Risk assessment and management method for multi-fetal pregnancy

Abstract

The invention discloses a risk assessment and management method for multi-fetal pregnancy, and relates to the technical field of multi-fetal pregnancy risk assessment, and the method comprises the following steps: constructing a time sequence behavior representation tensor of each fetus; based on the time sequence behavior representation tensor of each fetus, extracting a time-dependent growth evolution curve reflecting an individual development trend, so as to construct an inter-fetus coupling coefficient matrix representing a multi-fetal development interaction relationship; taking the growth evolution curve of each fetus and the extracted inter-fetus coupling coefficient matrix as model input for quantifying the development cooperation degree and potential pathological interference risk between different fetuses; adjusting drug administration parameters of the individualized drug regulation and control module through a preset intervention strategy; generating a prediction result for describing the risk level of the individual fetus; outputting a comprehensive health score and a corresponding intervention strategy suggestion; according to the invention, personalized intervention strategies and treatment schemes can be provided for each fetus.

Description

Technical Field

[0001] The present invention relates to the technical field related to multiple pregnancy risk assessment, and specifically to a risk assessment and management method for multiple pregnancy. Background Art

[0002] Multiple pregnancy, abbreviated as multiple, commonly known as multiple births, refers to the presence of more than or equal to two fetuses in the uterine cavity during one pregnancy. In contrast to "single pregnancy", pregnant women with multiple pregnancies have a higher probability of developing complications during pregnancy, such as gestational diabetes, pregnancy-induced hypertension, etc., and may also experience premature rupture of membranes, premature birth, and even intrauterine growth retardation and intrauterine distress of the fetus. In severe cases, it may endanger the life of the fetus. Therefore, the risk of multiple pregnancy is generally greater. If multiple pregnancy occurs, pregnant women need to go to the hospital for regular prenatal check-ups and monitor the growth and development of the fetus in a timely manner.

[0003] The existing multiple pregnancy risk assessment system mainly relies on overall pregnancy risk assessment, but ignores developmental imbalances between fetuses (such as fetal growth restriction) and mutual influences between fetuses (such as coupling effects between fetuses), resulting in the inability to effectively identify subtle differences and potential mutual influences between fetuses (such as the developmental restriction of one fetus may affect another fetus through placental blood flow). This lack of recognition of the coupling-feedback mechanism between multiple fetuses not only affects the accurate assessment of the growth status of each fetus, but also limits the implementation of early warning and personalized intervention strategies. Summary of the Invention

[0004] To address the deficiencies in the prior art, the present invention provides a method for risk assessment and management of multiple pregnancies.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: The present invention provides a method for risk assessment and management of multiple pregnancies, comprising the following steps: Collect multimodal physiological behavior time series data of each fetus in a pregnant woman with multiple pregnancies, standardize the multimodal physiological behavior time series data, and construct a time series behavior representation tensor for each fetus; Based on the temporal behavior representation tensor of each fetus, a time-dependent growth evolution curve reflecting individual developmental trends is extracted. Combined with the coupling relationship between fetal blood flow parameters, the physiological interaction indicators between multiple fetuses are analyzed to construct an inter-fetal coupling coefficient matrix that characterizes the developmental interactions among multiple fetuses. The growth evolution curve of each fetus and the extracted inter-fetal coupling coefficient matrix are used as model inputs and embedded into a multi-objective joint modeling structure based on time series modeling to quantify the degree of developmental coordination and potential pathological interference risks between different fetuses. Based on the interactive prediction indicators output by the multi-objective joint modeling structure, abnormal trends in multiple pregnancies are identified. Based on the tensor state and development trend of each fetus's current temporal behavior, the dosing parameters of the personalized drug regulation module are adjusted through preset intervention strategies. The tensor representing the fetus's temporal behavior throughout the gestational cycle is transmitted in real time to the cloud monitoring platform via edge acquisition devices with data upload capabilities. The artificial intelligence analysis model deployed on the cloud monitoring platform models and processes the data to generate prediction results describing the individual fetal risk level; The inter-fetal coupling coefficient matrix, the drug administration parameters of the drug regulation module, and the predicted results of the individual fetal risk level are taken as joint input and imported into the multi-parameter fusion analysis system. The feature fusion and state mapping methods are used to output the comprehensive health score and corresponding intervention strategy recommendations.

