BMI grouping and detection time point personalized optimization method and system for noninvasive prenatal detection of male fetuses

Abstract

The invention discloses a BMI grouping and detection time point personalized optimization method and system for noninvasive prenatal detection of male fetuses, and the method comprises the steps: obtaining the original data of male fetuses and pregnant women, and carrying out the preprocessing of the original data of male fetuses and pregnant women, and obtaining the preprocessed data; based on the preprocessed data, setting a dynamic weight coefficient, a time risk score and a chromosome concentration standard-reaching rate, constructing a comprehensive risk function, and defining a detection error constraint at the same time; carrying out mixed variable coding on a BMI grouping boundary and a detection time point, setting a core parameter of a simulated annealing algorithm, generating a new solution through neighborhood disturbance, receiving the solution according to a Metropolis criterion, and carrying out iteration until a convergence condition is met to obtain a preliminary optimization result; and based on the accuracy constraint, the error constraint, the sample size balance constraint and the clinical rationality constraint, screening an optimal BMI group and a scheme corresponding to a detection time point to obtain a personalized optimization scheme.

Description

Technical Field

[0001] This invention belongs to the field of non-invasive prenatal testing technology, specifically relating to a method and system for personalized optimization of BMI grouping and testing time points in non-invasive prenatal testing of male fetuses. Background Technology

[0002] NIPT (Non-invasive Prenatal Test) is a prenatal testing technique that involves collecting maternal blood, detecting cell-free fetal DNA fragments (including sex chromosomes), and analyzing for chromosomal abnormalities. Its purpose is to determine the fetal health status early. Typically, fetal sex chromosome concentration can be measured between 10 and 25 weeks of gestation. If the Y chromosome concentration in male fetuses reaches or exceeds 4%, and the X chromosome concentration in female fetuses is normal, the NIPT result is considered generally accurate; otherwise, the accuracy cannot be guaranteed. In practice, it is crucial to detect unhealthy fetuses as early as possible, otherwise, the treatment window may be shortened. Early detection (within 12 weeks) carries a lower risk; mid-term detection (13-27 weeks) carries a higher risk; and late detection (after 28 weeks) carries an extremely high risk. Pregnant women with high BMI experience a blood dilution effect, resulting in a later attainment of the target fetal cell-free DNA concentration (≥4%), while pregnant women with low BMI can be detected earlier to ensure the treatment window is met. Currently, using simple empirical grouping and a uniform testing time point for NIPT in all pregnant women can significantly impact testing accuracy and may lead to late detection of fetal abnormalities, shortening the treatment window and posing potential risks. Therefore, two core issues need to be addressed: first, to rationally group pregnant women with male fetuses by BMI and determine the BMI range for each group; and second, to match the optimal NIPT time point for each group, ultimately minimizing the potential risks to pregnant women. Summary of the Invention

[0003] This invention aims to address the shortcomings of existing technologies and provides the following solutions: A method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing for male fetuses includes the following steps: Obtain the raw data of pregnant women with male fetuses, and preprocess the raw data to obtain the preprocessed data; Based on the preprocessed data, dynamic weighting coefficients, time risk scores, and chromosome concentration compliance rates are set to construct a comprehensive risk function, while defining detection error constraints. The BMI grouping boundary and detection time point are encoded with mixed variables. The core parameters of the simulated annealing algorithm are set. New solutions are generated through neighborhood perturbation. Solutions are accepted according to the Metropolis criterion. The process is iterated until the convergence condition is met to obtain preliminary optimization results. Based on constraints of accuracy, error, sample size balance, and clinical rationality, the optimal BMI grouping and corresponding testing time points were selected to obtain a personalized optimization plan.

[0004] Preferably, the original data of the pregnant woman with a male fetus includes: the pregnant woman's BMI value, gestational age and Y chromosome concentration.

[0005] Preferred methods for obtaining preprocessed data include: Based on the original data of pregnant women with male fetuses, abnormal samples were removed by preset screening rules to obtain the detection data after removal. The data after removal were aggregated, and the median was used to calculate the effective value. Y Chromosome concentrations were calculated using arithmetic mean aggregation to represent gestational age, resulting in aggregated data. The aggregated data is standardized to eliminate dimensions, resulting in preprocessed data.

