Virtual patient physiological status real-time monitoring method and system based on digital twinning

Abstract

The invention relates to the field of physiological status real-time monitoring, provides a virtual patient physiological status real-time monitoring method and system based on digital twinning, and is used for solving the problems of non-linear combination deviation and multi-source data synchronization lag of virtual patient physiological status monitoring caused by dependence on a static model and one-way feedback. The method comprises the steps that a valve three-dimensional model is built based on aorta CT data, and a simulation model containing motion conditions is built in combination with a fluid mechanics equation; the blood flow pressure difference and vortex parameters of valve opening and closing are calculated in real time, an optical fiber virtual sensor is embedded in the topological surface of the valve to simulate blood flow deformation, and a valve state mechanical signal is output; parameters and signals are input into a digital twin engine, particle trajectories and refractive indexes are adjusted to generate a visual interface, blood flow velocity-pressure gradient spatio-temporal data are extracted, model boundary conditions are reversely corrected, and valve movement is synchronized with the physiological state of a patient. According to the technical scheme, heart valve movement is simulated in real time, and the real heartbeat rhythm of the patient is automatically matched.

Description

Technical Field

[0001] The present application relates to the technical field of real-time monitoring of physiological status, and in particular to a method and system for real-time monitoring of the physiological status of virtual patients based on digital twins. Background Art

[0002] In cardiovascular surgery rehearsal, it is necessary to simulate the intraoperative cardiac hemodynamic state in a virtual environment based on the patient's individual anatomical structure and real-time physiological parameters to evaluate the impact of the surgical plan on blood flow distribution, vascular wall stress and organ perfusion. Since cardiac blood flow is affected by the combined effects of physiological regulation and intraoperative instrument operation, it is necessary to build a virtual patient system that can integrate multimodal imaging data, real-time physiological feedback and hemodynamic models to support the prediction and optimization of complex scenarios such as blood flow mutations and thrombosis risks that may occur during surgery.

[0003] At present, a technical solution to this demand is a static hemodynamic simulation system based on three-dimensional reconstruction of medical images. The geometric model of the patient's heart and blood vessels is generated through preoperative CT or MRI images, and the blood flow velocity, pressure distribution and other parameters are simulated in combination with computational fluid dynamics. This solution drives the simulation by presetting fixed boundary conditions, and uses the finite element method to solve the hemodynamic equations to generate static prediction results of intraoperative blood flow status.

[0004] However, relying on preoperative static imaging data and hypothetical boundary conditions cannot reflect the impact of real-time physiological parameters on hemodynamics during surgery, resulting in significant deviations between simulation results and actual intraoperative blood flow parameters. In addition, static simulation lacks the ability to respond to the patient's individualized physiological feedback mechanism, making it difficult to capture instantaneous blood flow mutations or local turbulence phenomena triggered by surgical operations, limiting its reliability in complex surgical rehearsals. Summary of the invention

[0005] The present application provides a real-time monitoring method and system for the physiological state of virtual patients based on digital twins, which is used to solve the problems of nonlinear combination deviation and multi-source data synchronization lag caused by reliance on static models and unidirectional feedback in the virtual patient physiological state monitoring in the prior art.

[0006] In a first aspect, the present application provides a method for real-time monitoring of the physiological state of a virtual patient based on digital twins, comprising: Acquire aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, combine the three-dimensional geometric model with fluid mechanics control equations, and establish a simulation model including valve motion conditions; Based on the simulation model, the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve are calculated in real time; Embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor to generate a mechanical signal reflecting the opening and closing state of the valve; The pressure difference, eddy current parameters and mechanical signals are input into a rendering engine in the digital twin environment, and a visualization interface is generated by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; The spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface is obtained to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

[0007] Optionally, based on the simulation model, real-time calculation of pressure difference parameters and eddy flow parameters of blood flow during the opening and closing of the aortic valve includes: Extracting geometric deformation data of the aortic valve during opening and closing from the simulation model, and dividing a dynamic region synchronized with the opening and closing state of the valve in the three-dimensional geometric model based on the movement trajectory of the valve apex in the geometric deformation data; In the dynamic region, according to the difference in the movement trajectory of the valve apex, the pressure difference between adjacent valve apex is calculated, and the pressure distribution path of the valve surface is constructed through the pressure difference; Based on the sudden change position of the pressure difference in the pressure distribution path on the valve surface, marking the area where the blood flow direction suddenly changes as the vortex initial area, and counting the geometric center of the vortex initial area; Matching the geometric center of the vortex initial area to the corresponding time point according to the time series of the valve opening and closing state, calculating the displacement difference value of the geometric center between adjacent time points to determine the movement distance corresponding to each time point, and calculating the diffusion intensity based on the coverage change of the vortex initial area at different time points to generate vortex diffusion parameters including the movement distance and the diffusion intensity; The weighted superposition result between the intensity change of the pressure distribution path on the valve surface and the moving distance and diffusion intensity in the eddy flow diffusion parameter is used as a pressure difference parameter; The ratio of the moving distance to the diffusion intensity is taken as the eddy current parameter. Optionally, the geometric center of the vortex initial area is matched to the corresponding time point according to the time series of the valve opening and closing state, the displacement difference value of the geometric center between adjacent time points is calculated to determine the movement distance corresponding to each time point, and the diffusion intensity is calculated based on the coverage change of the vortex initial area at different time points to generate the vortex diffusion parameters including the movement distance and the diffusion intensity, including: Extracting continuous time points from the time series of the valve opening and closing states, and marking the geometric center position of the vortex initial area at each time point as the corresponding spatial coordinate; Arrange the spatial coordinates of adjacent time points in chronological order, calculate the straight-line distance difference of the spatial coordinates of the adjacent time points in three-dimensional space, and generate the displacement difference value of the geometric center between the adjacent time points; Based on the number of covered grids in the eddy initial region at different time points, the ratio of the number of covered grids between adjacent time points is calculated as the diffusion rate, wherein the number of covered grids is obtained by counting the total number of boundary grid units in the eddy initial region; The displacement difference value and the diffusion rate are respectively superimposed in time series to generate the movement distance and diffusion intensity corresponding to each time point; The result of the multiplication relationship between the moving distance and the diffusion intensity is used as the eddy current diffusion parameter including the moving distance and the diffusion intensity.

