A clinical diagnosis and treatment path generation method and system for tracking chest pain patients

Abstract

This application provides a method and system for generating clinical treatment pathways to track patients with chest pain. The method involves constructing a dynamic spatiotemporal data map by simultaneously acquiring respiratory rhythm phase difference, coronary artery calcium integral, and the slope of the high-sensitivity troponin curve. This data is then coupled with the asymmetric increase in ST-segment elevation amplitude, the attenuation gradient of left ventricular ejection fraction, and the release rate of B-type natriuretic peptide to generate risk transition signals. In the acute phase, the method dynamically binds the recanalization time window to the constraints of thrombolysis contraindications. In the subacute phase, a phase synchronization rule between enhanced CT scans and anticoagulant drug dosage is established through a multidisciplinary interface. The method integrates intravascular ultrasound plaque morphological mutation characteristics and cardiopulmonary bypass parameters, dynamically adjusting the revascularization sequence based on the myocardial oxygen consumption balance equation. Pathway self-optimization is achieved through the spatiotemporal superposition of interventional treatment delay parameters and drug metabolism curves. The technical solution provided in this application achieves dynamic optimization of chest pain treatment pathways and collaborative decision-making using multimodal parameters, improving the accuracy of risk stratification and the timeliness of treatment.

Description

Technical Field

[0001] This application relates to the field of diagnostic and treatment pathway generation technology, and in particular to a method and system for generating clinical diagnostic and treatment pathways for tracking patients with chest pain. Background Technology

[0002] With the increasing burden of cardiovascular disease and the growing demand for precision medicine, the chest pain center diagnosis and treatment system faces the dual challenges of multimodal data fusion and dynamic decision optimization.

[0003] In high-risk chest pain scenarios such as acute coronary syndrome and aortic dissection, it is necessary to construct an intelligent diagnosis and treatment pathway generation system with spatiotemporal parameter coordination and self-optimization capabilities; to achieve multi-dimensional spatiotemporal correlation modeling of respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve, breaking through the limitations of traditional single-time-point detection; to establish a dynamic coupling analysis mechanism between ST-segment elevation asymmetric growth and B-type natriuretic peptide release rate, improving the sensitivity of risk stratification; and to construct a spatial constraint model of vascular recanalization time window and thrombolysis contraindications, achieving spatiotemporal adaptation optimization of treatment plans.

[0004] In existing technical solutions, a 0 / 1 hour clinical decision-making pathway based on high-sensitivity troponin (hs-cTn) is used to accelerate chest pain assessment through a dynamic threshold algorithm; multimodal image integration technology is used to construct a three-dimensional diagnostic and treatment model by integrating coronary CTA, intravascular ultrasound, and extracorporeal circulation parameters; and an intelligent body orchestration system is used to optimize tuberculosis infection risk assessment and surgical pathway planning through deep learning.

[0005] However, existing technical solutions have the following technical defects: static data modeling cannot capture the dynamic correlation between respiratory phase difference and troponin curve, resulting in more than 20% of early ACS being missed; single-dimensional ST segment analysis ignores the spatiotemporal coupling effect of left ventricular ejection fraction decay gradient, causing more than 30% of risk misjudgments, and the accuracy of risk stratification and the timeliness of treatment are not high; traditional thrombolysis time window calculation does not integrate plaque morphological mutation characteristics, causing 15% of vascular recanalization timing to deviate from the optimal solution. Summary of the Invention

[0006] This application provides a method and system for generating clinical diagnosis and treatment pathways to track patients with chest pain, in order to solve the problems of low accuracy in risk stratification and low timeliness of treatment in the prior art.

[0007] In a first aspect, embodiments of this application provide a method for generating clinical treatment pathways to track patients with chest pain, including: By synchronously collecting respiratory rhythm phase difference, coronary artery calcium integral and high-sensitivity troponin curve slope within a preset time after chest pain onset, a spatiotemporal data map with dynamic priority labeling is established, and incremental weighting is applied to the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision-making baseline framework. Based on the spatiotemporal data map, the spatiotemporal correlation pattern between myocardial ischemia area and ECG lead abnormalities is identified. When the asymmetric increase of ST segment elevation amplitude is continuously detected, the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment pathway. Based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, the blood vessels are dynamically bound in the acute phase of the intensity distribution, and the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis is constructed to build an executable path topology. At the same time, in the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface. During the implementation of the executable path topology, the plaque morphological mutation characteristics of intravascular ultrasound and the circulatory auxiliary parameters of the extracorporeal life support device are integrated in real time. Based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation, the priority queue of treatment plans is reconstructed and the temporal dependence of revascularization surgery is dynamically adjusted. Interventional treatment delay parameters and dual antiplatelet drug metabolism curves are spatiotemporally superimposed according to the time-series dependency to generate a causal correlation map. The catheterization lab activation delay and drug efficacy verification results in the causal correlation map are then inversely mapped to the weighted reinforcement model of the clinical decision-making baseline framework to drive the self-optimization and updating of the path topology and construct a clinical diagnosis and treatment path generation strategy.

[0008] Optionally, based on the spatiotemporal data atlas, the spatiotemporal correlation pattern between myocardial ischemia areas and abnormal electrocardiogram leads is identified. When asymmetric increases in ST segment elevation amplitude are continuously detected, a coupling analysis linked to the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment pathway, including: The myocardial ischemia region in the spatiotemporal data map is located in three dimensions, and the abnormal fluctuation phase features of each lead channel are extracted. By aligning the multimodal signals in the abnormal fluctuation phase features with time series, a dynamic propagation relationship model is established. A dynamic slip monitoring window is constructed for the dynamic propagation relationship model. The ST segment elevation amplitude is decomposed into waveforms, and peak features, slope change rate and spatial lead distribution heterogeneity parameters are extracted. When the asymmetric growth index is detected to exceed the dynamic adjustment threshold within three consecutive windows, the cardiac function compensation reserve analysis channel is activated. The spatiotemporal trajectory of the left ventricular ejection fraction decay gradient and the fluctuation characteristics of the B-type natriuretic peptide release rate in the cardiac function compensation reserve analysis channel are acquired simultaneously. Through the multi-source data coupling engine of electrocardiogram-ultrasound-biomarkers, a dynamic weight model of ejection fraction change rate and B-type natriuretic peptide release rate is established, and the composite quantitative index is calculated. The asymmetric growth index and the composite quantitative index are subjected to spatiotemporal superposition analysis. When the spatiotemporal distribution of the two in the coronary artery blood supply area is detected to be in a critical coupling state, a directional risk transition vector is generated based on the dynamic matching rules of vascular lesion classification, inputting three-dimensional feature parameters such as acute exacerbation probability, compensation failure time window and targeted intervention sensitivity. Multi-level cross-validation is performed on the three-dimensional feature parameters. When the probability of acute exacerbation during the multi-level cross-validation process exceeds the standard deviation range of the preset baseline value, and the compensation failure time window overlaps with the current diagnosis and treatment stage, priority is given to intervening in the target coordinates and path switching time constraints to generate a risk transition signal to drive the diagnosis and treatment path switching.

[0009] Optionally, based on the intensity distribution of the risk transition signal driving the treatment pathway, the method dynamically binds blood vessels during the acute phase of the intensity distribution, establishes the temporal-spatial constraint relationship between the recanalization time window and thrombolysis contraindications, constructs an executable pathway topology, and simultaneously establishes a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve through a multidisciplinary parameter negotiation interface during the subacute phase of the intensity distribution, including: Based on the intensity distribution of the transition trigger signal, the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate are extracted, a two-dimensional constraint matrix is ​​constructed, the time axis of the two-dimensional constraint matrix is ​​used to map the change of reperfusion injury probability gradient, and the spatial axis of the two-dimensional constraint matrix is ​​used to associate the anatomical topology of coronary artery branches. Based on the gradient change characteristics in the two-dimensional constraint matrix, a dynamic matching rule between the CT angiography time window and the frequency of cardiac ultrasound examinations is established during the acute phase. The priority ranking of imaging examination nodes is driven by the real-time decay gradient of the hemodynamic parameters of the dynamic matching rule to generate a dynamic sequence chain of imaging examinations. The temporal distribution characteristics of the dynamic sequence chain of the imaging examination are injected into the thrombolysis contraindication matrix. The slope of the change in vascular occlusion rate in the thrombolysis contraindication matrix is ​​dynamically coupled with the myocardial survival threshold. The catheterization lab preparation stage, the anticoagulant loading time window, and the balloon dilation operation sequence are introduced to generate a parallel execution framework. During the subacute phase, the operation sequence in the parallel execution framework is spatiotemporally encoded with the blood flow reserve parameters of the enhanced CT scan. By inversely matching the metabolic rate of the anticoagulant drug with the contrast agent clearance curve, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established.

[0010] Optionally, the asymmetric growth index and the composite quantification index are subjected to spatiotemporal superposition analysis. When a critical coupling state is detected in the spatiotemporal distribution of the two in the coronary artery blood supply area, based on the dynamic matching rules of vascular lesion classification, three-dimensional feature parameters of acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity are input to generate a directional risk transition vector, including: A spatiotemporal convolution kernel group is introduced, and the parameters of the spatiotemporal convolution kernel group are calibrated by dynamic calibration of the branch vessel anatomical diameter, real-time pressure gradient and fractional flow reserve in the coronary artery supply area. Multi-scale feature extraction is performed on the asymmetric growth index and the composite quantization index to generate a vessel topology weight spatiotemporal superposition feature tensor. A dynamic threshold is obtained by calculating the moving average of the vascular endothelial shear stress waveform and the plaque morphological stability index. The vascular topology weight spatiotemporal superposition feature tensor is activated layer by layer by adjusting the dynamic threshold. When the feature activation intensity of at least two adjacent vascular segments exceeds the critical threshold of their corresponding branch vascular lesion classification, a spatiotemporal coupling state marker is triggered. The covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate, and microcirculation resistance index were extracted from the vascular lesion classification parameter library. A dynamic matching rule engine was established, and logistic regression was used to calculate the probability of acute exacerbation. The time window of compensation failure was predicted by the Kaplan-Meier modified model, and the sensitivity of targeted intervention was evaluated by the random forest classifier. The probability of acute exacerbation, the time window of compensation failure, and the sensitivity of targeted intervention were three-dimensionally normalized and encoded to generate a directional risk transition vector.

[0011] Optionally, during the implementation of the executable path topology, the plaque morphological mutation characteristics of intravascular ultrasound and the circulatory auxiliary parameters of the extracorporeal life support device are fused in real time. Based on the plaque rupture risk threshold and the myocardial oxygen consumption balance equation, the priority queue of treatment plans is reconstructed and the temporal dependency of revascularization surgery is dynamically adjusted, including: Plaque morphology mutation feature parameters were constructed by intravascular ultrasound imaging sequences. Multi-scale convolutional neural networks were used to extract plaque fibrous cap thickness gradient, calcification density distribution and lipid core volume ratio. Combined with arterial pressure waveform and ventricular assist flow parameters collected in real time by extracorporeal life support equipment, a plaque morphology-hemodynamic coupling feature vector was generated. The plaque morphology-hemodynamic coupling feature vector is input into the plaque rupture risk assessment model based on a random forest classifier. The product parameter of coronary artery perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated simultaneously. When the plaque rupture risk score exceeds the dynamic threshold and the oxygen consumption balance index deviates from the preset range, the blood revascularization surgery priority calculation channel is activated. In the blood revascularization surgery priority calculation channel, a weighted correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance is established. The sliding time window algorithm is used to dynamically sort the weighted correlation matrix. Combined with the catheterization lab resource occupancy status and anticoagulant drug metabolism half-life parameters, a time window and a priority queue for initiation time are generated. Based on the spatiotemporal constraints between the time window and the priority queue of thrombolysis and interventional procedures, a surgical timing dependency is established. The myocardial salvage index change curve under different timing combinations in the surgical timing dependency is predicted by a recurrent neural network, and the revascularization scheme corresponding to the maximum slope descent point is selected as the optimal path.

