, a comprehensive physiological index at the organ level is constructed. The comprehensive physiological index of an organ is calculated as: Physiological feature 1 × beta1 / CV1 + Physiological feature 2 × beta2 / CV2 + ... + Physiological feature... m ×beta m / CV mUltimately, comprehensive physiological indicators of multiple organs in the body can be calculated.

[0023] Preferably, the collaborative risk calculation module further includes: executing a collaborative combination search process: (a) Screen candidate organ systems based on the risk contribution intensity and stability of single organ systems, construct joint exposure groupings for each organ system in the candidate organ system set, and calculate additive scale interaction / synergistic risk indicators; (b) Based on the pre-defined consistency and sample support constraints, the interaction / cooperation risk indicators are resampled and verified, and combinations that do not meet the consistency constraints are pruned; (c) Only the binary combinations that have been pruned are extended to ternary and higher-order combinations, and the above evaluation and verification process is repeated to output the multi-organ synergistic combinations that satisfy the consistency constraints and their synergistic strength.

[0024] Preferably, in the synergistic risk calculation module, for the comprehensive physiological indicators of each organ, a restricted cubic spline (RCS) function is introduced into the Cox proportional hazards model to establish its relationship with the future onset of the disease; for organ-level physiological indicators with significant nonlinear relationships, the inflection point in the nonlinear relationship is used as the dividing point to divide all samples into low-risk and high-risk groups; for organ-level comprehensive physiological indicators without nonlinear relationships, the median is used as the dividing point to divide them into high-risk and low-risk groups; from all the organ-level comprehensive physiological indicators obtained after the above binarization, two are randomly selected to construct joint exposure groups: Group 1: Organ 1 high-risk + Organ 2 high-risk group; Group 2: Organ 1 high-risk + Organ 2 low-risk group; Group 3: Organ 1 low-risk + Organ 2 high-risk group; Group 4: Organ 1 low-risk + Organ 2 low-risk group; based on the Cox proportional hazards regression model, the existence of synergistic interaction between the two organs is calculated, and the magnitude of the synergistic interaction is expressed as the relative excess risk due to the interaction. The RERI (Reactive Risk Index) is calculated using group 4 as a reference group. It checks whether the disease risk value of group 1 is greater than the sum of the risks of groups 2 and 3. Using bootstrapping, 1000 random samples are taken, and the above process is repeated. At least 950 RERI values ​​greater than 0 in the 1000 calculations indicate a synergistic effect between the two organs on disease incidence risk. When the number of samples in any group within the joint exposure group corresponding to a candidate organ combination is less than a preset sample support threshold, the candidate organ combination is deemed not to meet the reliable estimation conditions and is removed from the candidate set, no longer entering the calculation of interactive / synergistic risk indicators or the higher-order combination expansion process. For the comprehensive physiological indicators of two organs with synergistic effects, a third binary organ comprehensive physiological indicator is added, and the above RERI calculation process is repeated to output whether there is a synergistic effect between the three organs.

[0025] Preferably, the risk assessment module outputs the following: a ranking of the relative importance of each organ system to the future risk of the target disease and corresponding confidence / stability scores; a list of multi-organ synergistic combinations that meet consistency constraints and their synergistic strength; for individual samples, calculating their joint risk scores at the organ system level and synergistic combination level, and stratifying them according to the joint risk scores, identifying samples whose joint risk scores exceed a preset threshold as high-risk groups for the disease.

