Metabolic product risk assessment and monitoring system for high altitude aviation workers

Abstract

The application discloses a metabolic product risk assessment and monitoring system for plateau aviation practitioners, and relates to the technical field of compound testing. The system comprises initialization determination modules, online separation and refinement modules, comparison verification and determination modules, application analysis modules and standardized and traceable analysis modules which are sequentially operated. The technical key points are as follows: the scheme can successfully identify the plateau-specific implicit exposure source through the implicit exposure source identification mechanism driven by the standardization index and the goodness-of-fit, and reversely optimize the operation scheduling. This effect depends on the high-quality and high-representative data provided by the first two stages. The three form an overall indivisible system from accurate measurement to efficient utilization to intelligent attribution. The system breaks through the limitations of traditional exposure assessment, realizes the transition from static description to dynamic traceability, and realizes the whole-process management from health monitoring to operation intervention.

Description

Technical Field

[0001] This invention relates to the field of compound testing technology, specifically to a system for assessing and monitoring the risks of metabolites produced by personnel working in high-altitude aviation industries. Background Technology

[0002] Compound testing refers to the technical process of qualitatively identifying and quantitatively analyzing specific chemical substances, aiming to determine their physicochemical properties such as type, structure, concentration, and purity. In the environmental and biomedical fields, compound testing is often used to detect the content of target pollutants or their metabolites in samples. Typical methods include chromatography, such as gas chromatography (GC), liquid chromatography (LC), and mass spectrometry (MS), combined with sample pretreatment techniques to achieve accurate determination of trace compounds in complex matrices.

[0003] In the risk assessment and monitoring of aromatic persistent organic pollutant (POPs) metabolites among aviation workers in high-altitude environments, existing technologies face multiple intertwined technical challenges: On the one hand, the PM2.5 matrix in high-altitude environments is complex, rich in interfering substances such as lipids and pigments, while the concentrations of target metabolites, such as 1-OH-Pyrene and 6-OH-BDE-47, are extremely low. Traditional GC-MS methods struggle to achieve both high sensitivity and high resolution, easily leading to co-elution or signal loss. On the other hand, two-dimensional coupled systems, such as LC-GC, often suffer from asynchrony between LC elution and GC injection times, causing condensation or retention of target analytes during transport, resulting in peak tailing or even quantification failure. Furthermore, the challenges faced by aviation workers... The amount of blood collected per person is relatively small, and traditional offline pretreatment steps are cumbersome, resulting in large fluctuations in recovery rates (e.g., PCB metabolite recovery rates are often between 70% and 120%). Furthermore, human error is easily introduced, distorting data for individuals with low exposure. More importantly, existing exposure assessments lack quantitative correlations with dynamic environmental factors, making it difficult to identify the key variables that truly drive exposure. For example, the actual exposure level of ground staff at an airport may vary significantly depending on the number of takeoffs and landings or atmospheric diffusion conditions on that day. However, traditional methods cannot capture such hidden exposure sources, leading to delayed risk warnings or mismatched protective measures. These problems collectively restrict the accuracy and sensitivity of occupational health monitoring in high-altitude aviation. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A risk assessment and monitoring system for metabolites of high-altitude aviation personnel is proposed. The system includes an initialization measurement module: a standard substance is selected as the development benchmark, spiked samples are configured and compiled to form a standard sample, and the standard sample is injected into a pre-constructed GC-MS platform for initial condition testing. Characteristic ion peak spectra of the target compound are obtained, a predefined traversal test is run, the signal-to-noise ratio-resolution coordinating function results after each traversal test are obtained, and the parameter combination in the corresponding traversal test that maximizes the results is extracted to obtain the optimal parameter combination.

[0006] Online separation and refinement module: Online liquid chromatography (OLLC) is used as the online separation unit. After gradient elution of the target column, interfering substances other than the target compound are retained on the target column. A dynamic matching window mechanism is introduced to determine the time synchronization index and execute the target control strategy according to the sub-constraint conditions.

[0007] Comparison and Validation Module: The constructed Online LC-GC-MS method is standardized and compared, a performance indicator dataset is obtained, the rule engine is run, and a judgment action is performed on the performance indicator dataset. If any indicator in the performance indicator dataset fails to meet the standard, a correction strategy for the benchmark is completed.

[0008] Application Analysis Module: Completes application processing of the validated Online LC-GC-MS method, introduces a sample utilization factor during the application processing, determines whether the current application processing needs to be optimized, and if so, runs the defined improvement strategy;

[0009] Standardization and source tracing analysis module: Standardizes the concentration data of target aromatic POPs and metabolites, derives the standardized index, and identifies different levels of exposure after comparison. Constructs a multiple regression model, fits it using the least squares method, and determines key influencing factors after significance testing. If the goodness of fit of the multiple regression model is less than the preset standard threshold, a variable supplementation mechanism is triggered. After completing the test, the analysis results are output.

[0010] Furthermore, the process of running the predefined traversal test is as follows:

[0011] Key performance indicators are extracted from characteristic ion peak spectra, including at least the signal-to-noise ratio (S / N) and the theoretical plate number (R).

[0012] The signal-to-noise ratio-resolution coordinating function is run based on the key performance indicators. The key performance indicators are accumulated and multiplied to obtain the required running result, which is the coordinating function value F_SR.

[0013] Different preset parameter combinations are tested; the parameter combinations include column temperature program, carrier gas flow rate and injection port temperature; if there are multiple target compounds, the cooperative function value F_SR of each target compound is calculated after each run, and the average value is taken as the final cooperative function value F_SR. The parameter combination that maximizes the final cooperative function value F_SR is extracted to obtain the optimal parameter combination.