[0012] Optionally, the plaque morphology-hemodynamic coupling feature vector is input into a plaque rupture risk assessment model based on a random forest classifier, and the product parameter of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated simultaneously. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the priority calculation channel for revascularization surgery is activated, including: The hyperparameters of the random forest classifier are dynamically Bayesian tuned. The importance of features is ranked based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform. The interaction term between calcification density distribution and ventricular auxiliary flow parameters is selected as the key input feature to generate a dynamic threshold curve for plaque rupture risk score. Based on the dynamic threshold curve of the plaque rupture risk score, the product parameter of coronary artery perfusion pressure and ventricular wall tension is calculated synchronously. The fluctuation amplitude and baseline offset of the product parameter are extracted by the sliding time window algorithm. A two-dimensional deviation index is established in combination with the dynamic threshold curve of the plaque rupture risk score. When the two-dimensional deviation index exceeds the preset warning line three times in a row, the activation command of the blood revascularization surgery priority calculation channel is triggered. In the priority calculation channel of the activation command of the revascularization surgery priority calculation channel, a spatiotemporal correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance is constructed. A dynamic weighted algorithm is used to fuse catheterization lab resource occupancy status parameters and anticoagulant drug metabolism half-life decay curves to generate a dynamic decision map. The interventional surgery initiation time in the dynamic decision graph is input into a recurrent neural network for temporal conflict detection. The optimal interventional path is selected through a myocardial salvage index prediction model. The plaque morphology mutation feature parameters and cyclic auxiliary parameters during the path execution process are fed back to the rupture risk assessment model for incremental learning and updating.

[0013] Optionally, the step of spatiotemporally superimposing the interventional treatment delay parameters and the metabolic curves of dual antiplatelet drugs according to the time-series dependency to generate a causal correlation map, and inversely mapping the catheterization lab activation delay and drug efficacy verification results in the causal correlation map to the weighted reinforcement model of the clinical decision-making baseline framework, driving the self-optimization update of the path topology, and constructing a clinical diagnosis and treatment path generation strategy, includes: The time series alignment and fusion of interventional treatment delay parameters and dual antiplatelet drug metabolism curves were used. The phase difference parameters between the drug concentration decay gradient and the catheterization lab preparation time were extracted through the sliding time window in the time series alignment process to generate a spatiotemporal superposition feature matrix. A causal relationship graph based on Bayesian network is constructed, using the delay parameter and the slope of the drug metabolism curve in the spatiotemporal superposition feature matrix as graph nodes, and the causal relationship graph between coronary artery blood flow recovery rate and platelet inhibition rate is identified by dynamic directed edge weights. The catheterization lab activation delay parameter and drug efficacy verification results in the causal association map are input into the random forest regression model to calculate the dynamic adjustment coefficient of the weighted enhancement model in the clinical decision baseline framework. The association weight between the coronary artery calcification integral and the slope of the high-sensitivity troponin curve is updated through the gradient descent model. A path topology self-optimization model is constructed using a deep reinforcement learning framework. The associated weights are dynamically coupled with real-time intravascular ultrasound plaque morphology parameters to generate a clinical diagnosis and treatment path generation strategy.

[0014] Secondly, embodiments of this application provide a clinical diagnosis and treatment pathway generation system for tracking patients with chest pain, including: The acquisition module is used to simultaneously acquire the respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve slope within a preset time after the onset of chest pain, establish a spatiotemporal data map with dynamic priority labels, and implement incremental weight enhancement on the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision baseline framework. The identification module is used to identify the spatiotemporal correlation pattern between myocardial ischemia area and abnormal electrocardiogram leads based on the spatiotemporal data map. When the asymmetric increase of ST segment elevation amplitude is continuously detected, the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment path. The construction module is used to dynamically bind blood vessels in the acute phase of the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, and construct the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis, based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path. At the same time, in the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface. The fusion module is used to fuse the plaque morphological mutation characteristics of intravascular ultrasound with the circulatory auxiliary parameters of the extracorporeal life support device in real time during the implementation of the executable path topology. Based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation, it reconstructs the priority queue of treatment plans and dynamically adjusts the temporal dependency of revascularization surgery. The mapping module is used to spatiotemporally superimpose interventional treatment delay parameters and dual antiplatelet drug metabolism curves according to the time-series dependency relationship to generate a causal relationship map. The catheterization lab activation delay and drug efficacy verification results in the causal relationship map are then inversely mapped to the weighted reinforcement model of the clinical decision baseline framework to drive the self-optimization update of the path topology and construct a clinical diagnosis and treatment path generation strategy.

[0015] Thirdly, embodiments of this application provide a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are to be invoked and executed by the processing component to implement a method for generating clinical diagnosis and treatment pathways for tracking patients with chest pain as described in the first aspect.

[0016] Fourthly, embodiments of this application provide a computer storage medium storing a computer program, which, when executed by a computer, implements a method for generating clinical treatment pathways for tracking patients with chest pain as described in the first aspect.

[0017] In this embodiment, a spatiotemporal data map with dynamic priority labels is established by synchronously collecting the respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve slope within a preset time after chest pain onset. Incremental weighting is applied to the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision baseline framework. Based on the spatiotemporal data map, the spatiotemporal correlation pattern between myocardial ischemia areas and abnormal ECG leads is identified. When asymmetric growth in ST segment elevation amplitude is continuously detected, a coupling analysis linked to the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the treatment pathway. According to the intensity distribution of the risk transition signal driving the treatment pathway, blood vessels are dynamically bound during the acute phase of the intensity distribution. The temporal-spatial constraint relationship between the recanalization time window and thrombolysis contraindications is constructed to build an executable path topology. The structure, simultaneously, establishes a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve through a multidisciplinary parameter negotiation interface during the subacute phase of the intensity distribution; during the implementation of the executable path topology, the plaque morphological mutation characteristics of intravascular ultrasound and the circulatory auxiliary parameters of the extracorporeal life support device are fused in real time, and the treatment plan priority queue is reconstructed and the temporal dependence of revascularization surgery is dynamically adjusted based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation; the interventional treatment delay parameters and the metabolism curve of dual antiplatelet drugs are spatiotemporally superimposed according to the temporal dependence to generate a causal correlation map, and the catheterization laboratory activation delay and drug efficacy verification results in the causal correlation map are inversely mapped to the weighted reinforcement model of the clinical decision baseline framework to drive the self-optimization update of the path topology and construct a clinical diagnosis and treatment path generation strategy.

[0018] The technical solution of this application has the following beneficial effects: By modeling the spatiotemporal correlation between respiratory rhythm phase difference and the slope of the high-sensitivity troponin curve, we overcome the limitations of traditional single-time-point detection and improve the sensitivity of early ACS identification. Based on the coupling analysis of ST-segment elevation asymmetric growth characteristics and B-type natriuretic peptide release rate, we establish a dynamic mapping relationship between myocardial ischemia area and electrocardiogram abnormalities, enhancing the early warning capability of risk transitions. Spatial constraint modeling of vascular recanalization time window and thrombolysis contraindications enables spatiotemporal adaptation optimization of acute phase treatment plans. By integrating intravascular ultrasound plaque morphological mutation characteristics and extracorporeal circulation parameters, we construct a dynamic adjustment mechanism for revascularization sequence to reduce the risk of plaque rupture. Through inverse mapping of causal correlation maps, we achieve closed-loop optimization of diagnosis and treatment pathways, forming a personalized clinical decision-making iterative upgrade system.

[0019] Furthermore, by locating the myocardial ischemia area in three-dimensional spatiotemporal space and extracting the abnormal fluctuation phase characteristics of leads, a dynamic propagation model with time series alignment of multimodal signals was established. A dynamic sliding monitoring window was used to decompose the ST segment elevation amplitude into waveforms. When the asymmetric growth index was detected to exceed the threshold in three consecutive windows, the cardiac function compensation reserve analysis channel was activated. Through the ECG-ultrasound-biomarker multi-source data coupling engine, the spatiotemporal change trajectories of the left ventricular ejection fraction decay gradient and the B-type natriuretic peptide release rate were simultaneously integrated to construct a dynamic weight model to generate a composite quantitative index. Finally, through spatiotemporal superposition analysis and multi-level cross-validation, a directional risk transition vector was generated in the critical coupling state of the coronary artery supply area, and a risk transition signal was generated based on the acute exacerbation probability, the compensation failure time window, and the intervention sensitivity parameters. This method achieves precise spatiotemporal correlation modeling between myocardial ischemia areas and electrocardiogram abnormalities. By linking dynamic monitoring of ST segment asymmetric growth with the analysis of cardiac functional compensatory reserve, it improves the early warning sensitivity of acute coronary syndrome. Combining a dynamic weighted model of ejection fraction decay gradient and B-type natriuretic peptide release rate, the spatiotemporal resolution of risk transition signals reaches the sub-millimeter level, reducing the false positive rate compared with traditional single-dimensional ST segment analysis methods. The path switching constraints established through a multi-level cross-validation mechanism can shorten the time window for identifying acute exacerbation events, significantly improving the timeliness of treatment for high-risk patients.

[0020] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a method for generating clinical treatment pathways to track patients with chest pain, provided in an embodiment of this application; Figure 2 This application provides a schematic diagram of the structure of a clinical diagnosis and treatment pathway generation system for tracking patients with chest pain, as shown in the embodiments of the present application. Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0024] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Figure 1 This application provides a flowchart of a method for generating clinical treatment pathways to track patients with chest pain, as shown in the embodiments of this application. Figure 1 As shown, the method includes: 101. By synchronously collecting the respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve slope within a preset time after the onset of chest pain, a spatiotemporal data map with dynamic priority labeling is established, and incremental weighting is applied to the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision-making baseline framework. In this step, the spatiotemporal data atlas is a comprehensive data structure that constructs a dynamic physiological change model encompassing both temporal and spatial dimensions by synchronously collecting key indicators such as the respiratory rhythm phase difference, coronary artery calcium integral, and the slope of the high-sensitivity troponin curve within a preset timeframe after chest pain onset. This atlas not only describes the trends of these physiological parameters but also reveals their interrelationships, providing a rich information foundation for clinical decision-making. Incremental weighting refers to the process of dynamically adjusting the rate of change of key indicators in the spatiotemporal data atlas. By assigning higher weights to important data, these data receive more attention in subsequent analysis and decision-making; this process helps improve the accuracy of diagnosis and the effectiveness of treatment plans. The clinical decision baseline framework is a preliminary treatment pathway framework generated based on the above data and weight adjustments. It provides a systematic evaluation standard to help physicians develop personalized treatment plans based on the patient's specific situation. This framework considers the influence of multiple factors and is continuously updated and optimized to adapt to different clinical needs.