[0006] The preferred comprehensive risk function is: in, S i,t This represents the overall risk score. α i Indicates dynamic weighting coefficients. A i,t Indicates a risk to accuracy. T t Indicates the risk score at a given point in time. R i,t This indicates the rate at which chromosome concentration meets the target. m i,t Indicates the first i Group at gestational age t hour Y Number of samples with chromosome concentration ≥4%, n i Indicates the first i Total number of samples in the group.

[0007] Preferably, the detection error constraint is: in, E i Indicates the detection error. σ Y,i Indicates the first i Group concentration standard deviation Yi,j Indicates the first i In the BMI group, the first j Valid samples Y Chromosome concentration, Indicates the first i Group Y Mean chromosome concentration.

[0008] Preferably, the method for generating new solutions through neighborhood perturbation includes: locally perturbing the current solution based on BMI boundary perturbation and time-point perturbation to generate new solutions. X’ ; The BMI boundary perturbation is: randomly selecting one boundary. b j ( j =1,2,3,4), according to the formula Update, verify after update ( b 0 = 18.5 b 5=47.0); The time-point perturbation is: randomly selecting one time point t. i (i=1~5), according to the formula Update, verify after update .

[0009] The preferred Metropolis criterion is: Calculate the sum of the risk scores between the new solution and the current solution. New solutions will be accepted according to the following rules: like Then the new solution will be accepted directly. X' ; like Then, based on probability P Accepted, the formula is: ,in, T This indicates the current temperature in the simulated annealing algorithm.

[0010] Preferably, the accuracy constraint is: The error constraint is: The sample size balance constraint is: The clinical rationale constraint is: the higher the BMI, the better the optimal time point. The later it gets.

[0011] The present invention also provides a BMI grouping and personalized optimization system for non-invasive prenatal testing of male fetuses, the system applying the above-mentioned method, including: a data preprocessing module, a function construction module, a preliminary optimization module and a screening optimization module; The data preprocessing module is used to acquire the raw data of pregnant women with male fetuses and to preprocess the raw data to obtain the preprocessed data. The function construction module constructs a comprehensive risk function based on the preprocessed data, setting dynamic weight coefficients, time risk scores, and chromosome concentration compliance rates, while defining detection error constraints. The preliminary optimization module is used to encode mixed variables for BMI grouping boundaries and detection time points, set the core parameters of the simulated annealing algorithm, generate new solutions through neighborhood perturbation, accept solutions according to the Metropolis criterion, iterate until the convergence condition is met, and obtain the preliminary optimization results. The screening and optimization module selects the optimal BMI grouping and corresponding testing time point based on accuracy constraints, error constraints, sample size balance constraints, and clinical rationality constraints, thus obtaining a personalized optimization plan.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This application adopts dynamic weighting coefficients and a comprehensive risk function to achieve a personalized dynamic balance between "accuracy and time risk" and adapt to the detection needs of different BMI groups. (2) This application uses simulated annealing algorithm to perform mixed variable synchronous optimization on the BMI group boundary and detection time point to avoid local optima and achieve global optima in the optimization results. (3) This application introduces a relative detection error constraint (E i ≤65%), construct a three-in-one optimization system of "compliance rate - time risk - detection stability" to improve detection reliability. (4) This application is for high BMI group (BMI≥36 kg / m 2 The detection compliance rate in the test was increased to over 80.9%, the average comprehensive risk was reduced by 34.9%, and the risk of misjudgment was significantly reduced. (5) The optimization results of this application are highly stable, and the boundary deviation is <0.5 kg / m when the initial grouping changes. 2 The time deviation is less than 0.3 weeks, which is suitable for diverse clinical sample distribution scenarios. Attached Figure Description

[0013] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0015] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] Example 1 In this embodiment, as Figure 1 As shown, a method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing for male fetuses includes the following steps: S1. Obtain the raw data of the pregnant woman with a male fetus, and preprocess the raw data to obtain the preprocessed data.