[0006] As a preferred technical solution of the present invention, the multimodal physiological behavior time series data includes fetal structural development data obtained by an ultrasonic imaging system, umbilical artery and venous blood flow rate data obtained by a placental Doppler blood flow monitoring system, and fetal electrocardiogram or electroencephalogram signal data obtained by a fetal bioelectric monitoring device.

[0007] As a preferred technical solution of the present invention, the time-dependent growth evolution curve is defined as a continuous time function of time and fetal physiological behavior state vector, and a personalized growth evolution function is constructed based on fetal structural development data, umbilical artery and venous blood flow rate data, fetal electrocardiogram or electroencephalogram signal data and time factors.

[0008] As a preferred technical solution of the present invention, the coupling effect between multiple fetuses is modeled by the functional relationship between the individual growth evolution function of each fetus and the inter-fetal coupling coefficient matrix. The specific steps include: Modeling of the interaction coupling effect: for each pair of fetuses, the interaction coupling effect between the fetuses is calculated using their respective growth evolution functions and the inter-fetal coupling coefficient matrix; Responsiveness measurement, through the model to calculate the interactive coupling responsiveness between fetuses, measures the transmission intensity and mutual influence of physiological resources between fetuses.

[0009] As a preferred technical solution of the present invention, the drug administration parameters of the drug control module are dynamically adjusted according to the developmental status of the fetus and real-time feedback. The adjustment rules of the drug administration parameters are controlled by the feedback regulation function. The specific steps include: Dynamic control mechanism, setting the feedback mechanism of drug release dosage and adjusting the drug dosage in real time according to the physiological state of the fetus; The intervention trigger time and feedback rate, the release of drug dosage is related to the intervention trigger time and feedback rate, and the drug administration plan is adjusted according to the model output.

[0010] As a preferred technical solution of the present invention, the prediction result of the fetal risk level adopts a weighted sum model, and the specific steps include: Integration of physiological and behavioral indicators, combining the physiological and behavioral time series data of each fetus; Weighted summation calculation, based on the weights trained in historical data, performs weighted summation on the physiological behavior time series data of each fetus to obtain a comprehensive risk assessment value.

[0011] As a preferred technical solution of the present invention, the generation of the comprehensive health score is achieved by the following steps: The time-dependent growth evolution curve of each fetus, the inter-fetal coupling coefficient matrix, medication parameters and the prediction results of fetal risk level are integrated; Feature normalization or tensor fusion methods are used for integration to ultimately generate a comprehensive health score for each fetus.

[0012] As a preferred technical solution of the present invention, the intervention strategy recommendation includes an independent drug administration route for single fetuses, risk warning information provided to medical personnel, and dynamic ranking recommendations for fetal monitoring priorities.

[0013] The beneficial effects of the present invention are: This invention comprehensively analyzes the time-series data of fetal physiological behavior to accurately identify the risks in multiple pregnancies and the mutual influence between fetuses, and can provide personalized intervention strategies and treatment plans for each fetus. The real-time prediction and multi-parameter fusion analysis of artificial intelligence not only improves the accuracy of risk assessment, but also realizes dynamic monitoring and precise intervention of fetal development status, bringing innovative solutions to the management of multiple pregnancies. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 The figure is a flow chart of the risk assessment and management method for multiple pregnancies according to the present invention. DETAILED DESCRIPTION

[0015] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0016] Example 1 like Figure 1 As shown, a method for risk assessment and management of multiple pregnancies includes the following steps: Collect multimodal physiological behavior time series data from each fetus in a pregnant woman with multiple pregnancies. This data covers multiple dimensions, such as fetal weight estimation, heart rate volatility, and blood flow velocity. These multimodal physiological behavior time series data are then standardized to construct a time series behavior tensor representing each fetus. This time series data has important physiological significance, reflecting the growth and development trends of the fetus at different stages of pregnancy. By continuously monitoring the physiological data of the fetus, doctors can understand the developmental status of the fetus and its physiological interactions with other fetuses in real time. For example, when a fetus experiences intrauterine growth restriction (IUGR), will it affect the growth and development of other fetuses through changes in placental blood flow?