[0027] In this embodiment, firstly, the system synchronously collects various physiological data within a preset time after the onset of chest pain; secondly, these data are integrated into a spatiotemporal data atlas; thirdly, incremental weighting is applied to the rate of change of key indicators in the atlas to ensure that important data receives more attention; finally, a clinical decision-making baseline framework is generated to provide a reference for subsequent diagnosis and treatment.

[0028] Suppose a patient is admitted to the hospital due to acute chest pain. Within the first hour after the onset of chest pain, the system simultaneously collects the patient's respiratory rhythm phase difference, coronary artery calcium score, and the slope of the high-sensitivity troponin curve. Next, the system integrates this data into a spatiotemporal data atlas. Then, the system performs incremental weighting on the rate of change of key indicators in the atlas, identifying the most significant change as the slope of the high-sensitivity troponin curve. Finally, a preliminary clinical decision-making baseline framework is generated to guide physicians in further examinations and treatments.

[0029] 102. Based on the spatiotemporal data map, identify the spatiotemporal correlation pattern between myocardial ischemia area and abnormal electrocardiogram leads. When asymmetric growth of ST segment elevation amplitude is continuously detected, activate the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate to generate a risk transition signal driving the diagnosis and treatment pathway. In this step, the ST segment is a portion of the electrocardiogram located between the QRS complex and the T wave. It represents the time interval between depolarization and the onset of repolarization in ventricular myocytes. Abnormalities in the ST segment (such as elevation or depression) are often associated with heart diseases such as myocardial ischemia or myocardial infarction.

[0030] B-type natriuretic peptide (BNP) is a hormone secreted by the heart, primarily responsible for regulating fluid balance and blood pressure. Elevated BNP levels are commonly associated with heart diseases such as heart failure and are a frequently used biomarker for heart failure in clinical practice.

[0031] Myocardial ischemia refers to the portion of the heart lacking oxygen due to insufficient blood supply, typically reflected by abnormalities in electrocardiogram (ECG) leads; identifying these areas is crucial for the early detection and intervention of acute cardiovascular events. Spatiotemporal correlation patterns describe the complex relationship between myocardial ischemia areas and ECG lead abnormalities, including temporal and spatial connections; identifying these patterns helps to more accurately pinpoint the problem and implement targeted measures. Coupling analysis, combining left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate, comprehensively assesses changes in cardiac function; this method provides more comprehensive information than a single indicator, thus supporting more effective clinical decision-making. Risk transition signals are warning signals triggered by the detection of asymmetric growth in ST-segment elevation, indicating the need for emergency intervention; the generation of this signal is based on multidimensional data analysis, ensuring timely response to potential cardiovascular risks.

[0032] In this embodiment, firstly, based on spatiotemporal data maps, the system identifies the spatiotemporal correlation pattern between myocardial ischemia areas and abnormal electrocardiogram leads; secondly, when continuous asymmetric growth of ST segment elevation is detected, coupling analysis is activated; thirdly, the cardiac function status is assessed by combining the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate; finally, a risk transition signal driving the diagnosis and treatment pathway is generated to prompt doctors to take emergency measures.

[0033] For example, continuing the previous example, assuming the patient's electrocardiogram shows abnormalities in multiple leads, the system identifies the spatiotemporal correlation pattern between the ischemic area of ​​myocardium and these leads; secondly, the system detects an asymmetric increase in ST segment elevation amplitude; thirdly, the system activates coupling analysis, combining the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate, and finds a significant decline in cardiac function; finally, a risk transition signal is generated, prompting the doctor to take immediate emergency intervention measures.

[0034] 103. Based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, dynamically bind blood vessels in the acute phase of the intensity distribution, and construct the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis to build an executable path topology. At the same time, in the subacute phase of the intensity distribution, establish a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve through a multidisciplinary parameter negotiation interface. In this process, the acute phase of intensity distribution refers to the stage where the patient's condition deteriorates rapidly. During this period, swift action is needed to prevent further deterioration, and data analysis during this phase is crucial for guiding immediate treatment. The recanalization time window is the optimal time window from symptom onset to recanalization surgery; missing this period may significantly reduce treatment effectiveness. The temporal-spatial constraints of thrombolysis contraindications consider the patient's specific conditions to determine suitability for thrombolytic therapy. This involves a comprehensive assessment of multiple factors to ensure the safety and effectiveness of treatment. The multidisciplinary parameter negotiation interface involves collaboration among experts from different departments to jointly determine the optimal treatment plan. This approach integrates multiple opinions and improves decision-making quality. The timing of enhanced CT scans involves selecting the appropriate time for enhanced CT scans to obtain more accurate imaging information, thereby guiding subsequent treatment. Timely imaging examinations are indispensable for accurate diagnosis. The anticoagulant dosage curve adjusts the dosage of anticoagulants according to the patient's real-time condition to maintain stable blood circulation. This requires precise monitoring and timely adjustments to avoid the risk of bleeding or thrombosis.

[0035] In this embodiment, firstly, the temporal-spatial constraint relationship between the vascular recanalization time window and the contraindications for thrombolysis is dynamically bound during the acute phase; secondly, an executable path topology is constructed; thirdly, during the subacute phase, the dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is determined through a multidisciplinary parameter negotiation interface; finally, all treatment measures are coordinated to improve the treatment effect.

[0036] For example, continuing the previous example, assuming the patient enters the acute phase, the system dynamically binds the time-space constraint relationship between the vascular recanalization time window and the contraindications for thrombolysis; secondly, the system constructs an executable path topology; thirdly, in the subacute phase, through a multidisciplinary parameter negotiation interface, the dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is determined; finally, all treatment measures are coordinated and consistent, improving the treatment effect.

[0037] 104. During the implementation of the executable path topology, the plaque morphological mutation characteristics of intravascular ultrasound and the circulatory auxiliary parameters of the extracorporeal life support device are integrated in real time. Based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation, the priority queue of treatment plan is reconstructed and the temporal dependency of revascularization surgery is dynamically adjusted. In this step, plaque morphological mutation characteristics refer to changes in plaque morphology detected by intravascular ultrasound, which may indicate an increased risk of plaque rupture. Monitoring these changes is crucial for preventing acute cardiovascular events. Circulatory support parameters of extracorporeal life support devices, including those used with ECMO and similar devices, are used to maintain blood circulation, especially in cases of severe cardiopulmonary insufficiency. Optimizing these parameters is essential for improving patient survival. The plaque rupture risk threshold sets a safety limit; exceeding this threshold indicates a risk of plaque rupture, requiring immediate intervention. This threshold is based on extensive clinical research and experience. The myocardial oxygen consumption balance equation calculates the balance between myocardial oxygen consumption and supply, ensuring adequate oxygen supply to the myocardium, which is critical for maintaining normal cardiac function. The treatment priority cohort ranks treatments according to the weight of various indicators to maximize therapeutic efficacy, requiring comprehensive consideration of the patient's specific condition and the effectiveness of various treatment methods. The temporal dependence of revascularization surgery refers to the time sequence and dependencies between different treatment steps, ensuring each step is performed at the appropriate time for optimal therapeutic effect. This requires meticulous time management and coordination.

[0038] In this embodiment, firstly, the system integrates the plaque morphological mutation characteristics from intravascular ultrasound with the circulatory auxiliary parameters of the extracorporeal life support device in real time; secondly, based on the plaque rupture risk threshold and the myocardial oxygen consumption balance equation, it reconstructs the priority queue of treatment plans; thirdly, it dynamically adjusts the temporal dependence of revascularization surgery to ensure that each treatment step is performed at the optimal time; finally, it generates an optimized treatment plan to guide doctors in performing effective revascularization surgery, thereby improving the patient's treatment effect and survival rate.

[0039] For example, continuing the previous example, assuming a patient experiences severe coronary artery stenosis in the acute phase, the system integrates in real-time the plaque morphological mutation characteristics detected by intravascular ultrasound with the circulatory support parameters of the ECMO device. Secondly, based on the plaque rupture risk threshold and the myocardial oxygen consumption balance equation, the system reconstructs the treatment priority queue. Thirdly, it dynamically adjusts the timing dependence of revascularization surgery to ensure that PCI (percutaneous coronary intervention) is performed at the optimal time. Finally, it generates an optimized treatment plan, guiding the physician to successfully perform revascularization surgery, significantly improving the patient's treatment outcome and survival rate.

[0040] 105. The interventional treatment delay parameters and the metabolism curves of dual antiplatelet drugs are spatiotemporally superimposed according to the time-series dependence to generate a causal relationship map. The catheterization lab activation delay and drug efficacy verification results in the causal relationship map are then inversely mapped to the weighted reinforcement model of the clinical decision-making baseline framework to drive the self-optimization and updating of the path topology and construct a clinical diagnosis and treatment path generation strategy.

[0041] In this step, the interventional treatment delay parameter refers to the time interval between deciding to perform interventional treatment and actually starting treatment. This time interval has a significant impact on treatment efficacy; a longer delay may lead to disease deterioration, so it is necessary to minimize this period. The dual antiplatelet drug metabolism curve describes the metabolic process and effects after the patient takes the dual antiplatelet drug, helping physicians assess the drug's onset time and duration of action, which is crucial for ensuring the drug's effectiveness and safety. Spatiotemporal overlay refers to integrating the interventional treatment delay parameter with the dual antiplatelet drug metabolism curve in both temporal and spatial dimensions to generate a causal relationship graph. This overlay helps to comprehensively understand the various factors in the treatment process and their interrelationships. The causal relationship graph is a data structure used to demonstrate the causal relationship between interventional treatment delay and drug onset verification results. It can help physicians better understand and predict treatment effects, thereby making more accurate decisions. The catheterization lab activation delay refers to the time interval between deciding to perform interventional treatment and the actual readiness of the catheterization lab. This delay directly affects the timeliness and effectiveness of treatment. A weighted reinforcement model, inversely mapped to the clinical decision baseline framework, can further optimize the treatment pathway. Drug efficacy validation results are actual validation results of the metabolic curves of dual antiplatelet drugs, used to confirm whether the drug has achieved the expected effect within the expected time. These results are fed back into the clinical decision baseline framework, driving the self-optimization and updating of the pathway topology.

[0042] In this embodiment, firstly, the system spatiotemporally superimposes the interventional treatment delay parameters and the metabolic curves of dual antiplatelet drugs according to their temporal dependence; secondly, it generates a causal relationship graph to show the causal relationship between interventional treatment delay and drug efficacy verification results; thirdly, it inversely maps the catheterization lab activation delay and drug efficacy verification results in the causal relationship graph to a weighted enhancement model of the clinical decision baseline framework; finally, based on this feedback information, it drives the self-optimization update of the path topology structure, constructs a clinical diagnosis and treatment path generation strategy, and ensures that each treatment step can be performed at the optimal time to improve the overall treatment effect.

[0043] For example, continuing the previous example, assuming that after a patient decides to undergo PCI, the system records the interventional treatment delay parameters and performs spatiotemporal overlay with the metabolism curves of dual antiplatelet drugs; secondly, the system generates a causal relationship graph, showing the causal relationship between the interventional treatment delay and the drug efficacy verification results; thirdly, the system inversely maps the catheterization lab activation delay and drug efficacy verification results in the causal relationship graph to the weighted reinforcement model of the clinical decision baseline framework; finally, based on this feedback information, the system drives the self-optimization update of the path topology structure to ensure that the catheterization lab is ready in the shortest possible time and confirms that the drug achieves the expected effect within the expected time, ultimately improving the overall treatment effect and patient survival rate.