[0008] Optionally, obtaining the spatiotemporal relationship data between the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient includes: Extracting the spatiotemporal relationship data of blood flow velocity and pressure gradient from the visualization interface, decomposing the hemodynamic characteristics in the spatiotemporal relationship data, and generating blood flow characteristic parameters reflecting the actual physiological state of the patient; Comparing the blood flow characteristic parameters with the boundary conditions of the current valve motion in the simulation model, identifying the matching deviation between the blood flow characteristic parameters and the boundary conditions, and generating boundary condition correction requirements; Based on the boundary condition correction requirements, the correlation between the valve motion trajectory and the blood flow characteristic parameters in the three-dimensional geometric model is analyzed to determine the influence weight of the valve motion on the blood flow characteristic parameters; Adjusting the boundary conditions of the simulation model according to the influence weights, and generating boundary condition corrections by constraining the coupling relationship between the amplitude of the valve opening and closing state and the blood flow velocity; The boundary condition correction amount is injected into the simulation model to correct the simulation model, and the temporal and spatial consistency of the valve motion and blood flow characteristic parameters of the corrected simulation model is verified to generate a simulation model that is synchronously mapped with the patient's physiological state.

[0009] Optionally, a fiber optic virtual sensor is embedded in the valve surface topological structure of the three-dimensional geometric model to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor to generate a mechanical signal reflecting the opening and closing state of the valve, including: Determining the node distribution positions of the optical fiber virtual sensor in the curvature change area and the stress concentration area of ​​the valve surface topological structure; A plurality of pressure sensing nodes are set in the node distribution position, and a deformation constraint condition of each pressure sensing node under the action of blood flow is established by binding each pressure sensing node with the local geometric coordinates of the valve surface topological structure; Calculating the displacement of each pressure sensing node by using the deformation constraint condition, and converting the displacement into a three-dimensional space coordinate change sequence of the pressure sensing node; Extracting displacement change data of each pressure sensing node according to the three-dimensional space coordinate change sequence, and constructing a deformation characteristic curve of the node based on the displacement change data; The deformation characteristic curve is converted into a mechanical signal waveform, and the mechanical signal waveform is characterized by the displacement amplitude, velocity change rate and acceleration change rate of the pressure sensing node to generate a mechanical signal reflecting the opening and closing state of the valve.

[0010] Optionally, a three-dimensional geometric model of the aortic valve is constructed based on the CT image data, and the three-dimensional geometric model is combined with a fluid mechanics control equation to establish a simulation model including valve motion conditions, including: Extracting contour information of the aortic valve from the aortic CT image data, and constructing a preliminary three-dimensional geometric structure of the valve based on the contour information, wherein the structure includes morphological characteristics and spatial position of the valve; Refining the preliminary three-dimensional geometric structure, supplementing the detailed features of the valve, and forming a complete three-dimensional geometric model of the aortic valve; Based on the three-dimensional geometric model, the geometric characteristic parameters of the valve surface are extracted, and the motion constraint conditions during the opening and closing process of the valve are defined in combination with the physiological motion law of the valve in the three-dimensional geometric model; Combining the three-dimensional geometric model with fluid mechanics control equations, and constructing a dynamic model of the interaction between blood flow and valve according to valve motion constraints, wherein the dynamic model includes the pressure effect of blood flow on the valve and the response of the valve to the blood flow; The pressure effect parameters of blood flow on the valve and the response parameters of the valve to blood flow in the dynamic model are bidirectionally coupled, and the motion constraint conditions are combined with the morphological characteristics of the three-dimensional geometric model to construct a simulation model including valve motion conditions. Optionally, the three-dimensional geometric model is combined with the fluid mechanics control equations to construct a dynamic model of the interaction between blood flow and valve according to the valve motion constraint conditions, including: Extracting the valve surface topology of the three-dimensional geometric model, obtaining the curvature distribution characteristics and node connection relationship in the valve surface topology, and generating a surface mesh for coupling fluid and structure; Matching the boundary conditions of the surface grid with the fluid mechanics control equation, analyzing the relationship between the fluid pressure gradient pulsation characteristics in the fluid mechanics control equation and the deformation response of the valve surface topological structure, and generating a grid topology for coupling the fluid and the structure; Extracting the valve opening and closing phase and motion trajectory feature points in the valve motion constraint conditions, matching the valve opening and closing phase and motion trajectory feature points with the mesh topology of the fluid-structure coupling, and generating fluid domain deformation parameters and valve motion boundary conditions; Inputting the fluid domain deformation parameters and the valve motion boundary conditions into the fluid mechanics control equation, synchronously solving the fluid mechanics control equation and the valve kinematic constraints, and generating the pressure effect of the blood flow on the valve and the response data of the valve to the blood flow; The response data is mapped to the valve surface topology of the three-dimensional geometric model, the position offset of the valve surface topology is corrected according to the response data, and a dynamic model including the interaction between blood flow and valve is generated by synchronous iterative updating of the position offset and pressure action. In the second aspect, the present application provides a real-time monitoring system for the physiological state of a virtual patient based on digital twins, comprising: A construction module is used to obtain aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, and combine the three-dimensional geometric model with fluid mechanics control equations to establish a simulation model including valve motion conditions; A calculation module, used for calculating the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve in real time based on the simulation model; A generating module, used for embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model, so as to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor, and generate a mechanical signal reflecting the opening and closing state of the valve; An input module, used to input the pressure difference, eddy current parameters and mechanical signals into a rendering engine in the digital twin environment, and generate a visualization interface by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; The correction module is used to obtain the spatiotemporal relationship data between the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

[0011] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a real-time monitoring method for the physiological state of a virtual patient based on digital twins as described in the first aspect above.

[0012] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a real-time monitoring method for the physiological state of a virtual patient based on digital twins as described in the first aspect.