[0018] The raw data for pregnant women with male fetuses includes: the pregnant woman's BMI, gestational age, and Y chromosome concentration. The methods for obtaining the preprocessed data include: based on the raw data of pregnant women with male fetuses, pre-defined screening rules are used to remove abnormal samples, resulting in the removed detection data; the removed detection data is then aggregated, and the median aggregation is used to calculate the effective value. Y Chromosome concentrations were aggregated using arithmetic mean to represent gestational age, resulting in aggregated data. The aggregated data were then standardized to eliminate dimensions, yielding preprocessed data.

[0019] In this embodiment: S11. Sample Selection and Inclusion Criteria: (1) Sample source: 1300 NIPT test samples from pregnant women with male fetuses were received. All samples met the clinical ethical standards.

[0020] (2) Inclusion criteria: singleton pregnancy, no history of chromosomal abnormalities, no pregnancy complications, gestational age at testing between 10.0 and 25.0 weeks (fetal sex chromosome concentration can usually be tested between 10 and 25 weeks of gestation), and BMI between 18.5 and 47.0 kg / m². 2 Between (excluding extreme BMI values ​​to improve universality).

[0021] (3) Data dimensions: The core data collected for each sample includes: the pregnant woman's BMI value (accurate to 0.1 kg / m²). 2 ), gestational age (accurate to 0.1 weeks, e.g., 11.2 weeks means 11 weeks and 1 day), and Y chromosome concentration (unit: %).

[0022] S12. Outlier Removal: Filter valid samples and remove invalid data according to quantification rules. The formula and operation are as follows: (1) Gestational age exclusion rule: Samples with a gestational age <10 weeks or >25 weeks are tested. The formula is as follows: Exclusion criteria: A total of 65 samples with abnormal gestational age were excluded (35 cases <10 weeks, 30 cases >25 weeks). (2) BMI exclusion rule: BMI < 18.5 kg / m 2 Or BMI > 47.0 kg / m 2 The formula is expressed as: Elimination conditions: A total of 28 samples with abnormal BMI (<18.5 kg / m²) were excluded. 2 18 cases, >47.0 kg / m 2 10 cases); (3) Abnormal concentration rejection rules: Y chromosome concentration (considered as detection failure), the formula is expressed as: rejection condition: Y < 0.5%; a total of 52 abnormal concentration samples were rejected; Final effective sample size: 1300 - 65 - 28 - 52 = 1155 cases.

[0023] S13. Aggregation of Multiple Test Data: For multiple test data of the same pregnant woman (a total of 148 cases, including 93 cases of multiple blood sampling tests and 55 cases of multiple tests from a single blood sampling test), the data is aggregated according to the following formula to avoid duplication and interference: S131. Calculation of effective Y chromosome concentration: Median aggregation (better against extreme values ​​than the mean) is used, and the formula is as follows: in, k Number of tests ( k ≥2), Y 1~ Y k For each test Y Chromosome concentration (unit: %) S132. Gestational age calculation: Arithmetic mean aggregation is used, the formula is: in, t i The gestational week for each test (unit: weeks).

[0024] S14. Data Standardization: Standardize the ranges for gestational age and BMI (for algorithm input only, without changing clinical significance) to eliminate dimensional differences. The formula is: Specifically, (1) standardization of gestational age: xmin =10.0、 x max =25.0, after standardization the value range [0,1]; (2) BMI standardization: x min =18.5、 x max =47.0, and the standardized value range is [0,1].

[0025] S2. Based on the preprocessed data, set dynamic weight coefficients, time risk scores, and chromosome concentration compliance rates to construct a comprehensive risk function, while defining detection error constraints.

[0026] In this embodiment, the core parameters are first defined: Y Chromosomal concentration target achievement rate ( R ): A certain BMI group at gestational age t hour, Y The proportion of samples with a chromosome concentration ≥4% to the total number of samples in the group. R The higher the value, the stronger the detection accuracy and the lower the accuracy risk; constraints R ≥80% (lower limit of clinically acceptable accuracy to avoid misdiagnosis).

[0027] Time risk score ( T (1) 10-12 weeks (early stage): T=0.1; (2) 13-17 weeks (early mid-stage): T=0.3; (3) 18-25 weeks (late mid-stage): T=0.6. Quantifying the risk of "early or late detection point", T The smaller the diameter, the longer the treatment window and the lower the time risk.