[0044] In summary, steps 101 to 105 cover the entire process from synchronously acquiring key physiological data, establishing spatiotemporal data atlases, identifying spatiotemporal correlation patterns between myocardial ischemia areas and abnormal electrocardiogram leads, generating risk transition signals driving treatment pathways, constructing executable pathway topologies, to real-time fusion of intravascular ultrasound data and extracorporeal life support device parameters and performing causal correlation analysis. The aim is to provide a systematic assessment and treatment framework to meet the needs of personalized medicine, ensuring patients receive the most suitable treatment at the optimal time, and improving diagnostic accuracy, treatment efficacy, and survival rates. Through these steps, the system can not only dynamically adjust the priority queue of treatment plans but also optimize the temporal-spatial relationship between interventional treatment delays and drug metabolism curves, thereby achieving precision medicine and efficient clinical decision support.

[0045] To further improve the accuracy of spatiotemporal correlation pattern recognition between myocardial ischemia areas and ECG lead abnormalities, in some embodiments, step 102 involves identifying the spatiotemporal correlation pattern between myocardial ischemia areas and ECG lead abnormalities based on the spatiotemporal data atlas. When asymmetric growth in ST segment elevation amplitude is continuously detected, a coupling analysis linked to the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the treatment pathway. This includes: extracting the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate based on the intensity distribution of the transition trigger signal; constructing a two-dimensional constraint matrix; mapping the reperfusion injury probability gradient change through the time axis of the two-dimensional constraint matrix; and associating the anatomical topology of coronary artery branches through the spatial axis of the two-dimensional constraint matrix; based on the gradient change characteristics in the two-dimensional constraint matrix, in the... A dynamic matching rule for CT angiography time windows and echocardiography frequency is established during the acute phase. The priority ranking of imaging examination nodes is driven by the real-time attenuation gradient of hemodynamic parameters in this dynamic matching rule to generate a dynamic sequence chain of imaging examinations. The temporal distribution characteristics of this dynamic sequence chain are injected into a thrombolysis contraindication matrix. The slope of the vascular occlusion rate change in the thrombolysis contraindication matrix is ​​dynamically coupled with the myocardial survival threshold. This is incorporated into the catheterization lab preparation phase, the anticoagulant loading time window, and the balloon dilation operation sequence to generate a parallel execution framework. During the subacute phase, the operation sequences in this parallel execution framework are spatiotemporally encoded with the blood flow reserve parameters of enhanced CT scans. A dynamic phase synchronization rule between enhanced CT scan time nodes and anticoagulant dose curves is established through inverse matching of the anticoagulant metabolism rate and the contrast agent clearance curve.

[0046] In this embodiment, three-dimensional spatiotemporal localization is a method to capture the specific location of myocardial ischemia areas and their temporal trends through precise spatial and temporal localization, ensuring accurate identification of lesion areas. This localization method helps to more accurately identify and treat cardiac problems. Abnormal fluctuation phase characteristics refer to the time points and phase characteristics of abnormal fluctuations appearing in each lead channel. Extracting these characteristics can better understand abnormalities in the electrocardiogram, which are usually manifested as waveform changes or irregular fluctuations within a specific time period. These characteristics provide important basis for subsequent analysis. The dynamic propagation relationship model is established by aligning multimodal signals through time series to describe the propagation path and patterns of abnormal fluctuations between different leads. This model can help identify potential myocardial ischemia areas and their spread trends, providing a comprehensive risk assessment. The dynamic slip monitoring window is a mechanism for real-time monitoring of ST segment elevation amplitude. By setting dynamically adjusted thresholds, it identifies abnormal growth trends. This method can improve the early warning capability for events such as acute myocardial infarction, ensuring timely intervention. The cardiac function compensation reserve analysis channel simultaneously acquires the spatiotemporal trajectory of the left ventricular ejection fraction decay gradient and the fluctuation characteristics of B-type natriuretic peptide release rate to assess cardiac function status and compensatory capacity. This channel provides comprehensive cardiac function assessment information to help physicians formulate optimal treatment plans. The composite quantitative index, calculated by a multi-source data coupling engine of ECG, ultrasound, and biomarkers, reflects cardiac function status and risk level. This index combines multiple parameters to provide a more comprehensive risk assessment and guide clinical decision-making. The directional risk transition vector, based on the dynamic matching rules of vascular lesion classification, inputs three-dimensional feature parameters such as acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity to generate a directional risk transition vector to guide clinical decision-making. This vector helps physicians determine the optimal treatment timing and plan. Multi-level cross-validation ensures the reliability and accuracy of test results through a multi-level data validation process. This process effectively reduces false positive and false negative results and improves diagnostic quality.

[0047] In this embodiment, firstly, the myocardial ischemia region in the spatiotemporal data atlas is located in three dimensions, and the abnormal fluctuation phase characteristics of each lead channel are extracted; secondly, a dynamic propagation relationship model is established by aligning the abnormal fluctuation phase characteristics over time; thirdly, a dynamic slip monitoring window is constructed, and the ST segment elevation amplitude is decomposed into waveforms to extract peak features, slope change rate, and spatial lead distribution heterogeneity parameters; when the asymmetric growth index is detected to exceed the dynamic adjustment threshold within three consecutive windows, the cardiac function compensation reserve analysis channel is activated; simultaneously, the spatiotemporal trajectory of the left ventricular ejection fraction decay gradient and the fluctuation characteristics of the B-type natriuretic peptide release rate are acquired, and analyzed via electrocardiogram- An ultrasound-biomarker multi-source data coupling engine is used to establish a dynamic weighted model of the ejection fraction change rate and the release rate of B-type natriuretic peptide, and to calculate a composite quantification index. The asymmetric growth index and the composite quantification index are spatiotemporally superimposed and analyzed. When the spatiotemporal distribution of the two in the coronary artery supply area shows a critical coupling state, a directional risk transition vector is generated based on the dynamic matching rules of vascular lesion classification, inputting three-dimensional feature parameters such as acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity. Finally, the three-dimensional feature parameters are cross-validated at multiple levels, and the target coordinates and path switching time constraints are prioritized to generate a risk transition signal to drive the switching of diagnosis and treatment pathways.

[0048] Here is a specific example: Assuming a patient presents with chest pain, the system first performs three-dimensional spatiotemporal localization of the ischemic area of ​​the myocardium and extracts abnormal fluctuation phase features from each lead. Secondly, the system establishes a dynamic propagation relationship model by aligning the abnormal fluctuation phase features over time. Thirdly, the system constructs a dynamic slip monitoring window, decomposes the ST segment elevation amplitude into waveforms, and extracts peak features, slope change rate, and spatial lead distribution heterogeneity parameters. When the asymmetric growth index exceeds the dynamic adjustment threshold within three consecutive windows, the cardiac function compensatory reserve analysis channel is activated. Simultaneously, the spatiotemporal trajectory of the left ventricular ejection fraction decay gradient and the fluctuation features of the B-type natriuretic peptide release rate are acquired. Through a multi-source data coupling engine combining ECG, ultrasound, and biomarkers, an ejection fraction analysis model is established. A dynamic weighted model of the rate of change and the release rate of B-type natriuretic peptide was used to calculate a composite quantitative index. The asymmetric growth index and the composite quantitative index were subjected to spatiotemporal superposition analysis. When the spatiotemporal distribution of the two in the coronary artery supply area showed a critical coupling state, a directional risk transition vector was generated based on the dynamic matching rules of vascular lesion classification, inputting three-dimensional feature parameters: acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity. Finally, the system performed multi-level cross-validation on the three-dimensional feature parameters. When the acute exacerbation probability exceeded the standard deviation range of the preset baseline value during the multi-level cross-validation process, and the compensation failure time window overlapped with the current treatment stage, a risk transition signal was generated to drive the switching of treatment pathways, guiding physicians to take emergency measures and successfully preventing the condition from worsening.

[0049] To further improve the accuracy of intensity distribution processing of risk transition signals driving the treatment pathway, in some embodiments, step 103 involves dynamically binding the time-space constraint relationship between the vessel recanalization time window and thrombolysis contraindications during the acute phase of the intensity distribution based on the intensity distribution of the risk transition signals driving the treatment pathway, constructing an executable path topology, and simultaneously establishing dynamic phase synchronization rules between the enhanced CT scan time node and the anticoagulant drug dose curve through a multidisciplinary parameter negotiation interface during the subacute phase of the intensity distribution. This includes: extracting the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vessel occlusion rate based on the intensity distribution of the transition trigger signal, constructing a two-dimensional constraint matrix, mapping the reperfusion injury probability gradient change through the time axis of the two-dimensional constraint matrix, and associating the anatomical topology of coronary artery branches through the spatial axis of the two-dimensional constraint matrix; based on the two-dimensional constraint matrix... The gradient change characteristics in the array are used to establish a dynamic matching rule between the CT angiography time window and the frequency of echocardiography examinations during the acute phase. The real-time attenuation gradient of the hemodynamic parameters in the dynamic matching rule drives the priority ranking of imaging examination nodes to generate a dynamic sequence chain of imaging examinations. The temporal distribution characteristics of the dynamic sequence chain of imaging examinations are injected into the thrombolysis contraindication matrix. The slope of the vascular occlusion rate change in the thrombolysis contraindication matrix is ​​dynamically coupled with the myocardial survival threshold. The catheterization lab preparation stage, anticoagulant loading time window, and balloon dilation operation sequence are introduced to generate a parallel execution framework. During the subacute phase, the operation sequence in the parallel execution framework is spatiotemporally encoded with the blood flow reserve parameters of the enhanced CT scan. By inversely matching the anticoagulant metabolism rate and the contrast agent clearance curve, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant dose curve is established.

[0050] In this embodiment, the two-dimensional constraint matrix is ​​a data structure combining time and space axes, used to extract the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate. This matrix maps the reperfusion injury probability gradient change along the time axis and correlates the anatomical topology of coronary artery branches along the space axis, providing a comprehensive data framework for analysis and decision-making. The dynamic matching rule is a method for establishing the CT angiography time window and the frequency of echocardiography examinations in the acute phase. This rule drives the priority ranking of imaging examination nodes through the real-time decay gradient of hemodynamic parameters, generating a dynamic sequence chain of imaging examinations. The dynamic sequence chain of imaging examinations is a time-distributed feature sequence generated based on the dynamic matching rule, prioritizing the time nodes of imaging examinations to form an ordered examination flow. This sequence chain not only considers the timing of imaging examinations but also incorporates the patient's real-time hemodynamic parameters. The parallel execution framework is an operational sequence framework generated in the subacute phase, combining the catheterization lab preparation stage, the anticoagulant loading time window, and the balloon dilation operation sequence. This framework uses spatiotemporal encoding of the blood flow reserve parameters from enhanced CT scans to ensure efficient coordination between various treatment steps. The dynamic phase synchronization rule is a method to establish the time point of enhanced CT scan and the dose curve of anticoagulant by inversely matching the metabolic rate of anticoagulant drugs with the clearance curve of scanning contrast agent. This rule ensures that enhanced CT scan and anticoagulant use can be coordinated during the subacute phase to improve the overall treatment effect.