[0017] Based on the temporal behavior representation tensor of each fetus, we extract a time-dependent growth evolution curve reflecting individual developmental trends. We then analyze physiological interaction indicators between multiple fetuses by combining the coupling relationship between fetal blood flow parameters. We further explore possible temporal correlations between fetuses at different developmental stages, thereby constructing an inter-fetal coupling coefficient matrix that characterizes the developmental interactions among multiple fetuses. For fetuses at risk of developmental abnormalities, the time-dependent growth evolution curve can help doctors predict the future growth trend of the fetus and promptly identify potential developmental delays or abnormalities. Doctors can then formulate appropriate treatment plans based on the predicted results for fetal developmental abnormalities.

[0018] The growth evolution curve of each fetus and the extracted inter-fetal coupling coefficient matrix are used as model inputs and embedded into a multi-objective joint modeling structure based on time series modeling to quantify the degree of developmental coordination and potential pathological interference risks between different fetuses. In multiple pregnancies, not only are there differences in individual development between fetuses, but if one fetus has abnormal blood flow or restricted nutrition, it may affect other fetuses through the physiological mechanisms of the placenta. The coupling coefficient matrix can quantify this interaction, providing a scientific basis for early intervention.

[0019] Based on the interactive prediction indicators output by the multi-objective joint modeling structure, abnormal trends in multiple pregnancies are identified, such as early warning signals such as fetal growth retardation or blood flow abnormalities. Based on the current temporal behavior of each fetus, the tensor state and its development trend are represented. Through preset intervention strategies, the dosing parameters of the personalized drug regulation module are adjusted to achieve personalized precision treatment in multiple pregnancies. This model can effectively identify abnormal trends in fetal development, such as developmental delay and abnormal placental blood flow, and provide early warnings, helping doctors quickly determine which fetuses require priority intervention and which fetuses are in a healthy development stage.

[0020] The drug regulation module can dynamically adjust the dosage and time of drug administration according to the physiological state of the fetus and the predicted results. Through the drug release function based on feedback regulation, the drug release dose can be dynamically adjusted to ensure the optimal development of the fetus.

[0021] The tensor representing the fetus's temporal behavior throughout the gestational cycle is transmitted in real time to the cloud monitoring platform via edge acquisition devices with data upload capabilities. The artificial intelligence analysis model deployed on the cloud monitoring platform models and processes the data to generate prediction results describing the individual fetal risk level; The cloud-based monitoring platform can not only process fetal physiological data in real time, but also provide dynamic health risk assessments and predictions based on artificial intelligence analysis results, helping doctors make more accurate treatment decisions.

[0022] The inter-fetal coupling coefficient matrix, the drug administration parameters of the drug regulation module, and the predicted results of the individual fetal risk level are used as joint input and imported into the multi-parameter fusion analysis system. The feature fusion and state mapping methods are used to output a comprehensive health score and corresponding intervention strategy recommendations. The strategy recommendations cover drug dosage adjustment, diagnosis and treatment process prompts, and follow-up plan recommendations.

[0023] Furthermore, the multimodal physiological behavior time series data includes fetal structural development data obtained by an ultrasonic imaging system, umbilical artery and venous blood flow rate data obtained by a placental Doppler blood flow monitoring system, fetal electrocardiogram or electroencephalogram signal data obtained by a fetal bioelectric monitoring device, and may also include genomic data obtained by prenatal genetic screening.

[0024] Specifically, the steps for normalizing time series data are as follows: Data collection: First, multimodal physiological behavior time series data of fetuses in multifetal pregnancies were collected, including estimated fetal weight, heart rate volatility, blood flow velocity, etc. Calculate the mean and standard deviation: For each fetal time series data feature (such as weight estimate, heart rate volatility, etc.), calculate the mean and standard deviation of the feature at all time points; Standardization operation: Apply the normalization formula to each data point to transform it into zero mean and unit variance; Construction of a temporal behavior representation tensor: The standardized temporal data of each fetus are merged to form a multi-dimensional temporal behavior representation tensor. In this tensor, the standardized data of each fetus at each moment are organized into a vector, and the temporal data of all fetuses can be uniformly modeled and analyzed.