[0028] Accuracy Risk Score ( A ): A =1- R . A The smaller the value, the lower the risk to accuracy (e.g., A=0.1 when R=90%, A=0.3 when R=70%), which directly relates to the reliability of the detection.

[0029] Weighting coefficients ( α (1) BMI < 28: α =0.5 (balance); (2) 28≤BMI<32: α =0.6 (slight emphasis); (3) 32≤BMI<36: α =0.7 (emphasis); (4) 36≤BMI<40: α =0.8 (high emphasis); (5) BMI≥40: α =0.9 (extremely high emphasis). (Passed) αAdjust the weights of the two objectives to ensure that clinical priorities are reasonable.

[0030] S21. Dynamic weighting coefficient α i (BMI Grouping Weights): The weights for "accuracy risk" and "time risk" are dynamically allocated based on the BMI range. The higher the BMI, the greater the accuracy weight. The formula and rules are as follows: in, α i For the first i The weighting coefficients for each BMI group (0 < α i <1), α i The larger the value, the more priority should be given to ensuring detection accuracy during optimization.

[0031] S22. Risk score at specific gestational ages: Quantifying the risk of the treatment window by gestational week segment, with the lowest risk in early detection and the highest risk in mid-to-late gestation, as shown in the formula below: in, T t To detect gestational age (unit: weeks), the value range is [0.1, 0.6]. The higher the value, the higher the risk of gestational age.

[0032] S23. Compliance Rate (Accuracy Quantitative Indicator): [Item] i The percentage of samples in each BMI group with a Y chromosome concentration ≥4% (clinical accuracy threshold) at gestational age is calculated using the following formula: in, m i,t Indicates the first i Group at gestational age t hour Y Number of samples with chromosome concentration ≥4%, n i Indicates the first i Total number of samples in the group.

[0033] S24. Comprehensive Risk Score Function (Core Optimization Objective): Integrating "Accuracy Risk" and "Time Risk," a single optimization objective function is constructed, with the following formula: in, S i,t This represents the overall risk score. α i Indicates dynamic weighting coefficients. A i,t Indicates a risk to accuracy. T t Indicates the risk score at a given point in time. R i,t Indicates the rate at which chromosome concentration targets are met; optimization objective To minimize the overall risk score; S25. Detection Error (Reliability Constraint): [The following text appears to be incomplete and requires further context:] i The relative detection error of each BMI group reflects the dispersion of the concentration data. The formula for the detection error constraint is: in, E i Indicates the detection error. σ Y,i Indicates the first i Group concentration standard deviation Y i,j Indicates the first i In the BMI group, the first j Valid samples Y Chromosome concentration, Indicates the first i Group Y Mean chromosome concentration.

[0034] S3. Encode the BMI grouping boundary and detection time point with mixed variables, set the core parameters of the simulated annealing algorithm, generate new solutions through neighborhood perturbation, accept solutions according to the Metropolis criterion, iterate until the convergence condition is met, and obtain the preliminary optimization results.

[0035] In this embodiment: S31. Optimize variable encoding: Integrate the "4 BMI grouping boundaries + 5 detection time points" into a mixed variable vector (containing discrete and continuous variables), and the encoding formula is as follows: ; (1) Grouping boundary constraints: This corresponds to 5 consecutive BMI groups: ; (2) Time constraints: ( t 1 t 5 represents the optimal time point for each group (unit: week).

[0036] S32. Initial Solution Generation: Generate an initial solution based on routine clinical grouping logic to ensure algorithm convergence efficiency. (1) Initial BMI grouping boundaries: b 1 = 25.98 b 2 = 31.26 b 3 = 36.54 b 4 = 41.82, corresponding to the initial groups: [18.5, 25.98), [25.98, 31.26), [31.26, 36.54), [36.54, 41.82), [41.82, 47.0]; (2) Initial time point calculation: determined according to the "earliest gestational age reached" average for each group, using the following formula: in, Indicates the first i Number of samples with a group concentration ≥ 4%, t i,j This represents the gestational week of the corresponding sample.

[0037] S33. Simulated Annealing Core Parameter Settings: The optimal parameters were determined through 50 sets of preliminary experiments, balancing search capability and convergence efficiency. The parameters are shown in Table 1. Table 1 S34. Neighborhood Solution Generation Rules: Methods for generating new solutions through neighborhood perturbations include: locally perturbing the current solution based on BMI boundary perturbations and time-point perturbations to generate new solutions. X' .