[0051] In this embodiment, firstly, the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate are extracted based on the intensity distribution of the transition trigger signal, and a two-dimensional constraint matrix is ​​constructed. Secondly, the temporal axis of this matrix is ​​mapped to the gradient change of reperfusion injury probability, and the spatial axis is associated with the anatomical topology of coronary artery branches. Thirdly, based on the gradient change characteristics in the two-dimensional constraint matrix, a dynamic matching rule between the CT angiography time window and the frequency of echocardiography is established in the acute phase. The priority ranking of imaging examination nodes is driven by the real-time attenuation gradient of hemodynamic parameters, generating a dynamic sequence chain of imaging examinations. Then, the temporal distribution characteristics of the dynamic sequence chain of imaging examinations are injected into the thrombolysis contraindication matrix, and the slope of the vascular occlusion rate change is dynamically coupled with the myocardial survival threshold. The catheterization lab preparation stage, the anticoagulant loading time window, and the balloon dilation operation sequence are introduced to generate a parallel execution framework. Finally, in the subacute phase, the operation sequence in the parallel execution framework is spatiotemporally encoded with the blood flow reserve parameters of the enhanced CT scan. By inversely matching the anticoagulant metabolism rate with the contrast agent clearance curve, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant dose curve is established.

[0052] Here is a specific example: Suppose a patient is rushed to the hospital due to sudden, severe chest pain at night. The system first extracts the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate based on the intensity distribution of the transition trigger signal, constructing a two-dimensional constraint matrix. Second, the system maps the reperfusion injury probability gradient change along the time axis of this matrix and associates the anatomical topology of coronary artery branches along the spatial axis. Third, based on the gradient change characteristics in the two-dimensional constraint matrix, a dynamic matching rule is established between the CT angiography time window and the frequency of echocardiography examinations in the acute phase. The system then prioritizes the imaging examination nodes by driving the real-time decay gradient of hemodynamic parameters, generating a dynamic sequence chain of imaging examinations. Finally, the system... The temporal distribution characteristics of the dynamic sequence chain of imaging examinations were injected into the thrombolysis contraindication matrix, dynamically coupling the slope of the vascular occlusion rate change with the myocardial survival threshold. This involved introducing the catheterization lab preparation phase, the anticoagulant loading time window, and the balloon dilation operation sequence, generating a parallel execution framework. Finally, in the subacute phase, the system spatiotemporally encoded the operation sequences in the parallel execution framework with the blood flow reserve parameters of the enhanced CT scan. By inversely matching the anticoagulant metabolism rate with the contrast agent clearance curve, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant dose curve was established, ensuring efficient coordination of each treatment step and significantly improving patient treatment outcomes and survival rates.

[0053] To further improve the accuracy of spatiotemporal superposition analysis of the asymmetric growth index and the composite quantification index, in some embodiments, step 102 involves performing spatiotemporal superposition analysis on the asymmetric growth index and the composite quantification index. When a critical coupling state is detected in the spatiotemporal distribution of the two in the coronary artery supply area, based on the dynamic matching rules of vascular lesion classification, three-dimensional feature parameters of acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity are input to generate a directional risk transition vector. This includes: introducing a spatiotemporal convolution kernel group; calibrating the parameters of the spatiotemporal convolution kernel group through dynamic calibration of the branch vessel anatomical diameter, real-time pressure gradient, and fractional flow reserve in the coronary artery supply area; extracting multi-scale features from the asymmetric growth index and the composite quantification index to generate a vascular topology weighted spatiotemporal superposition feature tensor; and using vascular endothelial shear stress... A dynamic threshold is obtained by calculating the moving average of the force waveform and the plaque morphology stability index. The spatiotemporal superposition feature tensor of the vascular topology weight is activated layer by layer through the adjustment of the dynamic threshold. When the feature activation intensity of at least two adjacent vascular segments exceeds the critical threshold of their corresponding branch vascular lesion classification, a spatiotemporal coupling state marker is triggered. The covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate and microcirculation resistance index in the vascular lesion classification parameter library are extracted to establish a dynamic matching rule engine. Logistic regression is used to calculate the probability of acute exacerbation. The compensation failure time window is predicted by the Kaplan-Meier modified model. The sensitivity of targeted intervention is evaluated by the random forest classifier. The probability of acute exacerbation, the compensation failure time window and the sensitivity of targeted intervention are three-dimensionally normalized and encoded to generate a directional risk transition vector.

[0054] In this embodiment, the spatiotemporal convolution kernel group is a method for dynamically calibrating the parameters of the spatiotemporal convolution kernel group by using the anatomical diameter of branch vessels in the coronary artery supply area, real-time pressure gradient, and fractional flow reserve. Multi-scale feature extraction is performed on the asymmetric growth index and composite quantization index to generate a spatiotemporal superposition feature tensor of vascular topology weights. This ensures the capture of key information at different temporal and spatial scales, improving the accuracy of feature extraction and the ability to understand complex data structures. Dynamic threshold adjustment is a mechanism calculated using the moving average of the vascular endothelial shear stress waveform and the plaque morphological stability index. This mechanism is used to activate the spatiotemporal superposition feature tensor of vascular topology weights layer by layer. When the feature activation intensity of at least two adjacent vascular segments exceeds the critical threshold of their corresponding branch vessel lesion classification, a spatiotemporal coupling state marker is triggered. This ensures accurate identification of potential risk areas at different time points and locations, allowing for timely intervention and improving diagnostic accuracy and treatment effectiveness. The dynamic matching rule engine, built upon logistic regression, a Kaplan-Meier modified model, and a random forest classifier, is used to assess the probability of acute exacerbations, the time window of compensatory failure, and the sensitivity to targeted intervention. By extracting the covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate, and microcirculation resistance index from a vascular lesion classification parameter database, it generates a directional risk transition vector to guide clinical decision-making, providing comprehensive risk assessment and helping physicians formulate optimal treatment plans, ensuring each step is performed at the optimal time. The directional risk transition vector is a comprehensive indicator generated by three-dimensional normalized encoding of the probability of acute exacerbations, the time window of compensatory failure, and the sensitivity to targeted intervention. This vector, generated based on dynamic matching rules for vascular lesion classification, guides clinical decision-making, helps physicians determine the optimal treatment timing and plan, ensures patients receive the most suitable treatment in the shortest possible time, and improves overall treatment efficacy and survival rates.

[0055] In this embodiment, a spatiotemporal convolutional kernel group is first introduced. The parameters of the spatiotemporal convolutional kernel group are calibrated by dynamically calibrating the branch vessel anatomical diameter, real-time pressure gradient, and fractional flow reserve in the coronary artery supply area. Multi-scale feature extraction is performed on the asymmetric growth index and composite quantization index to generate a spatiotemporal superposition feature tensor of vascular topology weights. Secondly, a dynamic threshold is obtained by calculating the moving average of the vascular endothelial shear stress waveform and the plaque morphology stability index. This dynamic threshold is used to adjust the spatiotemporal superposition feature tensor of vascular topology weights for layer-by-layer activation. When the feature activation intensity of at least two adjacent vascular segments exceeds a certain threshold, the activation is considered complete. When the corresponding branch vascular lesion classification reaches the critical threshold, a spatiotemporal coupling state marker is triggered. Next, the covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate, and microcirculation resistance index are extracted from the vascular lesion classification parameter library to establish a dynamic matching rule engine. Logistic regression is used to calculate the probability of acute exacerbation, the Kaplan-Meier modified model is used to predict the compensation failure time window, and a random forest classifier is used to evaluate the sensitivity of targeted intervention. Finally, the probability of acute exacerbation, the compensation failure time window, and the sensitivity of targeted intervention are three-dimensionally normalized and encoded to generate a directional risk transition vector to guide clinical decision-making.

[0056] Here is a specific example: Suppose a patient is admitted to the hospital due to sudden chest pain. The system first introduces a spatiotemporal convolutional kernel group. The parameters of the spatiotemporal convolutional kernel group are calibrated dynamically using the branch vessel anatomical diameter, real-time pressure gradient, and fractional flow reserve in the coronary artery supply area. Multi-scale feature extraction is performed on the asymmetric growth index and composite quantization index to generate a vascular topology weighted spatiotemporal superposition feature tensor. Secondly, the system obtains a dynamic threshold by calculating the moving average of the vascular endothelial shear stress waveform and the plaque morphology stability index. This dynamic threshold is used to adjust the spatiotemporal superposition feature tensor of the vascular topology weighted tensor for layer-by-layer activation. The system detects when the feature activation intensity of at least two adjacent vascular segments exceeds their corresponding... The critical threshold for branch vessel lesion classification triggered spatiotemporal coupling state labeling. Furthermore, the system extracted the covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate, and microcirculation resistance index from the lesion classification parameter library, established a dynamic matching rule engine, calculated the probability of acute exacerbation using logistic regression, predicted the compensation failure time window using a Kaplan-Meier modified model, and evaluated the sensitivity of targeted intervention using a random forest classifier. Finally, the system performed three-dimensional normalized encoding of the acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity to generate a directional risk transition vector, guiding physicians to take emergency measures and successfully preventing the condition from worsening.

[0057] To further improve the accuracy of real-time fusion of plaque morphological mutation features from intravascular ultrasound and circulatory auxiliary parameters from extracorporeal life support devices, and to reconstruct the priority queue of treatment plans and dynamically adjust the temporal dependence of revascularization surgery based on plaque rupture risk threshold and myocardial oxygen consumption balance equation, in some embodiments, the executable path topology implementation process described in step 104 includes: constructing plaque morphological mutation feature parameters through intravascular ultrasound image sequences; extracting plaque fibrous cap thickness gradient, calcification density distribution, and lipid core volume ratio using a multi-scale convolutional neural network; combining the arterial pressure waveform and ventricular auxiliary flow parameters acquired in real time by extracorporeal life support devices to generate a plaque morphology-hemodynamic coupling feature vector; inputting the plaque morphology-hemodynamic coupling feature vector into a plaque rupture risk assessment model based on a random forest classifier, and simultaneously calculating myocardial oxygen consumption... The product parameter of coronary perfusion pressure and ventricular wall tension in the oxygen balance equation activates the revascularization surgery priority calculation channel when the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range. A weighted correlation matrix between the plaque rupture risk gradient and the degree of myocardial oxygen consumption imbalance is established in this channel. A sliding time window algorithm is used to dynamically sort the weighted correlation matrix. Combined with catheterization lab resource occupancy status and anticoagulant drug metabolic half-life parameters, a time window and initiation time priority queue are generated. Based on the spatiotemporal constraints of thrombolysis and interventional procedures in the time window and initiation time priority queue, a surgical timing dependency is established. A recurrent neural network is used to predict the myocardial salvage index change curves under different timing combinations in the surgical timing dependency, and the revascularization scheme corresponding to the point of maximum slope descent is selected as the optimal path.