[0014] Furthermore, the dynamic matching window mechanism operates as follows:

[0015] The elution time of the target compound in LC and the preparation time of GC are obtained, and the result of dividing the elution time of the target compound in LC by the preparation time of GC is denoted as the time synchronization index T_s. The time synchronization index T_s is compared with the preset sub-constraint condition, which is: the time synchronization index T_s ∈ [T_mo1, T_mo2]. When the time synchronization index T_s < T_mo1, the target regulation strategy is executed to delay the GC start time; when the time synchronization index T_s > T_mo2, the target regulation strategy is executed to accelerate the LC elution speed.

[0016] Where [T_mo1, T_mo2] represent the tolerance interval.

[0017] Furthermore, when the time synchronization index T_s < T_mo1, the target control strategy is as follows: the absolute value of the difference between the elution time of the target compound in the LC and the time when the GC is ready to receive is recorded as the delay time Δt, and the GC heating program is triggered again after the delay time Δt after the LC starts running.

[0018] When the time synchronization index T_s > T_mo2, the target control strategy is as follows: set the maximum allowable LC elution time, combine the elution time of the target compound in LC as input, run the pre-constructed flow rate increase ratio derivation model, and output the percentage of flow rate increase Δ%. The flow rate increase ratio derivation model is run as follows: obtain the maximum allowable LC elution time t_target: t_target = T_mo2 × t_GC; where t_GC is the time for GC to prepare for reception; then, divide the elution time of the target compound in LC by the maximum allowable LC elution time to obtain the first ratio, and then subtract 1 from the first ratio to generate the flow rate increase Δ.

[0019] Furthermore, the standardized comparative verification process is as follows:

[0020] Prepare a uniform spiking matrix and prepare mixed standard solutions containing the target analytes; set up Q concentration gradients, prepare 3 parallel samples for each concentration, and run the Online LC-GC-MS method and the current standard method in parallel; calculate the linear correlation coefficient, limit of detection (LOD), spiked recovery rate, and relative standard deviation (RSD) for each target analyte; the performance index dataset should include at least the following: linear correlation coefficient, limit of detection, spiked recovery rate, and relative standard deviation.

[0021] Furthermore, the process of performing the judgment action on the performance metric dataset is as follows:

[0022] If the linear correlation coefficient does not meet the standard, the correction strategy is as follows: check whether the high concentration point has saturated response, and use weighted least squares (WLS) fitting instead; if the detection limit does not meet the standard, the correction strategy is as follows: switch GC-MS to MRM mode, optimize the LC-GC interface temperature until no adsorption occurs, and increase the PTV injection volume according to the preset gradient; if the spiked recovery rate does not meet the standard, the correction strategy is as follows: introduce an isotope internal standard to correct the matrix effect, add an enzymatic digestion step to the blood sample, and use a matrix-matched calibration curve instead; if the relative standard deviation does not meet the standard, the correction strategy is as follows: replace the LC column or regenerate it.

[0023] Furthermore, the criteria for determining whether the current application processing needs optimization are as follows:

[0024] Define the sample utilization factor U_f: U_f = V_used / V_total × R_recovery / 100;

[0025] In the formula, V_used represents the equivalent volume actually entering the instrument for analysis, and V_total represents the total plasma collection volume; R_recovery represents the spiked recovery rate obtained after verification by the comparison and judgment module; when the sample utilization factor U_f > the system set threshold, it is determined that the current application processing does not need optimization; otherwise, it is determined that the current application processing needs optimization, and the improvement strategy is to switch to microcolumn SPE.

[0026] Furthermore, the basis for identifying different levels of exposure after obtaining the standardized index and comparing it is as follows: the measured concentration of the target substance in any sample is denoted as C_i, the corresponding reference background value is denoted as C_ref, and the ratio between the two, C_i / C_ref, is denoted as the standardized index N_i; if the standardized index N_i exceeds the set value, it is defined as a high exposure level; otherwise, it is defined as a low exposure level; if the distribution of the standardized index N_i of any batch of samples shows an excessive deviation, a feedback adjustment mechanism is executed: the failure degree of the micro-column SPE used in the improvement strategy is checked, and regression correction is completed.

[0027] Furthermore, the basis for constructing the multiple regression model is: E=β_0+β_1×Th+β_2×Hr+β_3×An;

[0028] In the formula, E represents the individual exposure level, β_0 represents the intercept term, β_1, β_2, and β_3 are all partial regression coefficients, Th represents the average daily outdoor working time, Hr represents the altitude of the airport, and An represents the length of employment. The least squares method is used for fitting, and the key influencing factors are determined after significance testing. The criterion for significance testing is: the p-value corresponding to the partial regression coefficient of any variable is <0.05, and the key influencing factor is the corresponding variable.

[0029] Furthermore, the triggered variable replenishment mechanism is as follows:

[0030] When the goodness of fit of the multiple regression model is less than the preset standard threshold, it prompts the addition of potential covariates, including at least: wind speed Wq, daily average flight density Fm, and personal protective equipment usage rate Pt. The completed test action is: add the potential covariates to the multiple regression model in sequence to complete the test. The potential covariates that meet the conditions that the goodness of fit of the multiple regression model is not less than the preset standard threshold and the corresponding p value is <0.05 are identified as latent exposure sources.

[0031] This invention provides a system for assessing and monitoring the risk of metabolites in high-altitude aviation personnel, which has the following beneficial effects: (1) This scheme constructs a dual optimization method through the signal-to-noise ratio-resolution coordination function and the time synchronization index, and replaces the experience trial and error with the quantitative method to ensure that pg / mL level metabolites in high-altitude samples are stably detected. The dynamic matching window realizes the millisecond-level alignment of LC and GC, which greatly improves the transfer efficiency. The two together ensure the high fidelity of the front-end data, and solve the problems of the traditional GC-MS method in which it is difficult to balance sensitivity and separation under complex matrix, and the time mismatch of the two-dimensional coupling system leads to peak distortion or loss of target substances, laying a reliable foundation for all subsequent analyses.