[0025] Furthermore, current multiple pregnancy risk assessment systems generally use single-point or stage-by-stage static physiological indicators for analysis, ignoring the changing trends of the fetus's physiological behavior throughout the pregnancy cycle. In particular, in multiple pregnancy scenarios, they are unable to dynamically capture the evolutionary trajectories of each fetus at different developmental stages. Therefore, existing methods make it difficult to accurately model and quantitatively evaluate potential time-dependent risks, subtle developmental deviations, and coupled feedback effects between fetuses.

[0026] To address the above issues, the present invention proposes to use the "time-dependent growth evolution curve" as the core modeling element to dynamically track and model the development process of each fetus, thereby supporting high-precision modeling of the evolution path of the physiological state of individual fetuses, and further serving as the basic input for subsequent interaction analysis and risk prediction among multiple fetuses.

[0027] The time-dependent growth evolution curve is defined as a continuous time function of time and fetal physiological behavior state vector, and a personalized growth evolution function is constructed based on fetal structural development data, umbilical artery and venous blood flow rate data, fetal electrocardiogram or electroencephalogram signal data and time factors.

[0028] Specifically, the time-dependent growth evolution curve is defined as a continuous time function with respect to time and the fetal physiological behavior state vector, and is expressed in an implementation manner based on a linear model and feature weighting, and is specifically expressed as follows: ; Among them, x i represents the behavioral state vector of the i-th fetus at time t, including weight estimation, heart rate volatility, blood flow velocity, and other data such as estimated weight and fetal movement frequency. i (x i ,t) represents the individual growth evolution function obtained by fitting historical case data with current monitoring data, reflecting its comprehensive physiological health level, often expressed in the form of a standardized score (such as [0,1]) or risk score, and also expressed as the physiological behavior state vector of the i-th fetus at time t; w f It is a pre-trained weight vector used to characterize the importance of each physiological indicator's contribution to growth and development. The weight can be obtained through regression fitting of historical medical data. For example, if the growth rate of fetal biometric values ​​(such as head circumference, abdominal circumference, and femur length) measured by ultrasound is used as the target value, it can be trained using algorithms such as linear regression and ridge regression. b f is the bias term, and σ(·) is the activation function (Sigmoid function) used to map the output value to the [0,1] interval.

[0029] Specifically, when f i (x i,t)<0.3, it means that the development is severely restricted (urgent intervention is required). i (x i ,t)<0.7, it represents medium to low risk. i (x i ,t)>0.7, it indicates a good development state.

[0030] Furthermore, in cases of intrauterine growth restriction (IUGR) or abnormal blood flow in a fetus, this coupling mechanism may trigger a "chain reaction", affecting other fetuses and thus inducing systemic risks. Therefore, the present invention proposes a modeling scheme based on the inter-fetal coupling effect analysis mechanism to quantify and reveal the coordinated development status and potential pathological interference channels between multiple fetuses, thereby providing a basis for more accurate risk prediction and intervention strategy formulation.

[0031] The coupling effect between multiple fetuses is modeled through the functional relationship between the individual growth evolution function of each fetus and the inter-fetal coupling coefficient matrix. The specific steps include: Modeling of the interaction coupling effect: for each pair of fetuses, the interaction coupling effect between the fetuses is calculated using their respective growth evolution functions and the inter-fetal coupling coefficient matrix; Responsiveness measurement, through the model to calculate the interactive coupling responsiveness between fetuses, measures the transmission intensity and mutual influence of physiological resources between fetuses.