[0058] In this embodiment, plaque morphological mutation feature parameters are derived from a dataset constructed using intravascular ultrasound imaging sequences. A multi-scale convolutional neural network is employed to extract key features such as plaque fibrous cap thickness gradient, calcification density distribution, and lipid core volume ratio. These features are combined with arterial pressure waveforms and ventricular assist flow parameters acquired in real-time by extracorporeal life support equipment to generate a plaque morphology-hemodynamic coupling feature vector. This ensures the capture of the complex relationship between plaque morphology and hemodynamics, providing crucial information for subsequent analysis. The plaque rupture risk assessment model is a computational framework based on a random forest classifier, used to simultaneously calculate the product of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the priority calculation channel for revascularization surgery is activated, providing a comprehensive risk assessment to help physicians develop the optimal treatment plan. The weighted correlation matrix is ​​a data structure established in the priority calculation channel for revascularization surgery. It uses a sliding time window algorithm to dynamically sort the plaque rupture risk gradient and the degree of myocardial oxygen consumption imbalance. This matrix, combined with catheterization lab resource occupancy status and anticoagulant drug metabolic half-life parameters, generates time windows and a priority queue for initiation time, ensuring that each step is performed at the optimal time. The surgical sequence dependency is a sequence rule established based on the spatiotemporal constraints of thrombolysis and interventional procedures within the time window and priority queue for initiation time. A recurrent neural network predicts the myocardial salvage index change curves under different combinations, selecting the revascularization scheme corresponding to the point of maximum slope descent as the optimal path to ensure the best treatment effect in the shortest time.

[0059] In this embodiment, firstly, plaque morphological mutation feature parameters are constructed using intravascular ultrasound imaging sequences. A multi-scale convolutional neural network is then used to extract the plaque fibrous cap thickness gradient, calcification density distribution, and lipid core volume ratio. Combined with arterial pressure waveforms and ventricular assist flow parameters acquired in real-time by an extracorporeal life support device, a plaque morphology-hemodynamic coupling feature vector is generated. Secondly, this feature vector is input into a plaque rupture risk assessment model based on a random forest classifier. Simultaneously, the product parameter of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the model is activated. The system first establishes a priority calculation channel for revascularization surgery. Then, it establishes a weighted correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance within this channel. A sliding time window algorithm is used to dynamically sort the weighted correlation matrix. Combined with catheterization lab resource occupancy status and anticoagulant drug metabolic half-life parameters, a priority queue for time windows and initiation times is generated. Finally, based on the spatiotemporal constraints of thrombolysis and interventional procedures within the time windows and initiation time priority queues, a surgical timing dependency is established. A recurrent neural network is used to predict the myocardial salvage index change curves under different combinations, and the revascularization scheme corresponding to the point of maximum slope descent is selected as the optimal path.

[0060] Here is a specific example: Suppose a large hospital's chest pain center receives a patient experiencing severe nocturnal chest pain. First, plaque morphology mutation characteristic parameters are constructed using intravascular ultrasound imaging sequences. A multi-scale convolutional neural network is then used to extract the plaque fibrous cap thickness gradient, calcification density distribution, and lipid core volume ratio. Combined with real-time arterial pressure waveforms and ventricular assist flow parameters acquired by extracorporeal life support equipment, a plaque morphology-hemodynamic coupling feature vector is generated. Second, this feature vector is input into a plaque rupture risk assessment model based on a random forest classifier. Simultaneously, the product of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the system is activated. A priority calculation channel for revascularization surgery was established. Furthermore, a weighted correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance was created within this channel. A sliding time window algorithm was used to dynamically sort the weighted correlation matrix. Combined with catheterization lab resource occupancy status and anticoagulant drug metabolic half-life parameters, a priority queue for time windows and initiation times was generated. Finally, based on the spatiotemporal constraints of thrombolysis and interventional procedures within the time windows and initiation time priority queues, a surgical timing dependency was established. A recurrent neural network was used to predict the myocardial salvage index variation curves under different combinations. The revascularization scheme corresponding to the point of maximum slope decrease was selected as the optimal path, significantly improving patient treatment outcomes and survival rates.

[0061] To further improve the accuracy of the plaque rupture risk assessment model and dynamically adjust the priority and temporal dependence of revascularization surgery, in some embodiments, step 104 involves inputting the plaque morphology-hemodynamic coupling feature vector into the plaque rupture risk assessment model based on a random forest classifier. Simultaneously, the product parameter of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the priority calculation channel for revascularization surgery is activated. This includes: dynamically Bayesian tuning of the hyperparameters of the random forest classifier; constructing a feature importance ranking based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform; selecting the interaction term between calcification density distribution and ventricular auxiliary flow parameters as key input features; and generating a dynamic threshold curve for the plaque rupture risk score. Based on the dynamic threshold curve of the plaque rupture risk score, coronary perfusion is simultaneously calculated. The product parameter of pressure and ventricular wall tension is used to extract the fluctuation amplitude and baseline offset of the product parameter through a sliding time window algorithm. A two-dimensional deviation index is established by combining the dynamic threshold curve of the plaque rupture risk score. When the two-dimensional deviation index exceeds the preset warning line three times consecutively, the activation command of the revascularization surgery priority calculation channel is triggered. In the priority calculation channel of the revascularization surgery priority calculation channel activation command, a spatiotemporal correlation matrix between the plaque rupture risk gradient and the degree of myocardial oxygen consumption imbalance is constructed. A dynamic weighted algorithm is used to fuse the catheterization lab resource occupancy status parameters and the decay curve of the anticoagulant drug metabolic half-life to generate a dynamic decision map. The interventional surgery initiation time in the dynamic decision map is input into a recurrent neural network for temporal conflict detection. The optimal interventional path is selected through a myocardial salvage index prediction model. The plaque morphological mutation feature parameters and circulatory auxiliary parameters during the path execution process are fed back to the rupture risk assessment model for incremental learning and updating.

[0062] In this embodiment, the dynamic threshold curve for plaque rupture risk score is a data structure generated by dynamically Bayesian tuning of the hyperparameters of a random forest classifier. Feature importance ranking is constructed based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform, and the interaction term between calcification density distribution and ventricular auxiliary flow parameters is selected as a key input feature. This curve is used to generate the dynamic threshold for plaque rupture risk score, ensuring accurate identification of potential risk areas and timely intervention. The two-dimensional deviation index is a data index established by simultaneously calculating the product parameter of coronary perfusion pressure and ventricular wall tension based on the dynamic threshold curve of plaque rupture risk score. The fluctuation amplitude and baseline offset of the product parameter are extracted using a sliding time window algorithm, combined with the dynamic threshold curve of plaque rupture risk score. When the two-dimensional deviation index exceeds the preset warning line three times consecutively, the activation command for the revascularization surgery priority calculation channel is triggered, ensuring accurate identification of potential risk areas at different times and locations. The spatiotemporal correlation matrix is ​​a data structure constructed in the priority calculation channel for revascularization surgery. It employs a dynamic weighted algorithm to fuse catheterization lab resource occupancy status parameters with the decay curves of anticoagulant drug metabolism half-life, generating a dynamic decision map. This matrix ensures that each step is performed at the optimal time, optimizing the treatment pathway. Recurrent neural network temporal conflict detection is a detection method that inputs the interventional surgery initiation time from the dynamic decision map into a recurrent neural network. The optimal interventional path is selected through a myocardial salvage index prediction model, and plaque morphological mutation feature parameters and circulatory auxiliary parameters during path execution are fed back to the rupture risk assessment model for incremental learning and updates, ensuring the system can continuously improve and optimize treatment plans.

[0063] In this embodiment, the hyperparameters of the random forest classifier are first dynamically Bayesian tuned. Feature importance ranking is constructed based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform. The interaction term between calcification density distribution and ventricular auxiliary flow parameters is selected as a key input feature to generate a dynamic threshold curve for plaque rupture risk scoring. Secondly, based on this curve, the product parameter of coronary perfusion pressure and ventricular wall tension is simultaneously calculated. The fluctuation amplitude and baseline offset of the product parameter are extracted using a sliding time window algorithm. A two-dimensional deviation index is established by combining this with the dynamic threshold curve of the plaque rupture risk score. If the two-dimensional deviation index exceeds a preset warning level three times consecutively... The activation command for the revascularization surgery priority calculation channel is triggered at the time of the procedure. Next, a spatiotemporal correlation matrix between the plaque rupture risk gradient and the degree of myocardial oxygen consumption imbalance is constructed within the revascularization surgery priority calculation channel. A dynamic weighted algorithm is used to fuse catheterization lab resource occupancy status parameters with the decay curve of the anticoagulant drug's metabolic half-life, generating a dynamic decision map. Finally, the interventional surgery initiation time from the dynamic decision map is input into a recurrent neural network for temporal conflict detection. The optimal interventional path is selected using a myocardial salvage index prediction model, and the plaque morphological mutation feature parameters and cyclic auxiliary parameters during path execution are fed back to the rupture risk assessment model for incremental learning and updating.

[0064] Here is a specific example: Suppose a patient is rushed to the emergency room due to severe chest pain. First, the hyperparameters of the random forest classifier are dynamically Bayesian-tuned. A feature importance ranking is constructed based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform. The interaction term between calcification density distribution and ventricular assist flow parameters is selected as a key input feature, generating a dynamic threshold curve for plaque rupture risk scoring. Second, based on this curve, the product parameter of coronary perfusion pressure and ventricular wall tension is simultaneously calculated. The fluctuation amplitude and baseline offset of the product parameter are extracted using a sliding time window algorithm. A two-dimensional deviation index is established based on the dynamic threshold curve of the plaque rupture risk score. When the two-dimensional deviation index exceeds a preset warning line three times consecutively... This triggered the activation command of the revascularization surgery priority calculation channel. Next, a spatiotemporal correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance was constructed within the revascularization surgery priority calculation channel. A dynamic weighted algorithm was used to fuse catheterization lab resource occupancy status parameters and anticoagulant drug metabolic half-life decay curves to generate a dynamic decision map. Finally, the interventional surgery initiation time from the dynamic decision map was input into a recurrent neural network for temporal conflict detection. The optimal interventional path was selected using a myocardial salvage index prediction model, and the plaque morphological mutation characteristic parameters and cyclic auxiliary parameters during path execution were fed back to the rupture risk assessment model for incremental learning and updating, significantly improving patient treatment outcomes and survival rates.

[0065] To further improve the accuracy of spatiotemporal superposition analysis of interventional treatment delay parameters and dual antiplatelet drug metabolism curves, and to generate causal correlation maps to optimize clinical treatment pathways, in some embodiments, step 105 involves spatiotemporally superimposing the interventional treatment delay parameters and dual antiplatelet drug metabolism curves according to the time-series dependency relationship to generate a causal correlation map. The catheterization lab activation delay and drug efficacy verification results in the causal correlation map are then inversely mapped to the weighted enhancement model of the clinical decision baseline framework, driving self-optimization and updating of the pathway topology. This constructs a clinical treatment pathway generation strategy, including: using time-series alignment to fuse interventional treatment delay parameters and dual antiplatelet drug metabolism curves; and extracting the drug concentration decay gradient and catheterization lab preparation time through a sliding time window during the time-series alignment process. The phase difference parameters are used to generate a spatiotemporal superposition feature matrix. A causal relationship graph is constructed based on a Bayesian network, using the delay parameters and drug metabolism curve slopes in the spatiotemporal superposition feature matrix as graph nodes. The causal relationship graph between coronary artery blood flow recovery rate and platelet inhibition rate is identified by dynamic directed edge weights. The catheterization lab activation delay parameters and drug efficacy verification results in the causal relationship graph are input into a random forest regression model to calculate the dynamic adjustment coefficients of the weight reinforcement model in the clinical decision baseline framework. The association weights between coronary artery calcium integral and high-sensitivity troponin curve slope are updated through a gradient descent model. A path topology self-optimization model is constructed using a deep reinforcement learning framework, and the association weights are dynamically coupled with real-time intravascular ultrasound plaque morphology parameters to generate a clinical diagnosis and treatment path generation strategy.