[0038] The BMI boundary perturbation is: randomly selecting one boundary. b j ( j =1,2,3,4), according to the formula Update, verify after update ( b 0 = 18.5 b 5=47.0); The time-point perturbation is: randomly selecting one time point t. i (i=1~5), according to the formula Update, verify after update .

[0039] S35. Metropolis Criterion (Solution Acceptance Rule): Calculate the difference in the sum of risk scores between the new solution and the current solution. ,in, The formula for the overall risk mean of all groups is: New solutions will be accepted according to the following rules: like Then the new solution will be accepted directly. X’ ; like Then, based on probability P Accepted, the formula is: ,in, T This indicates the current temperature in the simulated annealing algorithm.

[0040] S36. Convergence Criteria: The algorithm terminates and outputs the optimal solution when both of the following conditions are met: (1) Current temperature T ≤ T end =10 -6 ; (2) The change in the optimal comprehensive risk score after 10 consecutive temperature iterations ( (This is the difference between the optimal solution in each iteration and the solution in the previous iteration).

[0041] S4. Based on accuracy constraints, error constraints, sample size balance constraints, and clinical rationality constraints, the optimal BMI grouping and corresponding testing time points are selected to obtain a personalized optimization plan.

[0042] In this embodiment: S41. Optimal treatment plan selection criteria: Select treatment plans that meet all of the following constraints to ensure clinical applicability: (1) The accuracy constraint is: ; (2) The error constraint is: ; (3) The sample size balance constraint is: ; (4) The clinical rationale constraint is: the higher the BMI, the better the optimal time point. The later it gets.

[0043] S42. Final optimization results: After 220 iterations, the optimal solution is shown in Table 2: Table 2 S43. Stability Verification: Change the initial number of groups (4 groups, 6 groups), repeat the S3 optimization process, and calculate the boundary and time point deviations, as shown in the following formula: in, , The boundary conditions and time points for the five schemes are shown in Table 3. Table 3 All deviations meet the requirement of "boundary deviation < 0.5 kg / m". 2 "Time deviation < 0.3 weeks" proves that the scheme has good stability.

[0044] Technical effect comparison experiment: 1. Comparison scheme setting: The existing clinical "one-size-fits-all" scheme was selected as the comparison: BMI grouping was [20,28), [28,32), [32,36), [36,40), [40,∞), and the testing time point was uniformly 12 weeks.

[0045] 2. Evaluation indicators and formulas: (1) Average compliance rate: ; (2) Average composite risk: ; (3) High BMI group (BMI≥36.0kg / m²) 2 The rate of improvement in compliance: .

[0046] 3. Comparison results: As shown in Table 4: Table 4 The results show that the proposed solution is significantly better than the existing solutions in terms of both accuracy and risk control, and the technical performance is significantly improved.

[0047] Example 2 In this embodiment, a BMI grouping and personalized optimization system for non-invasive prenatal testing of male fetuses includes: a data preprocessing module, a function construction module, a preliminary optimization module, and a screening optimization module.

[0048] The data preprocessing module acquires and preprocesses the raw data of pregnant women with male fetuses to obtain preprocessed data. The function construction module constructs a comprehensive risk function based on the preprocessed data, setting dynamic weight coefficients, time risk scores, and chromosome concentration attainment rates, while defining detection error constraints. The preliminary optimization module encodes mixed variables for BMI group boundaries and detection time points, sets the core parameters of the simulated annealing algorithm, generates new solutions through neighborhood perturbation, accepts solutions according to the Metropolis criterion, iterates until convergence conditions are met, and obtains preliminary optimization results. The screening and optimization module screens the optimal BMI group and corresponding detection time point scheme based on accuracy constraints, error constraints, sample size balance constraints, and clinical rationality constraints to obtain personalized optimization schemes.