[0032] (2) This scheme enhances the effective signal by using microcolumn SPE and uses the validated spiked recovery rate to dynamically evaluate the sample utilization factor. This not only improves the detection coverage, but also ensures that the subsequent output results are truly representative. It solves the contradiction between the small amount of blood samples and the extremely low concentration of metabolites among high-altitude aviation personnel, avoids the loss of pretreatment and the missed detection of low-exposure populations, and enhances the overall comprehensiveness of the detection.

[0033] (3) This scheme can successfully identify plateau-specific hidden exposure sources through the identification mechanism driven by the standardization index and goodness of fit, and optimize the operation scheduling in reverse. This effect depends on the high quality and high representative data provided in the first two stages. The three form an integral and inseparable system from accurate measurement to efficient utilization and then to intelligent attribution. It breaks through the limitations of traditional exposure assessment, realizes the leap from static description to dynamic source tracing, and realizes the whole process governance from health monitoring to operation intervention. Attached Figure Description

[0034] Figure 1 This is a simplified flowchart illustrating the technical approach in this invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figure 1 This embodiment provides a system for assessing and monitoring the risks of metabolic products from personnel working in high-altitude aviation sectors. The technical approach described in this embodiment is based on... Figure 1 The system, implemented or operated according to the process framework shown, uses online liquid chromatography-gas chromatography-mass spectrometry (LC-GC-MS) as its core analytical method. It focuses on the separation, detection, and exposure assessment of trace aromatic POPs in PM2.5 and their metabolites in human blood with high sensitivity and selectivity. It is particularly designed to address key issues such as strong matrix interference, low metabolite concentration, and abnormal physiological responses in the complex environment of high altitudes, and constructs a technical system with continuous, interconnected, and intelligent control capabilities.

[0037] The system mainly relies on the following modules and components to operate, specifically:

[0038] I. Initialization of the measurement module:

[0039] Standard substances were selected as the development benchmark. Spiked samples were prepared and compiled to form standard samples. The standard samples were then injected into a pre-constructed GC-MS platform for initial condition testing. Characteristic ion peak spectra of the target compounds were obtained. Predefined traversal tests were run, and the signal-to-noise ratio-resolution coordinating function results after each traversal test were obtained. The parameter combination in the corresponding traversal test that maximized the results was extracted to obtain the optimal parameter combination.

[0040] In this embodiment, the standard material can refer to: NIST SRM 1649a, such as the urban particulate matter standard reference material, used to simulate the real PM2.5 matrix; the spiked sample contains at least typical aromatic POPs and matrix interfering substances; the typical aromatic POPs include typical polycyclic aromatic hydrocarbons, polychlorinated biphenyls, polychlorinated naphthalenes, polybrominated biphenyls (ethers), and phthalates, etc.; the matrix interfering substances contain a large number of lipids, pigments, alkaloids, and polymers, etc.; these compounds have the characteristics of high toxicity, poor degradation and easy accumulation in organisms, and are key pollutants of concern in the high-altitude airport environment, which is targeted to a certain extent; the target compound refers to one or more of the above-mentioned compounds.

[0041] The GC-MS platform is based on a triple quadrupole mass spectrometer, i.e., GC-MS / MS, and is equipped with a programmed temperature inlet (PTV), a highly inert fused silica transfer line, and a high-purity helium carrier gas system. It is a conventional, existing setup platform, so it will not be described in detail here. The initial conditions were tested by setting the initial values ​​for the ion source temperature, electron energy, mass scan range, and inlet split ratio. The specific criteria are as follows:

[0042] Ion source temperature: initially set to 230℃ to ensure full ionization of the target compound and avoid thermal decomposition;

[0043] Electron energy: fixed at 70 eV, conforming to NIST mass library matching standards;

[0044] Mass scan range: set to m / z 50~500, covering the vast majority of aromatic POPs and their fragment ions;

[0045] Injection port split ratio: initially set at 1:10 to balance the injection volume with the risk of column overload.

[0046] The characteristic ion peak spectra of the target compound obtained are: total ion chromatogram (TIC) and extracted ion chromatogram (EIC).

[0047] The process of running a predefined traversal test is as follows:

[0048] (1) Extract key performance indicators from characteristic ion peak spectra, including: signal-to-noise ratio (S / N) and theoretical plate number (R);

[0049] The signal-to-noise ratio (S / N) is defined as the ratio of the maximum intensity of the target peak to the standard deviation of the noise of its neighboring baseline, reflecting the detection sensitivity; the theoretical plate number R is given by the formula: R = 16(t_R / W_b). 2The calculation is performed, where t_R is the retention time and W_b is the peak width, which is used to quantify the separation efficiency of the chromatographic column. Traditional methods usually optimize S / N or R separately, which can easily lead to an imbalance state of high sensitivity but co-elution or high separation but weak signal. Therefore, this embodiment innovatively proposes a signal-to-noise ratio-resolution synergistic function as a comprehensive evaluation index.

[0050] (2) Run the signal-to-noise ratio-resolution coordination function based on the key performance indicators, and multiply the key performance indicators to obtain the required running result, which is: the coordination function value F_SR;

[0051] Within a limited analysis time, sensitivity and resolution constitute a pair of coupled constraints. If only the S / N ratio is increased, it may lead to peak broadening or even overlap, reducing R. Conversely, if high R is excessively pursued, the analysis cycle will be lengthened, and some low-concentration components may have a decreased S / N due to diffusion loss. Therefore, it is necessary to find an operating point that maximizes the product of the two, that is, to achieve the maximum effective information output per unit time under given resources.