[0013] Beneficial effects of this application: In an embodiment of the present application, aortic CT image data of a virtual patient is obtained, a three-dimensional geometric model of the aortic valve is constructed based on the CT image data, the three-dimensional geometric model is combined with the fluid mechanics control equation, and a simulation model including valve motion conditions is established; based on the simulation model, the pressure difference and eddy parameters of blood flow during the opening and closing of the aortic valve are calculated in real time, and the eddy parameters are compared with a preset physiological threshold range; an optical fiber virtual sensor is embedded in the valve surface topological structure of the three-dimensional geometric model, and the optical fiber virtual sensor generates a mechanical signal reflecting the opening and closing state of the valve by simulating the deformation of the pressure sensing node under the action of blood flow; the pressure difference, eddy parameters and mechanical signals are input into the rendering engine in the digital twin environment, and a visualization interface is generated by adjusting the optical refractive index distribution of the blood flow particle trajectory and the pressure gradient field; the spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface are obtained, and the boundary conditions of the simulation model are corrected so that the valve motion of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

[0014] The present application obtains aortic CT image data of a virtual patient, constructs a three-dimensional geometric model of the aortic valve based on the CT image data, and combines the three-dimensional geometric model with the fluid mechanics control equation to establish a simulation model that includes valve motion conditions. This can achieve high-precision hemodynamic model construction based on individualized anatomical structures and improve the physiological matching degree of valve motion simulation. By calculating the pressure difference parameters and eddy flow parameters of blood flow during the opening and closing of the aortic valve in real time based on the simulation model, the key indicators of blood flow energy loss and turbulence characteristics can be dynamically captured. By embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model, the pressure sensing node can be simulated through the fiber optic virtual sensor. The deformation under the action of blood flow generates a mechanical signal reflecting the opening and closing state of the valve, and can establish a valve deformation feedback mechanism coupled with multiple physical fields; by inputting the pressure difference, eddy current parameters and mechanical signals into the rendering engine in the digital twin environment, adjusting the optical refractive index distribution of the blood flow particle trajectory and the pressure gradient field to generate a visualization interface, it is possible to achieve a holographic visualization expression of the blood flow dynamic process and mechanical characteristics; by obtaining the spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface, the boundary conditions of the simulation model are corrected, so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient, and a closed-loop optimized digital twin system can be constructed to ensure the dynamic consistency between the model prediction results and the real physiological behavior.

[0015] Furthermore, by extracting the geometric deformation data of the aortic valve opening and closing process and dividing the dynamic area based on the movement trajectory of the valve apex, the spatial synchronous mapping of the valve motion state and the dynamic characteristics of the three-dimensional model can be achieved, thereby improving the spatiotemporal accuracy of local blood flow dynamic analysis; by constructing a surface pressure distribution path driven by the pressure difference between adjacent valve apexes, the conduction law and energy distribution characteristics of the valve surface pressure gradient under the impact of blood flow can be revealed; by identifying the pressure mutation area to mark the initial position of the vortex and counting the geometric center, the turbulent core area caused by the sudden change in blood flow direction can be accurately located; by matching the vortex center in time series The displacement difference and coverage range change can quantify the coupling relationship between the energy transfer intensity and the spatial evolution trajectory during the eddy diffusion process; the pressure difference parameter is generated by superimposing the weighted results of the pressure distribution intensity and the eddy diffusion parameter, which can integrate the multi-dimensional mechanical characteristics of the local pressure gradient and the dynamic propagation of the eddy current; by establishing the ratio of the moving distance to the diffusion intensity as the eddy current parameter, the interference of the time dimension can be eliminated and a quantitative indicator characterizing the eddy energy dissipation efficiency can be constructed, ultimately forming a hemodynamic characteristic evaluation system that takes into account both spatial heterogeneity and temporal continuity, providing a high-resolution dynamic quantitative basis for the diagnosis of valvular dysfunction.

[0016] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flowchart of a method for real-time monitoring of the physiological state of a virtual patient based on digital twins provided in the present application is shown; Figure 2 A schematic diagram of the structure of a real-time monitoring system for the physiological state of a virtual patient based on digital twins provided in the present application is shown; Figure 3 A schematic diagram of the structure of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0020] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0021] Researchers have found that existing aortic valve hemodynamic simulation models have difficulty in mapping changes in the patient's physiological state in real time, and lack dynamic perception of the mechanical characteristics of the valve surface, resulting in large deviations between the simulation results and the actual blood flow parameters. Based on this, a method for modeling aortic valve digital twins is provided, which can achieve synchronous visualization analysis of valve motion and blood flow parameters through dynamic coupling of fiber optic virtual sensing and fluid mechanics. The technical solution of this application can be applied to preoperative planning of cardiovascular diseases and digital twin-assisted diagnosis and treatment scenarios.

[0022] The entire R&D process reflects the technical path of multimodal data fusion and dynamic boundary condition collaborative optimization, aiming to overcome the defects of the existing solutions, such as static modeling cannot adapt to physiological state fluctuations and simulation parameters are disconnected from real mechanical feedback. This method significantly improves the clinical credibility of aortic valve hemodynamic simulation through the closed-loop feedback mechanism of mechanical signals and fluid parameters, and provides a high-precision dynamic twin platform for personalized diagnosis and treatment.

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0024] Figure 1 A flowchart of a method for real-time monitoring of the physiological state of a virtual patient based on digital twins is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes: 101. Acquire aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, combine the three-dimensional geometric model with fluid mechanics control equations, and establish a simulation model including valve motion conditions; In this step, the aortic CT image data of the virtual patient refers to the digital aortic structure data set generated by medical image reconstruction technology. The three-dimensional geometric model refers to the three-dimensional structure model of the aortic valve established based on CT data. The fluid dynamics control equations refer to the Navier-Stokes equations and continuity equations that describe blood flow. The valve motion conditions refer to the boundary motion constraints that simulate the opening and closing of the valve leaflets.

[0025] In an embodiment of the present application, first, the aortic CT image data of a virtual patient is threshold segmented and morphologically optimized by medical image processing software (such as Mimics or 3D Slicer), the contour information of the aortic valve is extracted, and a two-dimensional slice layer is generated. Secondly, the two-dimensional slice layer is stacked into a three-dimensional geometric model using a three-dimensional reconstruction algorithm (such as MarchingCubes), and the surface is smoothed and the holes are repaired by a mesh optimization tool (such as MeshLab). Then, the surface mesh of the three-dimensional geometric model is imported into a fluid mechanics simulation software (such as ANSYS or OpenFOAM) to generate a body mesh (such as a dynamic layer mesh) containing the boundary conditions of the valve motion. Finally, the Navier-Stokes equations (fluid mechanics control equations) are combined with the periodic opening and closing conditions of the valve (such as the diastolic valve closure velocity) to establish a simulation model that can simulate blood flow.