[0032] Specifically, the coupling effect between multiple fetuses is modeled by the functional relationship between the individual growth evolution function of each fetus and the inter-fetal coupling coefficient matrix, which is expressed as follows: ; Among them, E is the total coupling response index between fetal groups, which is used to measure the system interaction strength between multiple pregnancies. In normal and healthy multiple pregnancy samples, the average stable value of E is in the range of [0.5, 1.2]. If E>1.5, it means that the system coupling response strength deviates significantly, indicating the risk of developmental imbalance. If the growth rate of abnormal E value (dE / dt>0.2 / week) can be used as the basis for triggering intervention signals; f i (x i ,t) represents the growth evolution function of the i-th fetus, g j (x j ,t) represents the auxiliary characteristic evolution function of the jth fetus, which is usually related to the fetus's metabolic state, placental blood flow and other physiological characteristics, and changes with time, that is, f i (x i ,t) is used to describe the growth process of the fetus, g j (x j ,t) describes the characteristics related to the physiological status of the fetus (such as blood flow, metabolism, etc.); Similarly, the auxiliary feature evolution function g j (x j ,t) can adopt the linear model with the same structure: ; The Sigmoid function may not be used here to retain its significance in characterizing the original changes in physiological characteristics.

[0033] C ij represents the interaction coupling coefficient between fetus i and fetus j, reflecting the strength of their association at the physiological level such as nutrient delivery, blood flow regulation, and pressure coupling. ij The recommended setting range is [0,1], where C ij =0, it means there is no significant interaction between fetus i and j, C ij =1, indicating that there is a strong coupling dependency between fetuses i and j (e.g., shared blood supply); C ij The initial value of can be set as follows: Clinical pathway dependent, e.g., fetuses in dichorionic-diamniotic configurations, C ij Usually less than 0.3; Data-driven method, using the time-delayed mutual information of fetal physiological state for estimation; n is the total number of fetuses in a multiple pregnancy (usually 2-4).

[0034] The specific workflow is as follows: Collect the behavioral state vector x of each fetus i (such as weight estimation, heart rate variability, blood flow velocity, etc.) and auxiliary features x j (such as umbilical vein, cerebral artery, placental blood oxygen, etc.); Fitting individual function f by historical cases and current data i (x i ,t) and g j (x j ,t), construct individual growth trend expression; C is constructed based on biological mechanisms such as blood flow coupling model and tissue perfusion delay model. ij , the initial value can be set from clinical priors or obtained by data-driven learning; Substitute the above three types of quantities into the formula to calculate E in real time; Set a risk threshold τ. When E>τ, the intervention system is triggered to alert medical staff that a fetus may be affecting other fetuses. E is passed as input into the subsequent multi-objective joint model to predict systemic risks and formulate personalized intervention plans.

[0035] For example, in a twin pregnancy, the growth rate of fetus A is A (x A ,t) is significantly lower than that of fetus B, but the fluctuation of blood flow parameters of B increases. AB =0.42 (suggesting that there is a shared area between the two in terms of placental distribution). When the system detects: f A (x A ,t) continued to decline, g B (x B ,t), the cerebral blood flow index RI increased, the overall E value increased from 1.1 to 1.6, and dE / dt increased rapidly; This may indicate that the developmental delay of fetus A has affected B, and it is necessary to adjust B's dosage regimen or increase the monitoring frequency to prevent systemic deterioration.

[0036] Furthermore, the drug administration parameters of the drug regulation module are dynamically changed through the feedback control function. In this way, the drug release strategy can be precisely adjusted according to the real-time monitoring of fetal growth evolution, coupling effects and health status, thereby maximizing the effect of personalized treatment.

[0037] The drug administration parameters of the drug control module are dynamically adjusted according to the fetal development status and real-time feedback. The adjustment rules of the drug administration parameters are controlled by the feedback regulation function. The specific steps include: Dynamic control mechanism, setting the feedback mechanism of drug release dosage and adjusting the drug dosage in real time according to the physiological state of the fetus; The intervention trigger time and feedback rate, the release of drug dosage is related to the intervention trigger time and feedback rate, and the drug administration plan is adjusted according to the model output.