[0052] (3) Conduct traversal tests on different preset parameter combinations; wherein, the parameter combination includes column temperature program, carrier gas flow rate and injection port temperature; wherein, the heating rate in the column temperature program is adjusted according to the set step value, such as 3℃ / min, 5℃ / min, 7℃ / min, the carrier gas flow rate is adjusted according to the set step value, such as 0.8, 1.0, 1.2 mL / min, and the injection port temperature is adjusted in the same way, such as 260℃, 280℃, 300℃.

[0053] (4) If there are multiple target compounds, the cooperative function value F_SR of each target compound will be automatically calculated after each run, and the average value will be taken as the final cooperative function value F_SR. The parameter combination that maximizes the final cooperative function value F_SR will be extracted to obtain the optimal parameter combination.

[0054] The selected parameter combination lays a solid foundation for the time synchronization and quantitative accuracy of the subsequent Online LC-GC interface. The introduced signal-to-noise ratio-resolution coordination function realizes the transformation from empirical trial and error to vectorized optimization, ensuring that the GC-MS platform has both high sensitivity and high resolution in the analysis of complex samples in high-altitude areas. It is a key guarantee for the technical consistency and reliability of the entire system.

[0055] II. Online Separation and Refinement Module:

[0056] Online liquid chromatography (OLLC) was used as the online separation unit. After gradient elution of the target column, interfering substances other than the target compound were retained on the target column. The optimal parameter combination obtained in S1 was used for processing. A dynamic matching window mechanism was introduced to determine the time synchronization index and execute the target control strategy according to the sub-constraint conditions.

[0057] Among them, online liquid chromatography (LC) serves as an online separation unit, namely the first-dimensional purification and selective enrichment unit. Its core task is to allow only target compounds with moderate polarity and small molecular weight to enter the GC through the interface, while retaining large molecular interferences on the LC column and removing them through the washing gradient.

[0058] To address the potential issues of excessively fast LC elution rates or excessively slow GC heating in actual operation, the aforementioned dynamic matching window mechanism was specifically introduced to reduce cross-contamination. Regarding the issue of excessively fast LC elution rates, the following explanations apply: If the LC gradient elution rate is too high, premature elution may occur before the GC has reached the target temperature (e.g., still below 300℃), leading to problems such as component condensation at the injection port, peak tailing or splitting, or lower quantitative results. Conversely, if the GC heating program is conservative, such as a rate of only 3℃ / min, while ensuring separation quality, the extended GC preparation time forces the LC to extend the gradient to wait for the GC, resulting in longer analysis cycles, reduced sample throughput, and increased risk of co-elution of high-boiling-point interfering substances.

[0059] The specific operational basis or process of the dynamic matching window mechanism is as follows:

[0060] The elution time of the target compound in the LC and the preparation time of the GC are obtained, and the result of dividing the elution time of the target compound in the LC by the preparation time of the GC is recorded as the time synchronization index T_s. The time synchronization index T_s is compared with the preset sub-constraint conditions. The sub-constraint conditions are: the time synchronization index T_s ∈ [T_mo1, T_mo2]. When the sub-constraint conditions are met, no response action is taken. When the time synchronization index T_s < T_mo1, the target regulation strategy is executed to delay the GC start time. When the time synchronization index T_s > T_mo2, the target regulation strategy is executed to accelerate the LC elution speed. Here, [T_mo1, T_mo2] represents the tolerance interval. In this embodiment, the two end values ​​in the tolerance interval can be set to [0.9, 1.1].

[0061] Specifically, the elution time of the target compound in LC is the actual elution time, in minutes. The GC preparation time is the time required from startup to reaching the optimal parameter combination determined in S1, in minutes, including processes such as heating, stabilization, and valve switching. This value has been obtained through no-load calibration in the S1 stage.

[0062] When the time synchronization index T_s < T_mo1, the benchmark control strategy is as follows: the absolute value of the difference between the elution time of the target compound in the LC and the time when the GC prepares to receive it is recorded as the delay time Δt. After the LC starts running, the GC heating program is triggered again after the delay time Δt. The basis for this is that the GC heating program has a programmable delay function, and the delayed start does not affect the final temperature curve, only changing the absolute time axis. When the time synchronization index T_s > T_mo2, the benchmark control strategy is as follows: the maximum allowable LC elution time is set, and the elution time of the target compound in the LC is used as input. A pre-built flow rate increase ratio derivation model is run, and the percentage of flow rate increase Δ% is output to guide the adjusted LC elution rate. It should be noted that the percentage of flow rate increase Δ% output must ensure that the final LC elution rate does not exceed the theoretical maximum value. If it does, subsequent adjustments can be made based on the theoretical maximum value corresponding to the LC elution rate.

[0063] The process of deriving the velocity increase ratio model is as follows:

[0064] Obtain the maximum allowable LC elution time t_target: t_target = T_mo2 × t_GC; where t_GC is the time for GC to prepare for reception; then, divide the elution time of the target compound in LC by the maximum allowable LC elution time to obtain the first ratio, and then subtract 1 from the first ratio to generate the flow rate increase Δ, and the output is the percentage of the flow rate increase Δ%.

[0065] The derivation is based on the following: The goal is to shorten the elution time of the target compound in LC to the upper limit of the allowable range, i.e., the maximum allowable LC elution time. Since the flow rate and retention time are approximately inversely proportional, the required increase in flow rate is equal to the relative reduction in time. The time reduction factor is the elution time of the target compound in LC divided by the maximum allowable LC elution time. The portion exceeding 1 is the required increase in flow rate, which is converted to a percentage, Δ%. This formula is used to accurately calculate the minimum necessary flow rate adjustment, eliminating GC vacancy without disrupting separation and improving system throughput and synchronization. Specifically, the system automatically adjusts the LC gradient program according to the percentage increase in flow rate Δ%, i.e., increasing the original flow rate by Δ%.