[0026] In the preoperative evaluation of a 58-year-old patient with aortic stenosis, the medical team obtained the imaging data of the aortic root through ECG-gated CT scanning (layer thickness 0.5mm, matrix 512×512), and used adaptive meshing technology to reconstruct a three-dimensional geometric model containing three leaflets, accurately presenting the leaflet calcification area (thickness 2.1-3.5mm) and the irregular shape of the valve ring. The model was imported into the blood flow-structure coupling simulation platform, and the peak flow velocity of the left ventricular systolic inlet was set to 2.1m / s, and the aortic pressure waveform matched the patient's measured 135 / 85mmHg characteristics. The leaflet material stiffness gradient (elastic modulus of calcified area 8MPa, normal tissue 2MPa) was defined through a bidirectional fluid-solid coupling algorithm, and a simulation environment for dynamic simulation of valve opening and closing (maximum opening angle 68°) was established to provide a basis for subsequent blood flow analysis.

[0027] 102. Based on the simulation model, calculate in real time the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve; In this step, the pressure difference parameter refers to the transient pressure difference peak value upstream and downstream of the valve. The vortex parameters include turbulence characteristic parameters such as vortex intensity and vortex core position.

[0028] In the embodiment of the present application, first, the fluid mechanics solver parameters (such as the transient time step of 0.001 seconds and the Newtonian fluid assumption) are set in the simulation model, and the physical parameters of the blood (such as density and viscosity) are loaded. Secondly, the control equation is solved by the finite volume method, and the pressure difference parameters (such as the pressure difference on both sides of the valve orifice) and eddy current parameters (such as vortex intensity) at different positions during the opening and closing of the valve are calculated in real time. Then, the flow field data (such as the peak systolic pressure) at key time points are extracted using post-processing tools (such as Paraview) to generate a pressure-time curve and an eddy current distribution cloud map. Finally, the calculation results are stored as a spatiotemporal data set for sensor simulation and visualization.

[0029] When the simulation model runs to mid-systole (valve opening reaches the maximum angle), the system solves the fluid mechanics equations in real time, and detects a peak pressure difference of 55mmHg (corresponding to severe stenosis) downstream of the valve, and forms a vortex core area with a diameter of 7mm on the back of the right coronary leaflet, and its rotation intensity parameter reaches 4.3m² / s². Particle tracking found that the vortex caused the blood retention time to be prolonged to 0.28 seconds (normal should be less than 0.15 seconds), indicating an increased risk of thrombosis. At the same time, in the early diastolic regurgitation stage, abnormal reverse flow velocity (0.6m / s) caused by delayed valve closure was detected. These parameters are updated at a rate of 150 frames per second to provide dynamic input for the virtual sensor network.

[0030] 103. Embed a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model, so as to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor, and generate a mechanical signal reflecting the opening and closing state of the valve; In this step, the fiber optic virtual sensor refers to the mechanical signal acquisition logic unit embedded in the digital model. The mechanical signal includes the stress distribution on the leaflet surface and the strain rate change.

[0031] In an embodiment of the present application, first, key topological nodes (such as leaflet edges and hinge points) are selected in the valve surface mesh of the three-dimensional geometric model as the embedding locations of the fiber optic virtual sensor. Secondly, based on the pressure difference data calculated by fluid mechanics, the deformation of the sensor node under the action of blood flow is simulated by finite element analysis (such as ABAQUS) (such as the node displacement of 0.1 mm for every 10 Pa increase in pressure). Then, the deformation is converted into a virtual mechanical signal (such as a change in voltage or light intensity), and a timestamp is bound to reflect the opening and closing state of the valve (such as a sudden increase in signal amplitude when closed). Finally, a mechanical signal sequence with spatial position labels is output for dynamic rendering in a digital twin environment.

[0032] 320 fiber optic virtual sensors were implanted on the surface of each leaflet of the three-dimensional model to simulate the strain sensing ability of real fiber Bragg gratings. When the left coronary leaflet was bent and deformed by 0.15 mm due to blood flow impact, the corresponding sensor node generated a mechanical signal with a frequency offset of 1.8 pm. After signal processing, the pressure distribution map of the leaflet surface (peak pressure 95 kPa) was reconstructed. At the moment of diastolic closure, two adjacent sensors at the edge of the non-coronary leaflet detected a sudden change in frequency within 0.6 ms (increase of 2.4 pm), triggering an early warning of valve incomplete closure. These mechanical signals are strictly synchronized with the simulation clock and input into the visualization system with an accuracy of 2500 samples per second.

[0033] 104. Input the pressure difference, eddy current parameters and mechanical signals into a rendering engine in the digital twin environment, and generate a visualization interface by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; In this step, the digital twin environment refers to the virtual simulation space that integrates the physical model with real-time data. The optical refractive index distribution refers to the visualization parameters that adjust the light propagation path according to the pressure gradient.

[0034] In an embodiment of the present application, first, the pressure difference parameters, eddy current parameters and mechanical signals are imported into a digital twin rendering engine (such as Unity or Unreal Engine), and the blood flow trajectory is simulated through a particle system (such as red particles represent high-speed flow, and blue represents eddy current). Secondly, the optical refractive index distribution is adjusted according to the pressure gradient field data (such as the refractive index in the high-pressure area increases, and the particle path deflection is enhanced), and a dynamic color mapping is generated (such as warm colors represent high pressure, and cold colors represent low pressure). Next, the mechanical signal is bound to the surface deformation of the valve, and the dynamic deformation of the three-dimensional model is driven by the vertex shader (such as the displacement of the sensor node triggers the opening and closing animation of the leaflet). Finally, all elements are integrated to generate an interactive visualization interface that supports multi-perspective observation (such as a cross-sectional view showing the flow field, and a surface view showing the mechanical signal intensity).

[0035] The pressure difference parameter is mapped to the color gradient of blood flow particles (red>50mmHg, yellow 30-50mmHg), and the vortex parameter is converted into the spiral density of the particle motion trajectory (dense white vortex lines are displayed in high vortex areas). The mechanical signal intensity (80-120kPa) of the fiber optic virtual sensor is superimposed on the valve surface through a translucent halo layer, and the halo brightness is proportional to the local pressure. When reflux is detected, the interface automatically activates the profile cutting function, generates a dynamic pressure isosurface (interval 10mmHg) in the aortic sinus, and highlights the orange warning box in the closure delay area. Doctors can trace back key phases (such as the turning point of valve opening and closing) through the interactive timeline to analyze the correlation between abnormal blood flow and structural deformation.