[0026] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A disease risk assessment system based on physiological data, characterized in that: It includes: The data acquisition module acquires physiological data representing the health level of various organ systems in the body extracted from clinical health checkups. Based on the physiological functions represented by each physiological data, it constructs a physiological indicator-organ system correlation matrix to represent the correlation strength between each physiological indicator and at least one organ system. The disease incidence risk calculation module extracts a set of physiological indicators belonging to each organ system based on the correlation matrix, establishes a regularized proportional risk model for predicting the future incidence risk of the target disease, estimates the risk contribution of each physiological indicator to the future incidence risk of the target disease while controlling for the influence of other physiological indicators within the same organ system, and outputs the corresponding risk contribution parameters. The feature recognition module divides the sample into candidate groups with different exposure levels for each physiological indicator, and adopts a confounding factor control strategy to achieve a balanced distribution of different groups on covariates such as age, gender, race, and education level; then, under multiple repeated sampling / resampling conditions, the risk contribution parameter of the physiological indicator is repeatedly estimated to obtain a stability measure of the risk contribution parameter. Using the stability metric as a screening constraint, only physiological indicators that have consistent risk contribution direction and stability that meet a preset threshold under repeated sampling conditions are retained, thus obtaining stable weights for subsequent comprehensive organ system characterization construction. The weighted feature generation module constructs a comprehensive physiological index for each organ system based on the selected physiological indicators, their weight coefficients, and stability within that organ system. This comprehensive physiological index is a weighted combination of the values ​​of each physiological indicator, their weight coefficients, and stability, to reduce the impact of sample perturbation on the comprehensive physiological index and improve cross-population applicability. It outputs comprehensive physiological indices for multiple organ systems as input for subsequent collaborative risk assessment and high-risk population identification. The collaborative risk calculation module establishes a risk model between the comprehensive physiological indicators of each organ system and the future risk of the target disease, and determines the stratification threshold according to the adaptive threshold determination rule when there is a nonlinear trend. The adaptive threshold determination rule determines the threshold based on the comprehensive optimal criterion of risk discrimination and stability measurement in the candidate threshold set, thereby dividing the samples into different risk levels. The risk assessment module, based on the comprehensive physiological indicators of the organ system and the synergistic combination of multiple organs, outputs: relative importance ranking, a list of synergistic combinations of multiple organs and their synergistic strength, and a joint risk score for individual samples.

2. The disease risk assessment system based on physiological data according to claim 1, characterized in that: The data acquisition module contains physiological data including: behavior, brain imaging, blood, and urine. The features extracted from these data include: physical examination data represented by blood pressure, grip strength, and vital capacity; blood biochemical data represented by alanine aminotransferase, bilirubin, and vitamin D; urine data represented by urea levels and creatinine; metabolomics data represented by amino acids and lipids; and metabolomics data represented by BCAN and GDF15.

3. The disease risk assessment system based on physiological data according to claim 2, characterized in that: In the data acquisition module, the same physiological indicator in the correlation matrix is ​​associated with multiple organ systems, and each association is assigned a corresponding association weight; the correlation matrix is ​​updated according to the stability of the contribution of physiological indicators to the organ system representation in the training data.

4. The disease risk assessment system based on physiological data according to claim 3, characterized in that: In the data acquisition module, physiological indicators are resampled in the training data. Based on the correlation structure or common variation patterns among physiological indicators in each organ system, the membership consistency score between each physiological indicator and each organ system is calculated, and the stability measure of the membership consistency score under the resampling condition is further calculated. The association weights in the association matrix are iteratively updated according to the stability measure, and the association weights of each physiological indicator on different organ systems are subjected to non-negativity constraints and normalization processing, so that the association weights are clustered towards organ systems with higher membership consistency and better stability. These organ systems include: liver, cardiovascular, lung, musculoskeletal, kidney, pancreas, brain, immune, and metabolic.

5. The disease risk assessment system based on physiological data according to claim 4, characterized in that: In the disease risk calculation module, for all physiological indicators belonging to the same organ or system, an elastic network proportional risk model is established to calculate the impact of each physiological feature on the future disease risk when other physiological features belonging to the organ are controlled out, and the corresponding weight coefficient beta is output from the model.

6. The disease risk assessment system based on physiological data according to claim 5, characterized in that: In the feature recognition module, each physiological indicator is divided into two groups based on the median threshold. Using a propensity score matching algorithm, two groups with no inter-group differences in age, gender, race, and education level (variables of interest) are identified from the two groups at a 1:1 ratio. The influence of this indicator on the future incidence of the disease is established using a Cox proportional hazards regression model, and the corresponding weight coefficients are output from the model. Depending on the amount of data, the matching ratio between the two groups is varied, and this sampling process is repeated a total of 1000 times at the same matching ratio. The coefficient of variation (cv) of the weight coefficients from these 1000 iterations is then used as a measure of the feature's stability. For each physiological characteristic, features with more than 950 weight coefficients greater than or less than 0 out of 1000 were retained for subsequent analysis.