[0066] By adopting the above technical solutions, this system constructs a dual optimization method through the signal-to-noise ratio-resolution coordination function and the time synchronization index T_s. The initialization measurement module replaces empirical trial and error with a quantitative approach to ensure the stable detection of pg / mL metabolites in high-altitude samples. The online separation and refinement module achieves millisecond-level alignment between LC and GC through a dynamic matching window, which greatly improves the transfer efficiency. Together, they ensure the high fidelity of the front-end data and solve the problems of traditional GC-MS methods in balancing sensitivity and resolution in complex matrices, as well as the peak distortion or target loss caused by time mismatch in two-dimensional coupling systems. This lays a reliable foundation for all subsequent analyses.

[0067] III. Comparison, Verification, and Judgment Module:

[0068] The constructed Online LC-GC-MS method is standardized and compared for verification. A performance metric dataset is obtained, the rule engine is run, and the judgment action is performed on the performance metric dataset. If any metric in the performance metric dataset fails to meet the standard, the correction strategy is completed. After correction, the comparison verification and judgment module is run again until any metric in the performance metric dataset meets the standard.

[0069] The standardized comparison and verification process is as follows:

[0070] First, a standardized spiking matrix was prepared: certified blank PM2.5 matrix filter membranes and blank human plasma were selected as the matrix; mixed standard solutions containing target analytes, such as PCB-153 and 1-OH-Pyrene, were prepared; Q concentration gradients were set, and since Q=6, the specific concentration gradients could be: 0.1, 0.2, 0.5, 1.0, 5.0, and 10.0 ng / mL; three parallel samples were prepared for each concentration to ensure statistical robustness; second, two methods were run in parallel: Method A, i.e., the Online LC-GC-MS method, and Method B, i.e., the current national standard method, such as HJ 836-2017, all samples were pretreated and analyzed by the same operator within the same batch to eliminate human or batch bias. Finally, performance indicators were calculated: linear correlation coefficient, limit of detection (LOD), spiked recovery, and relative standard deviation (RSD) were calculated for each target analyte. The linear correlation coefficient was derived from a 6-point calibration curve using the external standard method. The LOD was calculated as: LOD = 3.3 × σ / Sr, where σ is the standard deviation of 10 replicates in the blank, and Sr is the slope of the calibration curve. The spiked recovery was calculated as: spiked recovery = (measured value - blank value) / theoretical value × 100%. The measured value represents the concentration of the spiked sample, the blank value represents the concentration of the blank or unspecified sample, and the theoretical value refers to the theoretical concentration increment corresponding to the actual addition of standard solution to the sample. The relative standard deviation (RSD) was the mean of the peak area RSDs of 3 replicates at each of the 6 concentration points.

[0071] Therefore, the performance metrics dataset includes: linear correlation coefficient, detection limit, spiked recovery rate, and relative standard deviation.

[0072] The basis or process for making judgments on the performance index dataset is explained below:

[0073] When the linear correlation coefficient does not meet the standard, such as the linear correlation coefficient R 2 When the limit of detection (LOD) is <0.995, the correction strategy for the target is as follows: check whether the response is saturated at high concentration points (i.e., points exceeding the set concentration value), switch to weighted least squares (WLS) fitting, and use an autosampler to ensure accurate injection volume. When the limit of detection is not met, such as ≥0.05 ng / mL, the correction strategy for the target is as follows: switch GC-MS to MRM mode, optimize the LC-GC interface temperature until no adsorption occurs, and increase the PTV injection volume according to the preset gradient. When the spiked recovery rate is not met, such as <85% or >115%, the correction strategy for the target is as follows: introduce an isotope internal standard to correct the matrix effect; add an enzymatic digestion step to the blood sample to release bound metabolites; switch to matrix-matched calibration curves. When the relative standard deviation (RSD) is not met, such as >10%, the correction strategy for the target is as follows: replace or regenerate the LC column; calibrate the seal between the carrier gas regulator and the injection needle; insert QC control samples into each batch of samples to monitor stability.

[0074] By adopting the above technical solution and completing parallel verification, it was confirmed that the method has higher sensitivity, better accuracy and stronger reproducibility under complex high-altitude substrates, solving the loss and contamination problems caused by multiple preprocessing steps in traditional methods. At the same time, the verification process found that online two-dimensional coupling can effectively avoid the uncertainty brought about by offline purification, making the method more suitable for high-frequency and large-volume monitoring scenarios. This is an unexpected improvement in operational efficiency.

[0075] The specific results and differences are shown in Table 1:

[0076] Table 1: Comparison and Gap Analysis of Core Indicators

[0077] Performance indicators Method of the present invention Traditional standard method Gap Analysis [R 2 ]] 0.998 0.992 This method has better linearity because the first dimension of LC effectively removes matrix interference, resulting in a more stable response. LOD 0.04 0.20 The 80% reduction is attributed to the improved signal-to-noise ratio achieved through large-volume PTV injection and the MRM mode. Spiked recovery rate 92~105 70–120 Traditional methods are greatly affected by purification losses, while this method reduces operational steps by transferring data online. RSD 4.2 9.8 This method has a high degree of automation and minimal human error.

[0078] Based on the conclusions drawn from Table 1 above, this method is significantly superior to the current standard method in terms of sensitivity, accuracy, and precision. Combined with the above-mentioned benchmarking process, and after the benchmarking correction strategy, it can achieve more stable and efficient processing. Figure 1The specific implementation of the external standard method for quantification shown in the example is as follows: Once the method is validated, a fixed calibration scheme is adopted: using 6 concentration points, with 3 repetitions per point; fitting a linear equation Y=aX+b, where: Y is the peak area of ​​the characteristic ion of the target analyte; X is the concentration in ng / mL; a is the sensitivity, i.e., the response factor, with a value > 0; b is the system blank offset, which should ideally be close to 0; and R0 is required. 2 >0.995, and the blank sample had no interfering peak at the target retention time.