[0036] 105. Acquire the spatiotemporal relationship data between the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

[0037] In this step, the spatiotemporal relationship data refers to the time evolution trajectory of multidimensional parameters in three-dimensional space. Synchronous mapping refers to the virtual-reality dynamic calibration achieved through data assimilation.

[0038] In the embodiment of the present application, first, a timeline control is embedded in the visualization interface, allowing the user to replay or pause the simulation process, and synchronously display the pressure difference curve, vortex distribution and valve deformation. Secondly, the amplitude of the mechanical signal is mapped to the halo effect on the valve surface (e.g., the stronger the signal, the brighter the halo) through shader programming (such as GLSL) to enhance state perception. Then, a data probe tool is added, and the user can click anywhere to view real-time pressure, flow rate and deformation data. Finally, the visualization results are exported as dynamic videos or interactive VR scenes for surgical planning or medical education demonstrations.

[0039] Through the visualization interface, it was found that the systolic blood flow velocity in the valve calcification area reached 5.2m / s (15% higher than the initial setting value of the simulation), and the spatiotemporal relationship data of the area was extracted (velocity gradient 14m / s², pressure oscillation ±22mmHg), and the simulation inlet boundary conditions were reversed: the left ventricular ejection waveform was adjusted from the ideal curve to a patient-specific bimodal morphology (peak value 2.3m / s@150ms, 1.8m / s@300ms). At the same time, according to the spatial distribution characteristics of the regurgitant vortex, the material damping parameters of the leaflet closing phase were adjusted (from 0.18N·s / m to 0.25N·s / m), so that the simulated closing delay time was shortened from 40ms to 28ms (close to the actual ultrasound measurement value). After three iterations, the deviation between the simulation results and the actual pressure difference measured by the cardiac catheter was reduced from the initial 18% to 3%, achieving accurate synchronization between the digital twin and the patient's physiological state.

[0040] In summary, steps 101 to 105 realize the full-process closed-loop system of dynamic mapping of virtual aortic valve from medical images to digital twins. By integrating CT image three-dimensional modeling, fluid dynamics simulation, virtual sensor embedding and visualization interface generation, a high-precision mapping relationship between valve motion and hemodynamics is constructed. In the digital twin environment, based on the real-time interactive rendering of pressure difference, eddy parameters and mechanical signals, the dynamic process of valve and blood flow coupling is intuitively presented, and the boundary conditions of the simulation model are reversely corrected by spatiotemporal data, and an adaptive control mechanism of "simulation-feedback-optimization" is established. This method breaks through the limitations of traditional static models, ensures the dynamic synchronization of virtual valve motion and actual physiological state, and provides high-fidelity three-dimensional visualization and real-time mechanical monitoring capabilities for clinical decision-making and pathological analysis.

[0041] In order to establish the cross-scale correlation mechanism between the dynamic deformation of the aortic valve and the blood flow characteristics, the dynamic evolution of the valve geometric deformation and the surface pressure distribution path was simultaneously analyzed, the initial area of ​​the vortex caused by the sudden change in pressure gradient during the opening and closing of the valve was accurately located, the diffusion intensity and movement trajectory parameters of the vortex in the time and space dimensions were quantified, and the disturbance law of abnormal valve movement on the hemodynamic environment was revealed. A blood flow energy dissipation evaluation model based on the pressure difference-vortex dual parameter coupling was constructed, which provided a dynamic quantitative basis for the reverse reconstruction simulation of blood flow in heart valve diseases and the optimization of the fluid mechanics performance of valve interventional devices.

[0042] In some embodiments, as described in step 102, based on the simulation model, real-time calculation of the pressure difference parameter and vortex parameter of the blood flow during the opening and closing of the aortic valve includes: 201. Extracting geometric deformation data of the aortic valve during opening and closing from the simulation model, and dividing a dynamic region synchronized with the opening and closing state of the valve in the three-dimensional geometric model based on the movement trajectory of the valve apex in the geometric deformation data; In step 201, the dynamic region refers to a local region where the vertex motion trajectory is highly synchronized during the valve opening and closing process. The geometric deformation data is the coordinate information of the position of each point on the valve surface recorded in the simulation model as it changes with time.

[0043] In the embodiment of the present application, first, the geometric deformation data of the aortic valve opening and closing process (such as the displacement of the vertex coordinates over time) is derived from the simulation model, and the movement trajectory of each valve vertex is extracted through mesh processing software (such as Paraview) (such as the displacement of vertex A by 5mm during the systolic period). Secondly, based on the spatial similarity of the movement trajectory (such as the area with consistent displacement direction and amplitude), a clustering algorithm (such as K-means) is used to divide the valve surface into multiple dynamic areas (such as the leaflet center area and edge area). Finally, according to the phase of the valve opening and closing state (such as the fully closed period and the half-open period), the dynamic area is bound to the timestamp to generate a time-space synchronized area division mapping table.

[0044] 202. In the dynamic region, according to the difference in movement trajectories of the valve apexes, calculate the pressure difference between adjacent valve apexes, and construct a valve surface pressure distribution path through the pressure difference; In step 202, the pressure distribution path is a spatial distribution characteristic line formed by the pressure difference on the valve surface. The pressure difference is a local pressure gradient caused by the asynchronous movement of adjacent vertices.

[0045] In the embodiment of the present application, first, traverse the adjacent valve vertex pairs in the dynamic area (such as the distance between vertices A and B is 0.1mm), and read the corresponding pressure data (such as the pressure of vertex A is 120Pa, and the pressure of vertex B is 100Pa). Secondly, calculate the absolute value of the pressure difference between adjacent vertices (such as 20Pa), and generate a continuous pressure distribution path on the valve surface through an interpolation algorithm (such as a gradient line extending from a high-pressure area to a low-pressure area). Finally, mark the extreme points of pressure difference on the path (such as a difference of more than 50Pa) as key nodes, and construct a network topology diagram of the pressure distribution path on the valve surface.

[0046] 203. Based on the sudden change position of the pressure difference in the pressure distribution path on the valve surface, mark the area where the blood flow direction suddenly changes as the vortex initial area, and count the geometric center of the vortex initial area; In step 203, the eddy initial region is the starting position where the blood flow direction suddenly changes and causes rotation. The geometric center is the centroid coordinate of the eddy region.