[0038] Specifically, the drug administration parameters of the drug control module are dynamically adjusted through the following feedback regulation function: ; Where D(t) represents the drug release dose at time t, which is a key parameter dynamically adjusted according to the physiological changes of the fetus; A represents the regulation of peak release, providing rapid intervention capabilities in high-risk situations. That is, it is the maximum drug dose that needs to be released to achieve rapid intervention in high-risk situations. In acute high-risk situations (such as fetal growth restriction, insufficient placental blood flow, etc.), the A value needs to be set higher to provide sufficient drug intervention. Usually, this value can be set based on clinical experience and treatment requirements, for example, it can be set to 1.5-2 times the target drug dose; k1 is the rate coefficient of change of the regulation function, which determines the time response speed of drug release. It is set according to the response time window. In general, k1 needs to be adjusted according to the rate of change of fetal health. If there is a sudden change in fetal health (such as sudden placental dysfunction), a larger k1 value needs to be set to accelerate drug release. When the fetal health status is relatively stable, the k1 value should be smaller to avoid excessive intervention. Generally speaking, the value of k1 can be set between 0.1 and 1; t0 represents the time point when the intervention response is triggered, which is usually the time point when the fetal health status becomes abnormal. It is determined according to the clinical risk assessment system. When abnormal fetal physiological behavior is monitored, this value will be automatically adjusted to the time point when the risk occurs. B is the minimum basic maintenance dose, which is the minimum dose of the drug that must be maintained under any circumstances. It is used to ensure the stability of basic physiological functions. This dose is generally small and is set to the minimum value required to maintain the basic physiological functions of the fetus. Clinically, it is usually set at 10%-20% of the normal drug dose.

[0039] This regulatory function is designed to dynamically adjust the amount of drug released based on the fetus's real-time health status and provide rapid intervention capabilities at high-risk moments: Dynamic response, the core feature of this function is through exponential decay To reflect the time dependence of drug release, this design allows the drug release amount to be adjusted in real time according to the health status of the fetus. Especially when the health status of the fetus changes suddenly, this function can respond quickly, provide a peak value of drug intervention, and then gradually stabilize to the basic maintenance dose; Risk perception and regulation capabilities: When fetal risk signals appear (such as abnormal blood flow, fetal growth retardation, etc.), the setting of the A value ensures that the drug can be released quickly in the shortest possible time to respond to the crisis. By adjusting k1, the system can adjust the response speed of drug release, thereby providing more appropriate drug intervention plans under different health conditions; To ensure safety, the lowest basic drug dose B ensures that drug release is always maintained at a safe level without obvious risks, thus avoiding the side effects caused by overdose and ensuring the safety of both mother and fetus.

[0040] Furthermore, existing risk assessment methods for multiple pregnancies generally ignore the individual differences of each fetus and do not adequately consider the interactions between fetuses, which results in the inability to accurately predict health problems such as growth restriction and abnormal heart rate that may occur in each fetus. Therefore, the present invention proposes a risk prediction method based on multimodal physiological behavioral data and a weighted summation model, which can comprehensively evaluate the health status of each fetus, thereby providing a basis for medical intervention.

[0041] The prediction result of the fetal risk level adopts a weighted sum model, and the specific steps include: Integration of physiological and behavioral indicators, combining the physiological and behavioral time series data of each fetus; Weighted summation calculation, based on the weights trained in historical data, performs weighted summation on the physiological behavior time series data of each fetus to obtain a comprehensive risk assessment value.

[0042] Specifically, the prediction result of the fetal risk level adopts a weighted sum model, which is expressed as follows: ; Among them, P risk is the comprehensive risk estimate at the current time t. Since the physiological state of the fetus will change at different time points, the risk assessment value P risk It is a dynamically changing quantity. When performing weighted summation, the real-time monitoring data will be continuously updated according to the latest situation, and the final risk assessment value will also reflect the health status of the fetus in real time. i is the indicator importance weight obtained from historical data training. This coefficient represents the impact of each indicator on the final risk assessment. i The value of is obtained by training with historical data and can reflect the importance of this indicator in a specific pregnancy state. For example, the importance of fetal movement frequency may be low in a normal pregnancy, but in the case of fetal growth restriction, changes in fetal movement frequency may better reflect the fetal health status. In actual application, these weight coefficients can be dynamically adjusted according to different gestational periods, fetal health status, etc.

[0043] In actual use, each indicator (X i (t)) to make all input dimensions consistent, for example, the fetal heart rate is normalized (such as converting each minute into a standard value per unit time), the fetal movement frequency is normalized to make it consistent with the dimensions of other indicators, and the fetal weight estimate is standardized.