[0079] IV. Application Analysis Module:

[0080] The validated Online LC-GC-MS method is used to complete the application processing. During the application processing, a sample utilization factor is introduced to determine whether the current application processing needs to be optimized. If so, the defined improvement strategy is run until no optimization is needed.

[0081] In this embodiment, the application processing is applied to two core samples in a real high-altitude aviation scenario: PM2.5 filter membrane extract from the airport environment and venous plasma from aviation personnel. The specific process is as follows: PM2.5 sample processing: PM2.5 filter membranes collected from high-altitude airports are shredded; ultrasonic-assisted extraction is performed using methanol:dichloromethane = 1:1, twice for 30 minutes; the extracts are combined, concentrated to near dryness by nitrogen blowing, and redissolved with acetonitrile; filtered through a 0.22μm filter membrane before being processed; Plasma sample processing: venous blood is collected, anticoagulated with EDTA, and centrifuged to separate plasma; three times the volume of pre-cooled methanol / acetonitrile (1:1) is added for protein precipitation; vortexing for 1 minute, standing at 4℃ for 10 minutes, centrifuged at 13000 g for 10 minutes; the supernatant is collected, purified using an Oasis HLB solid-phase extraction column (SPE), and activated by: 3 mL methanol → 3 mL water; loading: plasma supernatant; rinsing: 2 mL 5% methanol aqueous solution. mL; Eluent: 3 mL acetonitrile; Eluent was dried under nitrogen and diluted to 100 μL with acetonitrile for LC-GC-MS analysis.

[0082] The criteria for determining whether the current application processing needs optimization are as follows:

[0083] In practice, there is a pair of mutually restrictive sub-constraints: Sub-constraint A (small sample size): the blood volume of high-altitude aviation personnel is usually ≤1 mL per blood sample. If the pretreatment loss is large, the effective volume available for analysis is insufficient; Sub-constraint B (extremely low concentration): metabolites are mostly in the pg / mL range, such as 6-OH-BDE-47, which requires a high recovery rate to be detected.

[0084] To quantify the aforementioned sub-constraint contradictions and guide optimization;

[0085] This invention defines the sample utilization factor U_f: U_f = V_used / V_total × R_recovery / 100;

[0086] In the formula, V_used represents the actual equivalent volume entering the instrument for analysis, considering the conversion between SPE elution volume and injection volume; V_total represents the total plasma collection volume; R_recovery represents the spiked recovery rate obtained after validation in the previous module; where R_recovery / 100 is a dimensionless ratio, so U_f is also a dimensionless number, reflecting the effective analytical signal that a unit of original sample can contribute; when the sample utilization factor U_f > the system set threshold, such as U_f > 0.8, it is determined that the current application processing does not need optimization; otherwise, it indicates that the sample or current processing efficiency is insufficient and needs optimization; at this time, it is determined that the current application processing needs optimization, and the improvement strategy is: to use microcolumn SPE, such as μElution Plate, with an elution volume of only 50μL; the corresponding technical basis is: reduce the elution volume → increase the equivalent concentration of V_used.

[0087] By adopting the above technical solutions, on the one hand, the pretreatment optimization driven by the sample utilization factor U_f can significantly improve the detection rate of metabolites in limited blood samples, ensuring that low-exposure populations can also be effectively monitored; on the other hand, it was found that microcolumn SPE combined with enzymatic digestion can simultaneously improve the total recovery rate of free and bound metabolites, making biomonitoring more comprehensive and enhancing the coverage of metabolic pathways.

[0088] Specifically, the system introduces a sample utilization factor U_f, enhances the effective signal through microcolumn SPE, and dynamically evaluates the sample utilization factor U_f by linking the spiked recovery rate obtained after verification by the comparison and judgment module. This not only improves the detection coverage, but also ensures that the subsequent output results (i.e., C_i and N_i obtained by the standardization and traceability analysis module) are truly representative. It resolves the contradiction between the small amount of blood samples and the extremely low concentration of metabolites among high-altitude aviation personnel, avoids missed detection of low-exposure populations due to preprocessing losses, and enhances the comprehensiveness of the overall detection.

[0089] V. Standardization and Traceability Analysis Module:

[0090] The concentration data of target aromatic POPs and metabolites are standardized to obtain a standardized index. After comparison, different levels of exposure are identified, a multiple regression model is constructed, and the least squares method is used for fitting. After significance testing, key influencing factors are determined. If the goodness of fit of the multiple regression model is less than the preset standard threshold, the variable supplementation mechanism is triggered. After the test is completed, the analysis results are output.

[0091] The basis for completing the standardization process is as follows:

[0092] To eliminate systematic errors caused by different sampling equipment, batches, and regions, the raw data is uniformly expressed and standardized. Specifically, for PM2.5 samples: the absolute mass of the filter extract (in ng) is divided by the sampling volume (in m³) to convert it to ng / m³. For plasma samples: metabolite concentration is directly expressed as pg / mL, as plasma is a homogeneous system and therefore no volume correction is required. The measured concentration of the target analyte in any sample is denoted as C_i, and the corresponding reference background value is denoted as C_ref. The ratio between the two, C_i / C_ref, is denoted as the standardization index N_i, where i represents the number of any target analyte. Similar to the target compounds mentioned above, they can be considered as the same. The corresponding reference background value can be selected from: the annual average of urban ambient air quality across the country, the average of non-airport control points in plateau regions, or the WHO health guidance value. It should be noted that the core function of the standardization index N_i is to serve as a unified scale for subsequent statistics or risk classification.