[0047] In the embodiment of the present application, first, the sudden change position of the pressure difference is identified in the pressure distribution path (such as a sudden increase of more than 50Pa in the difference between adjacent nodes), and the boundary of the sudden change area is expanded by the morphological expansion algorithm, and marked as the vortex initial area (such as a circular area with a diameter of 2mm). Secondly, the center of mass of the vertex coordinates in each vortex initial area is calculated (such as taking the average value of all vertex coordinates in the area) to obtain the geometric center position (such as coordinates x=10mm, y=5mm). Finally, the geometric center is bound to the area number and timestamp to generate a spatial distribution list of the vortex initial area.

[0048] 204. Match the geometric center of the vortex initial area to the corresponding time point according to the time series of the valve opening and closing state, calculate the displacement difference value of the geometric center between adjacent time points to determine the movement distance corresponding to each time point, and calculate the diffusion intensity based on the coverage change of the vortex initial area at different time points to generate vortex diffusion parameters including the movement distance and the diffusion intensity; In step 204, the eddy center coordinates at each time point are arranged in chronological order, and the Euclidean distance between adjacent frames is calculated. At the same time, the rate of change of the eddy region area is measured. Finally, a data table containing timestamps, movement distances, and diffusion intensity is generated.

[0049] In the embodiment of the present application, first, the geometric center of the vortex initial area is matched according to the time series of the valve opening and closing state (such as one frame every 0.1 seconds), and the center displacement difference between adjacent time points (such as t=0.1s and t=0.2s) is calculated (such as moving from x=10mm to x=12mm, a displacement of 2mm). Secondly, the coverage area changes of the same vortex initial area at different time points are counted (such as expanding from 2mm² to 5mm²), and the diffusion intensity is calculated (such as an area change rate of 150%). Finally, the displacement difference value is bound to the diffusion intensity to generate an vortex diffusion parameter table containing the moving distance (2mm) and the diffusion intensity (150%).

[0050] 205. Taking the weighted superposition result of the intensity change of the pressure distribution path on the valve surface and the moving distance and diffusion intensity in the eddy current diffusion parameter as a pressure difference parameter; In step 205, weighted superposition is a calculation method that proportionally combines the pressure path strength and the eddy diffusion parameter. The pressure difference parameter is a composite index that integrates the pressure gradient and the eddy influence.

[0051] In the embodiment of the present application, first, the intensity change value of the pressure distribution path on the valve surface is extracted (such as the pressure difference on the path increases from 50Pa to 80Pa, and the intensity change rate is 60%). Secondly, the moving distance (2mm) and the diffusion intensity (150%) in the eddy diffusion parameter are weighted superimposed according to the preset weights (such as the moving distance weight 0.6, the diffusion intensity weight 0.4) (such as 2×0.6+ 150%×0.4=1.2+0.6=1.8). Finally, the superposition result is multiplied by the pressure path intensity change value (such as 1.8×60%=1.08) to generate a comprehensive pressure difference parameter (1.08) for quantifying the impact effect of abnormal blood flow on the valve.

[0052] 206. The ratio of the moving distance to the diffusion intensity is used as an eddy current parameter. In step 206, the eddy current parameter is a comprehensive index for quantifying the eddy current motion characteristics. The ratio of the moving distance to the diffusion intensity reflects the balance between the eddy current moving speed and the expansion speed.

[0053] In the embodiment of the present application, first, the moving distance (2mm) and the diffusion intensity (150%) are extracted from the eddy diffusion parameters. Secondly, the ratio of the two (2mm / 150%≈1.33mm / %) is calculated as the eddy parameters. Finally, the eddy parameters and the pressure difference parameters (1.08) are input into the health assessment model to determine the risk level of valve lesions (e.g., the higher the ratio, the more unstable the eddy current).

[0054] Here is a specific example: In the scenario of computer-assisted evaluation of the aortic valve of patients with cardiovascular disease, a hospital conducted a preoperative hemodynamic simulation on a patient with severe aortic stenosis. In the three-dimensional valve model constructed based on the patient's CT image, the system extracted the dynamic deformation data of the three leaflet vertices expanding from the closed position (spacing 2.8mm) to 18.3mm when the valve opened during systole, and divided the dynamic pressure response area centered on the valve ring. The simulation showed that in the middle of valve opening (0.25 seconds), the pressure difference between the adjacent vertices at the junction of the right coronary leaflet and the non-coronary leaflet reached 32mmHg, forming a high-pressure gradient zone extending from the valve ring to the aortic sinus. Its edge area was marked as the vortex initial area due to the sudden drop in pressure of 18mmHg, and the geometric center was located at the coordinates of the aortic sinus (X35, Y62, Z18). As the valve entered the maximum opening state at 0.35 seconds, the vortex center moved 9.7mm toward the ascending aorta, and the coverage area spread from the initial 15mm² to 42mm², and the diffusion intensity coefficient was calculated to be 0.83. The system superimposed the change in the intensity of the pressure gradient band (32→18mmHg) and the eddy current displacement distance with a weight of 0.6:0.4 to generate a pressure difference parameter of 68 (normal value <30) to characterize the abnormal valve function. At the same time, the presence of pathological eddy currents was determined based on the ratio of 9.7mm displacement to 42mm² diffusion range of 0.23 (threshold 0.15). The quantitative results were highly consistent with the abnormal blood flow acceleration area caused by the calcification of the valve leaflets found by transesophageal ultrasound during the operation, and ultimately guided the surgical team to use a 25mm artificial mechanical valve for replacement. The postoperative simulation parameters showed that the pressure difference parameter dropped to 22 and the eddy current ratio returned to 0.09, effectively avoiding the risk of paravalvular flow complications that may be missed by traditional evaluation methods.

[0055] In summary, steps 201 to 206 implement a precise quantification method for blood flow parameters based on dynamic tracking of valve geometric deformation. The dynamic area is divided by extracting the difference in the movement trajectory of the valve apex, and the pressure distribution path is correlated with the geometric characteristics of the initial area of ​​the vortex for analysis, and an innovative vortex diffusion parameter calculation model based on the number of covered grids and spatial displacement is proposed. Through the weighted superposition of pressure difference parameters and vortex parameters, a composite index reflecting the spatiotemporal evolution of blood flow disturbances is constructed, which solves the problem of insufficient capture of complex vortex dynamics by traditional single parameter methods. This technology significantly improves the ability to quantify the blood flow disorder phenomenon during the opening and closing of the valve, and provides a new dimension of data support for the early identification and evaluation of hemodynamic abnormalities.