[0066] In this embodiment, the spatiotemporal superposition feature matrix is ​​a data structure generated by merging interventional treatment delay parameters and dual antiplatelet drug metabolism curves through time-series alignment. A sliding time window is used to extract the phase difference parameter between the drug concentration decay gradient and the catheterization lab preparation time. This matrix can capture the complex relationship between drug metabolism and catheterization lab preparation time, providing important evidence for subsequent analysis. The causal association map is constructed based on a Bayesian network, using the delay parameters in the spatiotemporal superposition feature matrix and the slope of the drug metabolism curve as map nodes. Dynamically directed edge weights are used to identify the causal relationship between coronary artery blood flow recovery rate and platelet inhibition rate. This map helps identify the interactions between different factors, guiding clinical decision-making. The dynamic adjustment coefficient is calculated using a random forest regression model to enhance the weighted model within the clinical decision-making baseline framework. A gradient descent model is used to update the association weights between the coronary artery calcium integral and the slope of the high-sensitivity troponin curve. This method ensures accurate identification of potential risk areas at different times and locations, allowing for timely intervention. The path topology self-optimization model is constructed using a deep reinforcement learning framework, which dynamically couples the associated weights with real-time intravascular ultrasound plaque morphology parameters to generate a clinical treatment path generation strategy. This model can achieve seamless connection between multiple treatment steps, ensuring that each step is performed at the optimal time, thereby improving the overall treatment effect.

[0067] In this embodiment, firstly, time-series alignment is used to fuse interventional treatment delay parameters and dual antiplatelet drug metabolism curves. A sliding time window during the time-series alignment process is used to extract the phase difference parameters between the drug concentration decay gradient and the catheterization lab preparation time, generating a spatiotemporal superposition feature matrix. Secondly, a causal relationship graph is constructed based on a Bayesian network. The delay parameters and the slope of the drug metabolism curve in the spatiotemporal superposition feature matrix are used as graph nodes. Dynamic directed edge weights are used to identify the causal relationship between coronary artery blood flow recovery rate and platelet inhibition rate. Thirdly, the catheterization lab activation delay parameters and drug efficacy verification results from the causal relationship graph are input into a random forest regression model to calculate the dynamic adjustment coefficients of the weighted reinforcement model in the clinical decision baseline framework. The correlation weights between the coronary artery calcium integral and the slope of the high-sensitivity troponin curve are updated using a gradient descent model. Finally, a path topology self-optimization model is constructed using a deep reinforcement learning framework. The correlation weights are dynamically coupled with real-time intravascular ultrasound plaque morphology parameters to generate a clinical treatment path generation strategy.

[0068] Here is a specific example: Suppose a patient is admitted to the hospital due to persistent chest tightness and palpitations. The system first uses time-series alignment to fuse interventional treatment delay parameters with the metabolic curves of dual antiplatelet drugs. Through a sliding time window during the time-series alignment process, the phase difference parameters between the drug concentration decay gradient and the catheterization lab preparation time are extracted, generating a spatiotemporal superposition feature matrix. Secondly, the system constructs a causal relationship graph based on a Bayesian network. Using the delay parameters and the slope of the drug metabolism curves in the spatiotemporal superposition feature matrix as graph nodes, dynamic directed edge weights are used to identify the causal relationship between coronary artery blood flow recovery rate and platelet inhibition rate. Secondly, the system inputs the catheterization lab activation delay parameters and drug efficacy verification results from the causal association map into a random forest regression model to calculate the dynamic adjustment coefficients of the weighted reinforcement model in the clinical decision-making baseline framework. The system also updates the association weights between the coronary artery calcium integral and the slope of the high-sensitivity troponin curve using a gradient descent model. Finally, the system uses a deep reinforcement learning framework to construct a path topology self-optimization model, dynamically coupling the association weights with real-time intravascular ultrasound plaque morphology parameters to generate a clinical treatment path generation strategy, which significantly improves patient treatment outcomes and survival rates.

[0069] Figure 2 This application provides a schematic diagram of a clinical treatment pathway generation system for tracking patients with chest pain, as shown in the embodiments of this application. Figure 2 As shown, the device includes: The acquisition module 21 is used to establish a spatiotemporal data map with dynamic priority labels by synchronously acquiring the respiratory rhythm phase difference, coronary artery calcium integral and high-sensitivity troponin curve slope within a preset time after the onset of chest pain, and to implement incremental weight enhancement on the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision baseline framework. The identification module 22 is used to identify the spatiotemporal correlation pattern between myocardial ischemia area and abnormal electrocardiogram lead based on the spatiotemporal data map. When the asymmetric increase of ST segment elevation amplitude is continuously detected, the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment path. Module 23 is used to dynamically bind blood vessels in the acute phase of the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, and construct the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis, based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path. At the same time, in the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface. The fusion module 24 is used to fuse the plaque morphological mutation characteristics of intravascular ultrasound with the circulatory auxiliary parameters of the extracorporeal life support device in real time during the implementation of the executable path topology structure, and reconstruct the priority queue of treatment plans and dynamically adjust the temporal dependency of revascularization surgery based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation. The mapping module 25 is used to spatiotemporally superimpose the interventional treatment delay parameters and the metabolism curve of dual antiplatelet drugs according to the time-series dependency relationship to generate a causal relationship map, and to inversely map the catheterization laboratory activation delay and drug efficacy verification results in the causal relationship map to the weighted reinforcement model of the clinical decision baseline framework, driving the self-optimization update of the path topology structure and constructing a clinical diagnosis and treatment path generation strategy.

[0070] Figure 2 The aforementioned clinical treatment pathway generation system for tracking patients with chest pain can execute... Figure 1 The implementation principle and technical effects of the clinical treatment pathway generation method for tracking patients with chest pain described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit performs its operations in the clinical treatment pathway generation system for tracking patients with chest pain described in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0071] In one possible design, Figure 2 The clinical treatment pathway generation system for tracking patients with chest pain, as shown in the embodiment, can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0072] The processing component 32 is used to: establish a spatiotemporal data map with dynamic priority labels by synchronously collecting the respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve slope within a preset time after the onset of chest pain; and to implement incremental weight enhancement on the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision baseline framework; based on the spatiotemporal data map, identify the spatiotemporal correlation pattern between myocardial ischemia areas and ECG lead abnormalities; when asymmetric growth of ST segment elevation amplitude is continuously detected, activate the coupling analysis linked to the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate to generate a risk transition signal driving the treatment pathway; and, based on the intensity distribution of the risk transition signal driving the treatment pathway, dynamically bind blood vessels in the acute phase of the intensity distribution, and construct an executable path extension based on the temporal-spatial constraint relationship between the recanalization time window and thrombolysis contraindications. The system employs a multidisciplinary parameter negotiation interface to establish dynamic phase synchronization rules between enhanced CT scan time nodes and anticoagulant drug dose curves during the subacute phase of the intensity distribution. During the implementation of the executable path topology, it integrates plaque morphological mutation characteristics from intravascular ultrasound with circulatory auxiliary parameters from extracorporeal life support devices in real time. Based on plaque rupture risk thresholds and myocardial oxygen consumption balance equations, it reconstructs the treatment priority queue and dynamically adjusts the temporal dependence of revascularization surgery. Interventional treatment delay parameters and dual antiplatelet drug metabolism curves are spatiotemporally superimposed according to the temporal dependence to generate a causal correlation map. The catheterization lab activation delay and drug efficacy verification results in the causal correlation map are then inversely mapped to the weighted reinforcement model of the clinical decision baseline framework, driving the self-optimization update of the path topology and constructing a clinical diagnosis and treatment path generation strategy.

[0073] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0074] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0075] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0076] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0077] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0078] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0079] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a method for generating clinical treatment pathways to track patients with chest pain.

[0080] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0081] The device embodiments described above are merely illustrative. 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; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0082] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for generating clinical treatment pathways to track patients with chest pain, characterized in that, include: By synchronously collecting respiratory rhythm phase difference, coronary artery calcium integral and high-sensitivity troponin curve slope within a preset time after chest pain onset, a spatiotemporal data map with dynamic priority labeling is established, and incremental weighting is applied to the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision-making baseline framework. Based on the spatiotemporal data map, the spatiotemporal correlation pattern between myocardial ischemia area and ECG lead abnormalities is identified. When the asymmetric increase of ST segment elevation amplitude is continuously detected, the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment pathway. Based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, the blood vessels are dynamically bound in the acute phase of the intensity distribution, and the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis is constructed to build an executable path topology. At the same time, in the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface. During the implementation of the executable path topology, the plaque morphological mutation characteristics of intravascular ultrasound and the circulatory auxiliary parameters of the extracorporeal life support device are integrated in real time. Based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation, the priority queue of treatment plans is reconstructed and the temporal dependence of revascularization surgery is dynamically adjusted. Interventional treatment delay parameters and dual antiplatelet drug metabolism curves are spatiotemporally superimposed according to the time-series dependency to generate a causal correlation map. The catheterization lab activation delay and drug efficacy verification results in the causal correlation map are then inversely mapped to the weighted reinforcement model of the clinical decision-making baseline framework to drive the self-optimization and updating of the path topology and construct a clinical diagnosis and treatment path generation strategy.

2. The method according to claim 1, characterized in that, Based on the spatiotemporal data map, the spatiotemporal correlation pattern between myocardial ischemia areas and ECG lead abnormalities is identified. When asymmetric increases in ST segment elevation amplitude are continuously detected, a coupling analysis linked to the left ventricular ejection fraction decay gradient and B-type natriuretic peptide release rate is activated to generate risk transition signals driving the diagnosis and treatment pathway, including: The myocardial ischemia region in the spatiotemporal data map is located in three dimensions, and the abnormal fluctuation phase features of each lead channel are extracted. By aligning the multimodal signals in the abnormal fluctuation phase features with time series, a dynamic propagation relationship model is established. A dynamic slip monitoring window is constructed for the dynamic propagation relationship model. The ST segment elevation amplitude is decomposed into waveforms, and peak features, slope change rate and spatial lead distribution heterogeneity parameters are extracted. When the asymmetric growth index is detected to exceed the dynamic adjustment threshold within three consecutive windows, the cardiac function compensation reserve analysis channel is activated. The spatiotemporal trajectory of the left ventricular ejection fraction decay gradient and the fluctuation characteristics of the B-type natriuretic peptide release rate in the cardiac function compensation reserve analysis channel are acquired simultaneously. Through the multi-source data coupling engine of electrocardiogram-ultrasound-biomarkers, a dynamic weight model of ejection fraction change rate and B-type natriuretic peptide release rate is established, and the composite quantitative index is calculated. The asymmetric growth index and the composite quantitative index are subjected to spatiotemporal superposition analysis. When the spatiotemporal distribution of the two in the coronary artery blood supply area is detected to be in a critical coupling state, a directional risk transition vector is generated based on the dynamic matching rules of vascular lesion classification, inputting three-dimensional feature parameters such as acute exacerbation probability, compensation failure time window and targeted intervention sensitivity. Multi-level cross-validation is performed on the three-dimensional feature parameters. When the probability of acute exacerbation during the multi-level cross-validation process exceeds the standard deviation range of the preset baseline value, and the compensation failure time window overlaps with the current diagnosis and treatment stage, priority is given to intervening in the target coordinates and path switching time constraints to generate a risk transition signal to drive the diagnosis and treatment path switching.