[0049] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for personalized optimization of BMI grouping and testing time points in non-invasive prenatal testing for male fetuses, characterized in that, Includes the following steps: Obtain the raw data of pregnant women with male fetuses, and preprocess the raw data to obtain the preprocessed data; Based on the preprocessed data, dynamic weighting coefficients, time risk scores, and chromosome concentration compliance rates are set to construct a comprehensive risk function, while defining detection error constraints. The BMI grouping boundary and detection time point are encoded with mixed variables. The core parameters of the simulated annealing algorithm are set. New solutions are generated through neighborhood perturbation. Solutions are accepted according to the Metropolis criterion. The process is iterated until the convergence condition is met to obtain preliminary optimization results. Based on constraints of accuracy, error, sample size balance, and clinical rationality, the optimal BMI grouping and corresponding testing time points were selected to obtain a personalized optimization plan.

2. The method for BMI grouping and personalized optimization of testing time points for non-invasive prenatal testing of male fetuses according to claim 1, characterized in that, Raw data for male fetuses include: the pregnant woman's BMI, gestational age, and Y chromosome concentration.

3. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 1, characterized in that, Methods for obtaining preprocessed data include: Based on the original data of pregnant women with male fetuses, abnormal samples were removed by preset screening rules to obtain the detection data after removal. The data after removal were aggregated, and the median was used to calculate the effective value. Y Chromosome concentrations were calculated using arithmetic mean aggregation to represent gestational age, resulting in aggregated data. The aggregated data is standardized to eliminate dimensions, resulting in preprocessed data.

4. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 1, characterized in that, The overall risk function is: in, S i,t This represents the overall risk score. α i Indicates dynamic weighting coefficients. A i,t Indicates a risk to accuracy. T t Indicates the risk score at a given point in time. R i,t This indicates the rate at which chromosome concentration meets the target. m i,t Indicates the first i Group at gestational age t hour Y Number of samples with chromosome concentration ≥4%, n i Indicates the first i Total number of samples in the group.

5. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 4, characterized in that, The detection error constraint is: in, E i Indicates the detection error. σ Y,i Indicates the first i Group concentration standard deviation Y i,j Indicates the first i In the BMI group, the first j Valid samples Y Chromosome concentration, Indicates the first i Group Y Mean chromosome concentration.

6. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 4, characterized in that, Methods for generating new solutions through neighborhood perturbations include: applying local perturbations to the current solution based on BMI boundary perturbations and time-point perturbations to generate new solutions. X’ ; The BMI boundary perturbation is: randomly selecting one boundary. b j ( j =1,2,3,4), according to the formula Update, verify after update ( b 0 = 18.5 b 5=47.0); The time-point perturbation is: randomly selecting one time point t. i (i=1~5), according to the formula Update, verify after update .

7. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 6, characterized in that, The Metropolis guidelines are: Calculate the sum of the risk scores between the new solution and the current solution. New solutions will be accepted according to the following rules: like Then the new solution will be accepted directly. X' ; like Then, based on probability P Accepted, the formula is: ,in, T This indicates the current temperature in the simulated annealing algorithm.

8. The method for personalized optimization of BMI grouping and testing time point in non-invasive prenatal testing of male fetuses according to claim 5, characterized in that, The accuracy constraint is: The error constraint is: The sample size balance constraint is: The clinical rationale constraint is: the higher the BMI, the better the optimal time point. The later it gets.

9. A BMI grouping and personalized optimization system for non-invasive prenatal testing of male fetuses, wherein the system applies the method described in any one of claims 1-8, characterized in that, include: Data preprocessing module, function construction module, preliminary optimization module, and screening optimization module; The data preprocessing module is used to acquire the raw data of pregnant women with male fetuses and to preprocess the raw data to obtain the preprocessed data. The function construction module constructs a comprehensive risk function based on the preprocessed data, setting dynamic weight coefficients, time risk scores, and chromosome concentration compliance rates, while defining detection error constraints. The preliminary optimization module is used to encode mixed variables for BMI grouping boundaries and detection time points, set the core parameters of the simulated annealing algorithm, generate new solutions through neighborhood perturbation, accept solutions according to the Metropolis criterion, iterate until the convergence condition is met, and obtain the preliminary optimization results. The screening and optimization module selects the optimal BMI grouping and corresponding testing time point based on accuracy constraints, error constraints, sample size balance constraints, and clinical rationality constraints, thus obtaining a personalized optimization plan.

Images