[0056] In some embodiments, as described in step 204, the geometric center of the vortex initial area is matched to the corresponding time point according to the time series of the valve opening and closing state, the displacement difference value of the geometric center between adjacent time points is calculated to determine the movement distance corresponding to each time point, and the diffusion intensity is calculated based on the coverage change of the vortex initial area at different time points to generate the vortex diffusion parameters including the movement distance and the diffusion intensity, including: 301. Extracting continuous time points from the time series of the valve opening and closing states, and marking the geometric center position of the vortex initial area at each time point as the corresponding spatial coordinate; In step 301, continuous time points refer to equally spaced sampling moments intercepted from the valve motion cycle. The vortex initial area refers to the first rotating structure area formed when the fluid leaves the valve edge. The geometric center position refers to the coordinates of the spatial symmetric center point of the area calculated by mathematical methods.

[0057] In the embodiment of the present application, first, continuous time points (for example, one time point per second) are extracted from the time series of the valve opening and closing state, and the geometric center position data of the vortex initial area in each time point is read. The three-dimensional coordinates (X, Y, Z) of the geometric center are extracted through data processing software (such as Paraview), and marked as the spatial coordinates of the corresponding time point (for example, the time point t=1 second corresponds to the coordinates [10,5,2]). Then, the time points and coordinates are stored in a time series database (such as the time series database InfluxDB) in chronological order to form a "time-coordinate" mapping table to provide input for subsequent displacement calculations.

[0058] 302. Arrange the spatial coordinates of adjacent time points in chronological order, calculate the straight-line distance difference between the spatial coordinates of the adjacent time points in three-dimensional space, and generate the displacement difference value of the geometric center between the adjacent time points; In step 302, the displacement difference value refers to the linear movement of the eddy current center in the three-dimensional space at adjacent moments. Spatial coordinate arrangement refers to organizing the time series data into an ordered coordinate set according to the acquisition order.

[0059] In an embodiment of the present application, first, the spatial coordinates of adjacent time points are taken out from the "time-coordinate" mapping table in chronological order (such as the coordinates [10,5,2] at t=1 second and the coordinates [12,6,3] at t=2 seconds). Secondly, the straight-line distance difference between the two is calculated using the three-dimensional space distance formula (for example, √[(12-10)²+(6-5)²+(3-2)²]=√6≈2.45 mm, √ represents the square root). Then, the straight-line distance difference is stored in a displacement difference value list, and the corresponding adjacent time points are associated (such as the displacement difference from t=1 to t=2 seconds is 2.45 mm), and a displacement difference value sequence is generated to quantify the instantaneous speed of the vortex center movement.

[0060] 303. Based on the number of covered grids in the eddy initial region at different time points, a ratio of the number of covered grids between adjacent time points is calculated as a diffusion rate, wherein the number of covered grids is obtained by counting the total number of boundary grid units in the eddy initial region; In step 303, the number of covered grids refers to the total number of grids occupied by the eddy region after the three-dimensional space is divided into unit cubic grids. The diffusion rate refers to the spatial expansion ratio of the eddy region per unit time.

[0061] In the embodiment of the present application, first, based on the grid division of the three-dimensional geometric model (such as dividing the initial eddy current area into a cubic grid with a side length of 1 mm), the total number of grid units covered by the initial eddy current area at each time point is counted (for example, 50 grids are covered at t = 1 second). Secondly, the ratio of the number of covered grids between adjacent time points is calculated (such as 75 grids are covered at t = 2 seconds, and the ratio is 75 / 50 = 1.5). Finally, the ratio is used as the diffusion rate (such as a diffusion rate of 1.5 means that the coverage area is expanded by 50%), and a diffusion rate sequence is generated to reflect the speed of diffusion of the eddy current area over time.

[0062] 304. Superimposing the displacement difference value and the diffusion rate in time series respectively to generate a movement distance and diffusion intensity corresponding to each time point; In step 304, the moving distance refers to the length of the spatial trajectory of the eddy center accumulated over time. The diffusion intensity refers to the cumulative amount of the kinetic energy of the eddy region expansion.

[0063] In the embodiment of the present application, first, the displacement difference value sequence is accumulated (for example, the displacement is 2.45 mm from t=1 to t=2 seconds, and the displacement is 3 mm from t=2 to t=3 seconds, then the moving distance at t=3 seconds is 2.45+3=5.45 mm), and the cumulative moving distance corresponding to each time point is generated. Secondly, the diffusion rate sequence is weighted averaged (such as taking the average of the diffusion rates of the three most recent time points) to generate the diffusion intensity at each time point (for example, the average of 1.5 corresponds to the diffusion intensity of "medium"). Finally, the moving distance and the diffusion intensity are bound to each time point to form a "time-moving distance-diffusion intensity" parameter table.

[0064] 305. The result of the multiplication operation between the moving distance and the diffusion intensity is used as an eddy current diffusion parameter including the moving distance and the diffusion intensity.

[0065] In step 305, the eddy diffusion parameter refers to a composite dynamic index that integrates the motion trajectory and the expansion capability.

[0066] In the embodiment of the present application, first, the moving distance (such as 5.45 mm) and the diffusion intensity (such as 1.5) at each time point are extracted from the parameter table. Secondly, the two are multiplied according to the product relationship (for example, 5.45×1.5≈8.175) to generate the eddy current diffusion parameter (such as parameter value 8.175) containing the moving distance and the diffusion intensity. Finally, the parameter value is mapped to the eddy current diffusion level (such as 0-5 for low risk and 5-10 for medium risk), and the time series is associated and output as an eddy current diffusion risk assessment report for clinical or engineering analysis.