[0044] Furthermore, the generation of the comprehensive health score is achieved by the following steps: The time-dependent growth evolution curve of each fetus, the inter-fetal coupling coefficient matrix, medication parameters and the prediction results of fetal risk level are integrated; By integrating the data using feature normalization or tensor fusion methods, a comprehensive health score for each fetus is ultimately generated, providing clinicians with a quantifiable and personalized health status reference and supporting the intelligent generation of drug regulation and medical decision-making pathways.

[0045] Specifically, the comprehensive health score is expressed by the following integrated function: ; Here, S is the final health assessment score integrating all key factors, and f(·) is a function with feature normalization, tensor fusion, and nonlinear mapping, which supports multi-factor integrated prediction.

[0046] In general, the range of the health score S will vary depending on the specific clinical application. For the comprehensive scoring function of the present invention, the health score S typically ranges from 0 to 1, depending on the model design and the weights and inputs of different variables. For example, if the fetus is in good physiological condition and has low risk, the score S will be close to 1, indicating good health. If the score is close to 0, it indicates that the fetus is at serious health risk and requires urgent intervention.

[0047] Suppose we have a deep neural network model to represent f(·), its basic structure can be expressed as: ; Among them, σ is the activation function, usually Sigmoid or ReLU, to ensure that the output score is between 0 and 1, and α and β are weight coefficients used to adjust the impact of drug intervention and risk score on the final result.

[0048] Furthermore, the present invention constructs an individual fetal growth evolution function and an inter-fetal coupling coefficient matrix to accurately characterize the individual fetal development trend and the dynamic interaction relationship between fetuses in the form of time series modeling. Further, with the help of an artificial intelligence engine, the physiological risk factors output by the model are integrated and analyzed to generate a comprehensive health score S in real time. Based on the health score results, an intervention strategy generation module is constructed to make intelligent recommendations on possible intervention methods during multiple pregnancies.

[0049] The intervention strategy recommendations include independent dosing routes for singleton fetuses, risk warning information provided to medical personnel, and dynamic ranking recommendations for fetal monitoring priorities.

[0050] The above score S is used as the input parameter of the intervention strategy decision module to achieve the following three intelligent strategy outputs: Independent medication route recommendation, when the system is based on the comprehensive health score S and individual risk assessment value P risk When a fetus is judged to be in a high-risk range, the drug intervention feedback control model is activated, i.e. D(t); Risk warning information (provided to medical personnel), the system is based on the coupling influence coefficient C between the fetus ij and individual risk assessment value P risk To determine whether a risk warning needs to be issued, for example, when the coupling influence coefficient C between the two fetuses is ij Exceeding the threshold 0.5 and the risk value P riskWhen the value is greater than the threshold, the system will automatically trigger a risk warning, prompting medical staff to pay attention to the health of the fetus; Dynamic sorting of fetal monitoring priorities is recommended, with each fetus’s P risk , relative coupling influence and relative coupling influence comprehensive calculation monitoring priority R i : ; Among them, σ i is the physiological variation rate of the fetus, which describes the fluctuation of the fetal physiological state. For a certain physiological index x i (t) (e.g., fetal heart rate) is sampled N times within the time window [t-Δt,t], and its variation rate is defined as: ; Among them, x i (t k ) is expressed as at time point t k Physiological data, It is expressed as the mean of multiple groups of physiological data within the time window Δt; C i is the coupling influence of the fetus, that is, the influence intensity with other fetuses, expressed as: ; λ1, λ2, λ3 are the corresponding weight coefficients, preferably, λ1∈[0.4,0.6], clinical risk dominant weight; λ2∈[0.2,0.3], coupling influence dominant weight; λ3∈[0.1,0.2], behavioral volatility contribution, the system according to R i Dynamically adjust the frequency of fetal monitoring recommendations (for example, at least three monitoring sessions are recommended within 24 hours).

[0051] Take, for example, a pregnant woman with twins who receives real-time fetal monitoring from an edge device: The system detected an increase in the umbilical artery pulsatility index (PI) and decreased fetal movement in fetus A; At the same time, it was found that there was a significant positive coupling between fetus B and A (C AB =0.72); The system triggers a coupling warning and recommends that medical staff increase the frequency of B's ​​monitoring; For A, start the drug path feedback function, initially setting A=5mg / h, k1=0.4; At the same time, the fetus priority monitoring ranking is output: B>A (because it is susceptible to coupling conduction).