3. The method according to claim 1, characterized in that, The method involves dynamically binding blood vessels during the acute phase of the intensity distribution based on the intensity distribution of the risk transition signal driving the treatment pathway, establishing a temporal-spatial constraint relationship between the recanalization time window and thrombolysis contraindications, and constructing an executable pathway topology. Simultaneously, during the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface, including: Based on the intensity distribution of the transition trigger signal, the evolution characteristics of thrombolysis contraindications and the nonlinear growth curve of vascular occlusion rate are extracted, a two-dimensional constraint matrix is ​​constructed, the time axis of the two-dimensional constraint matrix is ​​used to map the change of reperfusion injury probability gradient, and the spatial axis of the two-dimensional constraint matrix is ​​used to associate the anatomical topology of coronary artery branches. Based on the gradient change characteristics in the two-dimensional constraint matrix, a dynamic matching rule between the CT angiography time window and the frequency of cardiac ultrasound examinations is established during the acute phase. The priority ranking of imaging examination nodes is driven by the real-time decay gradient of the hemodynamic parameters of the dynamic matching rule to generate a dynamic sequence chain of imaging examinations. The temporal distribution characteristics of the dynamic sequence chain of the imaging examination are injected into the thrombolysis contraindication matrix. The slope of the change in vascular occlusion rate in the thrombolysis contraindication matrix is ​​dynamically coupled with the myocardial survival threshold. The catheterization lab preparation stage, the anticoagulant loading time window, and the balloon dilation operation sequence are introduced to generate a parallel execution framework. During the subacute phase, the operation sequence in the parallel execution framework is spatiotemporally encoded with the blood flow reserve parameters of the enhanced CT scan. By inversely matching the metabolic rate of the anticoagulant drug with the contrast agent clearance curve, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established.

4. The method according to claim 2, characterized in that, The process involves performing a spatiotemporal overlay analysis of the asymmetric growth index and the composite quantification index. When a critical coupling state is detected in the spatiotemporal distribution of the two in the coronary artery supply area, based on the dynamic matching rules of vascular lesion classification, and inputting three-dimensional feature parameters such as acute exacerbation probability, compensation failure time window, and targeted intervention sensitivity, a directional risk transition vector is generated, including: A spatiotemporal convolution kernel group is introduced, and the parameters of the spatiotemporal convolution kernel group are calibrated by dynamic calibration of the branch vessel anatomical diameter, real-time pressure gradient and fractional flow reserve in the coronary artery supply area. Multi-scale feature extraction is performed on the asymmetric growth index and the composite quantization index to generate a vessel topology weight spatiotemporal superposition feature tensor. A dynamic threshold is obtained by calculating the moving average of the vascular endothelial shear stress waveform and the plaque morphological stability index. The vascular topology weight spatiotemporal superposition feature tensor is activated layer by layer by adjusting the dynamic threshold. When the feature activation intensity of at least two adjacent vascular segments exceeds the critical threshold of their corresponding branch vascular lesion classification, a spatiotemporal coupling state marker is triggered. The covariance matrices of lesion calcification density gradient, fibrous cap thickness attenuation rate, and microcirculation resistance index were extracted from the vascular lesion classification parameter library. A dynamic matching rule engine was established, and logistic regression was used to calculate the probability of acute exacerbation. The time window of compensation failure was predicted by the Kaplan-Meier modified model, and the sensitivity of targeted intervention was evaluated by the random forest classifier. The probability of acute exacerbation, the time window of compensation failure, and the sensitivity of targeted intervention were three-dimensionally normalized and encoded to generate a directional risk transition vector.

5. The method according to claim 1, characterized in that, During the implementation of the executable path topology, the plaque morphological mutation characteristics from intravascular ultrasound are fused in real time with the circulatory auxiliary parameters of the extracorporeal life support device. Based on the plaque rupture risk threshold and the myocardial oxygen consumption balance equation, the priority queue of treatment plans is reconstructed and the temporal dependency of revascularization surgery is dynamically adjusted, including: Plaque morphology mutation feature parameters were constructed by intravascular ultrasound imaging sequences. Multi-scale convolutional neural networks were used to extract plaque fibrous cap thickness gradient, calcification density distribution and lipid core volume ratio. Combined with arterial pressure waveform and ventricular assist flow parameters collected in real time by extracorporeal life support equipment, a plaque morphology-hemodynamic coupling feature vector was generated. The plaque morphology-hemodynamic coupling feature vector is input into the plaque rupture risk assessment model based on a random forest classifier. The product parameter of coronary artery perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated simultaneously. When the plaque rupture risk score exceeds the dynamic threshold and the oxygen consumption balance index deviates from the preset range, the blood revascularization surgery priority calculation channel is activated. In the blood revascularization surgery priority calculation channel, a weighted correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance is established. The sliding time window algorithm is used to dynamically sort the weighted correlation matrix. Combined with the catheterization lab resource occupancy status and anticoagulant drug metabolism half-life parameters, a time window and a priority queue for initiation time are generated. Based on the spatiotemporal constraints between the time window and the priority queue of thrombolysis and interventional procedures, a surgical timing dependency is established. The myocardial salvage index change curve under different timing combinations in the surgical timing dependency is predicted by a recurrent neural network, and the revascularization scheme corresponding to the maximum slope descent point is selected as the optimal path.

6. The method according to claim 5, characterized in that, The plaque morphology-hemodynamic coupling feature vector is input into a plaque rupture risk assessment model based on a random forest classifier. Simultaneously, the product parameter of coronary perfusion pressure and ventricular wall tension in the myocardial oxygen consumption balance equation is calculated. When the plaque rupture risk score exceeds a dynamic threshold and the oxygen consumption balance index deviates from a preset range, the priority calculation channel for revascularization surgery is activated, including: The hyperparameters of the random forest classifier are dynamically Bayesian tuned. The importance of features is ranked based on the covariance matrix of the plaque fibrous cap thickness gradient and the phase difference of the arterial pressure waveform. The interaction term between calcification density distribution and ventricular auxiliary flow parameters is selected as the key input feature to generate a dynamic threshold curve for plaque rupture risk score. Based on the dynamic threshold curve of the plaque rupture risk score, the product parameter of coronary artery perfusion pressure and ventricular wall tension is calculated synchronously. The fluctuation amplitude and baseline offset of the product parameter are extracted by the sliding time window algorithm. A two-dimensional deviation index is established in combination with the dynamic threshold curve of the plaque rupture risk score. When the two-dimensional deviation index exceeds the preset warning line three times in a row, the activation command of the blood revascularization surgery priority calculation channel is triggered. In the priority calculation channel of the activation command of the revascularization surgery priority calculation channel, a spatiotemporal correlation matrix between plaque rupture risk gradient and myocardial oxygen consumption imbalance is constructed. A dynamic weighted algorithm is used to fuse catheterization lab resource occupancy status parameters and anticoagulant drug metabolism half-life decay curves to generate a dynamic decision map. The interventional surgery initiation time in the dynamic decision graph is input into a recurrent neural network for temporal conflict detection. The optimal interventional path is selected through a myocardial salvage index prediction model. The plaque morphology mutation feature parameters and cyclic auxiliary parameters during the path execution process are fed back to the rupture risk assessment model for incremental learning and updating.

7. The method according to claim 1, characterized in that, The interventional treatment delay parameters and the metabolic curves of dual antiplatelet drugs are spatiotemporally superimposed according to the time-series dependency to generate a causal correlation map. The catheterization lab activation delay and drug efficacy verification results in the causal correlation map are then inversely mapped to the weighted reinforcement model of the clinical decision-making baseline framework to drive self-optimization and updating of the path topology, thus constructing a clinical treatment path generation strategy, including: The time series alignment and fusion of interventional treatment delay parameters and dual antiplatelet drug metabolism curves were used. The phase difference parameters between the drug concentration decay gradient and the catheterization lab preparation time were extracted through the sliding time window in the time series alignment process to generate a spatiotemporal superposition feature matrix. A causal relationship graph based on Bayesian network is constructed, using the delay parameter and the slope of the drug metabolism curve in the spatiotemporal superposition feature matrix as graph nodes, and the causal relationship graph between coronary artery blood flow recovery rate and platelet inhibition rate is identified by dynamic directed edge weights. The catheterization lab activation delay parameter and drug efficacy verification results in the causal association map are input into the random forest regression model to calculate the dynamic adjustment coefficient of the weighted enhancement model in the clinical decision baseline framework. The association weight between the coronary artery calcification integral and the slope of the high-sensitivity troponin curve is updated through the gradient descent model. A path topology self-optimization model is constructed using a deep reinforcement learning framework. The associated weights are dynamically coupled with real-time intravascular ultrasound plaque morphology parameters to generate a clinical diagnosis and treatment path generation strategy.

8. A clinical treatment pathway generation system for tracking patients with chest pain, characterized in that, include: The acquisition module is used to simultaneously acquire the respiratory rhythm phase difference, coronary artery calcium integral, and high-sensitivity troponin curve slope within a preset time after the onset of chest pain, establish a spatiotemporal data map with dynamic priority labels, and implement incremental weight enhancement on the rate of change of key indicators in the spatiotemporal data map to generate a clinical decision baseline framework. The identification module is used to identify the spatiotemporal correlation pattern between myocardial ischemia area and abnormal electrocardiogram leads based on the spatiotemporal data map. When the asymmetric increase of ST segment elevation amplitude is continuously detected, the coupling analysis linked with the left ventricular ejection fraction attenuation gradient and B-type natriuretic peptide release rate is activated to generate a risk transition signal driving the diagnosis and treatment path. The construction module is used to dynamically bind blood vessels in the acute phase of the intensity distribution of the risk transition signal of the driving diagnosis and treatment path, and construct the time-space constraint relationship between the recanalization time window and the contraindication of thrombolysis, based on the intensity distribution of the risk transition signal of the driving diagnosis and treatment path. At the same time, in the subacute phase of the intensity distribution, a dynamic phase synchronization rule between the enhanced CT scan time node and the anticoagulant drug dose curve is established through a multidisciplinary parameter negotiation interface. The fusion module is used to fuse the plaque morphological mutation characteristics of intravascular ultrasound with the circulatory auxiliary parameters of the extracorporeal life support device in real time during the implementation of the executable path topology. Based on the plaque rupture risk threshold and myocardial oxygen consumption balance equation, it reconstructs the priority queue of treatment plans and dynamically adjusts the temporal dependency of revascularization surgery. The mapping module is used to spatiotemporally superimpose interventional treatment delay parameters and dual antiplatelet drug metabolism curves according to the time-series dependency relationship to generate a causal relationship map. The catheterization lab activation delay and drug efficacy verification results in the causal relationship map are then inversely mapped to the weighted reinforcement model of the clinical decision baseline framework to drive the self-optimization update of the path topology and construct a clinical diagnosis and treatment path generation strategy.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a method for generating clinical treatment pathways for tracking patients with chest pain as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements a method for generating clinical treatment pathways for tracking patients with chest pain as described in any one of claims 1 to 7.

Images