7. The disease risk assessment system based on physiological data according to claim 6, characterized in that: In the weighted feature generation module, for each organ or system, a comprehensive physiological index at the organ level is constructed based on the weights and stability measures of the physiological features selected for that organ. The comprehensive physiological index is calculated as: Organ Comprehensive Physiological Index = Physiological Feature 1 × beta1 / CV1 + Physiological Feature 2 × beta2 / CV2 + ... + Physiological Feature m ×beta m / CV m .

8. The disease risk assessment system based on physiological data according to claim 7, characterized in that: The collaborative risk calculation module further includes: executing a collaborative combination search process. (a) Screen candidate organ systems based on the risk contribution intensity and stability of single organ systems, construct joint exposure groupings for each organ system in the candidate organ system set, and calculate additive scale interaction / synergistic risk indicators; (b) Based on the pre-defined consistency and sample support constraints, the interaction / cooperation risk indicators are resampled and verified, and combinations that do not meet the consistency constraints are pruned; (c) Only the binary combinations that have been pruned are extended to ternary and higher-order combinations, and the above evaluation and verification process is repeated to output the multi-organ synergistic combinations that satisfy the consistency constraints and their synergistic strength.

9. The disease risk assessment system based on physiological data according to claim 8, characterized in that: In the collaborative risk calculation module, for the comprehensive physiological indicators of each organ, a restricted cubic spline function is introduced into the Cox proportional hazards model to establish its relationship with the future onset of the disease. For organ-level physiological indicators with significant nonlinear relationships, the inflection point in the nonlinear relationship is used as the dividing point to divide all samples into low-risk and high-risk groups. For organ comprehensive physiological indicators without nonlinear relationships, the median is used as the dividing point to divide them into high-risk and low-risk groups. From all the organ comprehensive physiological indicators obtained after the above binarization, two are randomly selected to construct joint exposure groups: Group 1: Organ 1 high-risk + Organ 2 high-risk group; Group 2: Organ 1 high-risk + Organ 2 low-risk group; Group 3: Organ 1 low-risk + Organ 2 high-risk group; Group 4: Organ 1 low-risk + Organ 2 low-risk group. Based on the Cox proportional hazards regression model, we calculate whether there is a synergistic interaction between two organs. The magnitude of the synergistic interaction is represented by the relative excess risk (RERI) caused by the interaction. Using group 4 as the reference group, we calculate whether the disease risk value of group 1 is greater than the sum of the risks of groups 2 and 3. Using bootstrapping, 1000 random samples are taken and the above process is repeated. If at least 950 of the 1000 calculated RERIs are greater than 0, it indicates that there is a synergistic effect between the two organs on the risk of disease incidence. When the number of samples in any group of the joint exposure group corresponding to the candidate organ combination is less than the preset sample support threshold, the candidate organ combination is determined not to meet the reliable estimation conditions, is removed from the candidate set, and will not be included in the calculation of interaction / synergistic risk indicators or the higher-order combination expansion process. For the comprehensive physiological indicators of two organs with synergistic effects, the above RERI calculation process is repeated by adding a third binary organ comprehensive physiological indicator to output whether there is a synergistic effect between the three organs.

10. The disease risk assessment system based on physiological data according to claim 9, characterized in that: The risk assessment module outputs the following: a ranking of the relative importance of each organ system to the future risk of developing the target disease and corresponding confidence / stability scores; a list of multi-organ synergistic combinations that meet consistency constraints and their synergistic strength; for individual samples, calculating their joint risk scores at the organ system level and synergistic combination level, and stratifying them according to the joint risk scores, identifying samples whose joint risk scores exceed a preset threshold as high-risk groups for the disease.

Images