[0093] The process of identifying different levels of exposure after comparison is explained below:

[0094] If the standardized index N_i exceeds the set value, for example, N_i > 1, it is defined as a high exposure level; otherwise, it is defined as a low exposure level. If the distribution of the standardized index N_i of any batch of samples shows an excessive deviation, i.e., the increase exceeds 50%, a feedback adjustment mechanism is executed: the failure degree of the microcolumn SPE used in the improvement strategy is checked, and regression correction is completed. For example, if the failure degree of the microcolumn SPE is found to be low, i.e., the elution solvent is incorrectly prepared or the acetonitrile water content exceeds the standard, the data can be corrected and returned to normal; if the failure degree of the microcolumn SPE is found to be high, it is replaced.

[0095] The basis for constructing the multiple regression model is: E = β_0 + β_1 × Th + β_2 × Hr + β_3 × An;

[0096] To analyze the driving factors of exposure, E in the formula represents the individual exposure level, expressed as 1-OH-Pyrene pg / mL in plasma; β_0 represents the intercept term, and β_1, β_2, and β_3 are partial regression coefficients reflecting the independent contributions of each factor to exposure; Th represents the average daily outdoor working time in hours, derived from the scheduling system; Hr represents the altitude of the airport in kilometers; and An represents the length of employment in days. The least squares method was used for fitting, and key influencing factors were determined after significance testing. The significance test criterion was: if a variable includes any one of the following—average daily outdoor working time, airport altitude, and length of employment—and the p-value corresponding to the partial regression coefficient β_2 of the airport altitude Hr is <0.05, it indicates that the variable has a statistically significant impact on the exposure level E. The key influencing factor determined is the airport altitude Hr, i.e., the dominant exposure source. It should be noted that the multivariate regression model constructed above is designed for individuals under high exposure levels.

[0097] The triggered variable replenishment mechanism is as follows:

[0098] When the goodness of fit of the multiple regression model is less than the preset standard threshold, it indicates that the model's explanatory power is insufficient and there are important missing variables. At this time, the system will automatically prompt to supplement potential covariates, including at least: wind speed Wq, daily average flight density Fm, and personal protective equipment usage rate Pt. The completed test action is as follows: add the potential covariates to the multiple regression model in sequence to complete the test. The potential covariates that meet the conditions that the goodness of fit of the multiple regression model is not less than the preset standard threshold and the corresponding p value is <0.05 are identified as latent exposure sources. The output analysis results are: explicit exposure source + latent exposure source, so as to generate subsequent warnings.

[0099] Specifically, the aforementioned standardization and source tracing analysis modules do not operate in isolation, but are deeply coupled with the data output from the preceding modules. This provides high-sensitivity detection capabilities, ensuring that low-concentration metabolites can be detected and avoiding underestimation of exposure due to non-detection. Method recovery and precision validation ensure the reliability of C_i data and prevent spurious correlations. Furthermore, the sample utilization factor U_f ensures strong representativeness of individual exposure data and reduces noise interference. Finally, the standardization index N_i and unified units make data from different individuals and airports comparable, supporting cross-group modeling and ultimately enabling effective identification of hidden exposure sources. Only with the groundwork laid by the preceding modules can high-quality, highly representative, and highly sensitive data be ensured, thus achieving a true reflection of the exposure-factor relationship. Otherwise, hidden exposure sources may simply be illusions caused by measurement errors.

[0100] By adopting the above technical solution, and through the hidden exposure source identification mechanism driven by the standardized index N_i and goodness of fit, plateau-specific hidden exposure sources can be successfully identified, and operational scheduling can be optimized in reverse. This effect relies on the high-quality and highly representative data provided in the first two stages. The three form an inseparable system from accurate measurement to efficient utilization and then to intelligent attribution, breaking through the limitations of traditional exposure assessment, realizing the leap from static description to dynamic source tracing, and achieving closed-loop governance from health monitoring to operational intervention.

[0101] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A risk assessment and monitoring system for metabolic products of aviation personnel operating in high-altitude areas, characterized in that, The system includes an initialization measurement module: a standard substance is selected as the development benchmark, spiked samples are configured and summarized to form a standard sample, and the standard sample is injected into a pre-constructed GC-MS platform to perform initial condition testing, obtain the characteristic ion peak spectrum of the target compound, run a predefined traversal test, obtain the signal-to-noise ratio-resolution coordinating function running result after each traversal test, extract the parameter combination in the corresponding traversal test that maximizes the running result, and obtain the optimal parameter combination; Online separation and refinement module: Online liquid chromatography (OLLC) is used as the online separation unit. After gradient elution of the target column, interfering substances other than the target compound are retained on the target column. A dynamic matching window mechanism is introduced to determine the time synchronization index and execute the target control strategy according to the sub-constraint conditions. Comparison and Validation Module: The constructed Online LC-GC-MS method is standardized and compared, a performance indicator dataset is obtained, the rule engine is run, and a judgment action is performed on the performance indicator dataset. If any indicator in the performance indicator dataset fails to meet the standard, a correction strategy for the benchmark is completed. Application Analysis Module: Completes application processing of the validated Online LC-GC-MS method, introduces a sample utilization factor during the application processing, determines whether the current application processing needs to be optimized, and if so, runs the defined improvement strategy; Standardization and source tracing analysis module: Standardizes the concentration data of target aromatic POPs and metabolites, derives the standardized index, and identifies different levels of exposure after comparison. Constructs a multiple regression model, fits it using the least squares method, and determines key influencing factors after significance testing. If the goodness of fit of the multiple regression model is less than the preset standard threshold, a variable supplementation mechanism is triggered. After completing the test, the analysis results are output.

2. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 1, characterized in that, The process of running a predefined traversal test is as follows: Key performance indicators are extracted from characteristic ion peak spectra, including at least the signal-to-noise ratio (S / N) and the theoretical plate number (R). The signal-to-noise ratio-resolution coordinating function is run based on the key performance indicators. The key performance indicators are accumulated and multiplied to obtain the required running result, which is the coordinating function value F_SR. Different preset parameter combinations are tested; the parameter combinations include column temperature program, carrier gas flow rate and injection port temperature; if there are multiple target compounds, the cooperative function value F_SR of each target compound is calculated after each run, and the average value is taken as the final cooperative function value F_SR. The parameter combination that maximizes the final cooperative function value F_SR is extracted to obtain the optimal parameter combination.

3. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 1, characterized in that, The dynamic matching window mechanism operates as follows: The elution time of the target compound in LC and the preparation time of GC are obtained, and the result of dividing the elution time of the target compound in LC by the preparation time of GC is denoted as the time synchronization index T_s. The time synchronization index T_s is compared with the preset sub-constraint condition, which is: the time synchronization index T_s ∈ [T_mo1, T_mo2]. When the time synchronization index T_s < T_mo1, the target regulation strategy is executed to delay the GC start time; when the time synchronization index T_s > T_mo2, the target regulation strategy is executed to accelerate the LC elution speed. Where [T_mo1, T_mo2] represent the tolerance interval.

4. The system for risk assessment and monitoring of metabolites of aviation personnel operating in high-altitude areas according to claim 3, characterized in that, When the time synchronization index T_s < T_mo1, the target control strategy is as follows: the absolute value of the difference between the elution time of the target compound in LC and the time when GC is ready to receive is recorded as the delay time Δt, and the GC heating program is triggered again after the delay time Δt after LC starts running. When the time synchronization index T_s > T_mo2, the target control strategy is as follows: set the maximum allowable LC elution time, combine the elution time of the target compound in LC as input, run the pre-constructed flow rate increase ratio derivation model, and output the percentage of flow rate increase Δ%. The process of the flow rate increase ratio derivation model is as follows: obtain the maximum allowable LC elution time t_target: t_target = T_mo2 × t_GC; where t_GC is the time for GC to prepare for reception; then, divide the elution time of the target compound in LC by the maximum allowable LC elution time to obtain the first ratio, and then subtract 1 from the first ratio to generate the flow rate increase Δ.

5. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 1, characterized in that, The standardized comparative verification process is as follows: Prepare a uniform spiking matrix and prepare a mixed standard solution containing the target analyte; set up Q concentration gradients, prepare 3 parallel samples for each concentration, and run the Online LC-GC-MS method and the current standard method in parallel; For each target analyte, calculate the linear correlation coefficient, limit of detection (LOD), spiked recovery rate, and relative standard deviation (RSD). The performance index dataset should include at least the following: linear correlation coefficient, LOD, spiked recovery rate, and relative standard deviation.

6. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 5, characterized in that, The process of performing the judgment action on the performance metric dataset is as follows: If the linear correlation coefficient does not meet the standard, the correction strategy is as follows: check whether the high concentration point has saturated response, and use weighted least squares (WLS) fitting instead; if the detection limit does not meet the standard, the correction strategy is as follows: switch GC-MS to MRM mode, optimize the LC-GC interface temperature until no adsorption occurs, and increase the PTV injection volume according to the preset gradient; if the spiked recovery rate does not meet the standard, the correction strategy is as follows: introduce an isotope internal standard to correct the matrix effect, add an enzymatic digestion step to the blood sample, and use a matrix-matched calibration curve instead; if the relative standard deviation does not meet the standard, the correction strategy is as follows: replace the LC column or regenerate it.

7. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 1, characterized in that, The criteria for determining whether the current application processing needs optimization are as follows: Define the sample utilization factor U_f: U_f = V_used / V_total × R_recovery / 100; In the formula, V_used represents the equivalent volume actually entering the instrument for analysis, and V_total represents the total plasma collection volume; R_recovery represents the spiked recovery rate obtained after verification by the comparison and judgment module; when the sample utilization factor U_f > the system set threshold, it is determined that the current application processing does not need optimization; otherwise, it is determined that the current application processing needs optimization, and the improvement strategy is to switch to microcolumn SPE.

8. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 7, characterized in that, The basis for identifying different levels of exposure after obtaining the standardized index and comparing them is as follows: the measured concentration of the target substance in any sample is denoted as C_i, the corresponding reference background value is denoted as C_ref, and the ratio between the two, C_i / C_ref, is denoted as the standardized index N_i; if the standardized index N_i exceeds the set value, it is defined as a high exposure level; otherwise, it is defined as a low exposure level; if the distribution of the standardized index N_i of any batch of samples shows an excessive deviation, a feedback adjustment mechanism is executed: the failure degree of the micro-column SPE used in the improvement strategy is checked, and regression correction is completed.

9. The system for risk assessment and monitoring of metabolic products of high-altitude aviation personnel according to claim 1, characterized in that, The basis for constructing the multiple regression model is: E = β_0 + β_1 × Th + β_2 × Hr + β_3 × An; In the formula, E represents the individual exposure level, β_0 represents the intercept term, β_1, β_2, and β_3 are all partial regression coefficients, Th represents the average daily outdoor working time, Hr represents the altitude of the airport, and An represents the length of employment. The least squares method is used for fitting, and the key influencing factors are determined after significance testing. The criterion for significance testing is: the p-value corresponding to the partial regression coefficient of any variable is <0.05, and the key influencing factor is the corresponding variable.

10. The system for risk assessment and monitoring of metabolites of high-altitude aviation personnel according to claim 9, characterized in that, The triggered variable replenishment mechanism is as follows: When the goodness of fit of the multiple regression model is less than the preset standard threshold, it prompts the addition of potential covariates, including at least: wind speed Wq, daily average flight density Fm, and personal protective equipment usage rate Pt. The completed test action is: add the potential covariates to the multiple regression model in sequence to complete the test. The potential covariates that meet the conditions that the goodness of fit of the multiple regression model is not less than the preset standard threshold and the corresponding p value is <0.05 are identified as latent exposure sources.

Images