[0052] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for risk assessment and management of multiple pregnancies, characterized in that: The following steps are involved: Collect and standardize the multimodal physiological behavior time series data of each fetus in multiple pregnancies to construct a time series behavior representation tensor for each fetus; Based on the temporal behavior representation tensor, the time-dependent growth evolution curve is extracted, and the coupling coefficient matrix reflecting the multi-fetal interaction is constructed by combining the coupling relationship of physiological parameters between fetuses. Inputting the growth evolution curve and coupling coefficient matrix into a multi-objective joint modeling structure to quantify the developmental synergy between fetuses and the risk of pathological interference; Identify pregnancy abnormality trends based on interactive indicators output by the multi-objective joint modeling structure, and adjust drug administration parameters based on the temporal status and development trends of each fetus; The tensor representing the temporal behavior throughout the pregnancy cycle is uploaded to the cloud-based artificial intelligence analysis platform, and processed to obtain the individual fetal risk level prediction; The coupling coefficient matrix, medication parameters and risk level prediction are integrated to output a comprehensive health score and intervention strategy recommendations.

2. A method for risk assessment and management of multiple pregnancies according to claim 1, characterized in that: The multimodal physiological behavior time series data includes fetal structural development data obtained by an ultrasonic imaging system, umbilical artery and vein blood flow rate data obtained by a placental Doppler blood flow monitoring system, and fetal electrocardiogram or electroencephalogram signal data obtained by a fetal bioelectric monitoring device.

3. A method for risk assessment and management of multiple pregnancies according to claim 2, characterized in that: The time-dependent growth evolution curve is defined as a continuous time function of time and fetal physiological behavior state vector, and a personalized growth evolution function is constructed based on fetal structural development data, umbilical artery and venous blood flow rate data, fetal electrocardiogram or electroencephalogram signal data and time factors.

4. A method for risk assessment and management of multiple pregnancies according to claim 3, characterized in that: The coupling effect between multiple fetuses is modeled through the functional relationship between the individual growth evolution function of each fetus and the inter-fetal coupling coefficient matrix. The specific steps include: Modeling of the interaction coupling effect: for each pair of fetuses, the interaction coupling effect between the fetuses is calculated using their respective growth evolution functions and the inter-fetal coupling coefficient matrix; Responsiveness measurement, through the model to calculate the interactive coupling responsiveness between fetuses, measures the transmission intensity and mutual influence of physiological resources between fetuses.

5. A method for risk assessment and management of multiple pregnancies according to claim 4, characterized in that: The drug administration parameters of the drug control module are dynamically adjusted according to the fetal development status and real-time feedback. The adjustment rules of the drug administration parameters are controlled by the feedback regulation function. The specific steps include: Dynamic control mechanism, setting the feedback mechanism of drug release dosage and adjusting the drug dosage in real time according to the physiological state of the fetus; The intervention trigger time and feedback rate, the release of drug dosage is related to the intervention trigger time and feedback rate, and the drug administration plan is adjusted according to the model output.

6. A method for risk assessment and management of multiple pregnancies according to claim 5, characterized in that: The fetal risk level is predicted using a weighted sum model, and the specific steps include: Integration of physiological and behavioral indicators, combining the physiological and behavioral time series data of each fetus; Weighted summation calculation, based on the weights trained in historical data, performs weighted summation on the physiological behavior time series data of each fetus to obtain a comprehensive risk assessment value.

7. A method for risk assessment and management of multiple pregnancies according to claim 6, characterized in that: The comprehensive health score is generated by the following steps: The time-dependent growth evolution curve of each fetus, the inter-fetal coupling coefficient matrix, medication parameters and the prediction results of fetal risk level are integrated; Feature normalization or tensor fusion methods are used for integration to ultimately generate a comprehensive health score for each fetus.

8. The method for risk assessment and management of multiple pregnancies according to claim 7, wherein: The intervention strategy recommendations include independent dosing routes for singleton fetuses, risk warning information provided to medical personnel, and dynamic ranking recommendations for fetal monitoring priorities.

Images