[0067] Here is a specific example: In the postoperative evaluation scenario of patients undergoing aortic valve replacement, a hospital conducted a CT image hemodynamic simulation analysis on a patient with a bioprosthetic valve implanted. Based on the dynamic valve opening and closing data one year after surgery, the system captured the initial vortex area formed at the junction of the left coronary valve and the right coronary valve when the valve leaflets closed at 0.12 seconds in the diastolic period, and its geometric center was located at the three-dimensional coordinates (X28, Y45, Z12). As the time series advanced to 0.18 seconds, the center of the vortex moved distally along the ascending aorta wall to (X31, Y49, Z15), and the linear displacement between adjacent time points reached 4.3 mm. At the same time, the coverage grid expanded from 56 to 89, and the diffusion rate was calculated to be 1.59. At the 0.24 second time point, the vortex center continued to displace 3.8 mm to (X34, Y52, Z17), and the coverage grid increased to 123, corresponding to a diffusion rate of 1.38. The system fuses the cumulative displacement of 8.1 mm during 0.12-0.24 seconds with the diffusion intensity coefficient (1.59×1.38=2.19) to generate an eddy diffusion parameter of 17.7 (8.1×2.19), which is significantly higher than the normal threshold of 5.0. The quantitative result is highly consistent with the paravalvular micro-regurgitation area detected by ultrasound Doppler, revealing the abnormal hemodynamic characteristics caused by local tissue hyperplasia at the suture edge of the biological valve. Based on this, the clinical team formulated a targeted follow-up plan, and found that the parameter dropped to 6.3 during the review 18 months after the operation, which verified the therapeutic effect of adaptive remodeling of hyperplastic tissue and avoided the risk of premature secondary surgical intervention.

[0068] In summary, steps 301 to 305 achieve the ability to accurately model the spatiotemporal evolution of three-dimensional eddy diffusion characteristics. The vector coupling parameters of eddy diffusion intensity and movement distance are established by multiplying the displacement difference value of the geometric center coordinates at consecutive time points with the diffusion rate of the covering grid. This method fully describes the position migration law and scale expansion trend of the eddy core area through dynamic tracking of the three-dimensional spatial coordinate sequence and statistical analysis of the grid coverage range, breaking through the limitations of traditional fixed observation points in describing the dynamic characteristics of the eddy. This technology enhances the modeling accuracy of the eddy evolution process in complex blood flow scenarios, and provides a quantitative basis with high spatiotemporal resolution for the evaluation of blood flow energy loss caused by valvular lesions.

[0069] In some embodiments, as described in step 105, obtaining the spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient includes: 401. Extracting the spatiotemporal relationship data of blood flow velocity and pressure gradient from the visualization interface, decomposing the hemodynamic characteristics in the spatiotemporal relationship data, and generating blood flow characteristic parameters reflecting the actual physiological state of the patient; In step 401, the spatiotemporal relationship data refers to a three-dimensional dynamic data set including blood flow velocity distribution and pressure gradient changes. The hemodynamic characteristics refer to key indicator parameters reflecting the heart's pumping function extracted from the spatiotemporal data. In the embodiment of the present application, first, the spatiotemporal relationship data of blood flow velocity and pressure gradient (such as velocity vector diagram and pressure cloud diagram) are extracted from the data interface of the visualization interface, and the discrete data is converted into a continuous time series through a spatiotemporal interpolation algorithm. Secondly, the periodicity, pulsation and other hemodynamic characteristics (such as systolic velocity peak and diastolic pressure valley) in the data are separated by feature decomposition technology (such as principal component analysis or wavelet transform). Finally, the decomposed features are normalized in combination with clinical standard parameters (such as normal cardiac output range) to generate blood flow characteristic parameters (such as abnormal velocity index and pressure gradient fluctuation rate) that reflect the actual physiological state of the patient, and stored as a structured parameter table.

[0070] 402. Compare the blood flow characteristic parameters with the boundary conditions of the current valve motion in the simulation model, identify the matching deviation between the blood flow characteristic parameters and the boundary conditions, and generate boundary condition correction requirements; In step 402, the boundary condition correction requirement refers to the parameter adjustment direction indicated by the difference between the simulation model calculation result and the actual measurement data. The matching deviation refers to the model prediction error existing in the spatial distribution and time evolution of the characteristic parameters. In the embodiment of the present application, first, the boundary conditions of the current valve movement (such as the leaflet opening and closing angle, closing speed) are read from the simulation model, and aligned with the blood flow characteristic parameters of step 401 according to the time window. Secondly, the differences between the two are compared item by item through residual calculation or dynamic time warping algorithm (DTW) (such as the actual flow rate peak is 15% lower than the simulation value), and the matching deviation is identified (such as insufficient flow rate due to delayed valve closure). Finally, according to the deviation type (amplitude deviation, timing deviation), the boundary condition correction requirements (such as "increase the leaflet opening and closing angle by 10%" or "advance the closing phase by 0.1 second") are generated to form a correction requirement list.

[0071] 403. Based on the boundary condition correction requirement, analyze the correlation between the valve motion trajectory and the blood flow characteristic parameters in the three-dimensional geometric model, and determine the influence weight of the valve motion on the blood flow characteristic parameters; In step 403, the influence weight refers to a quantitative index of the sensitivity of the change in the valve motion trajectory to the blood flow parameters. The correlation relationship refers to the nonlinear coupling mechanism between the leaflet displacement and the blood flow characteristic parameters. In the embodiment of the present application, first, the valve motion trajectory data (such as the hinge point displacement curve) is extracted from the three-dimensional geometric model and aligned with the blood flow characteristic parameters (such as the pressure gradient) in time and space. Secondly, the influence weight of the valve motion parameters (such as the opening and closing speed) on the blood flow characteristic parameters (such as the pressure gradient) is calculated through gray correlation analysis or multivariate regression model (such as the weight of the opening and closing speed on the pressure gradient is 0.6). Finally, the key correction items are determined according to the weight sorting (such as giving priority to adjusting the opening and closing angles of the leaflets with high weights), and an influence weight distribution table is generated to guide the priority of subsequent boundary condition adjustments.

[0072] 404. Adjust the boundary conditions of the simulation model according to the influence weights, and generate boundary condition corrections by constraining the coupling relationship between the amplitude of the valve opening and closing state and the blood flow velocity; In step 404, the boundary condition correction refers to the parameter value range that needs to be adjusted to achieve model calibration. The coupling relationship constraint refers to limiting the parameter adjustment range by establishing a physical correlation equation between the valve motion amplitude and the blood flow velocity.

[0073] In the embodiments of the present application, first, according to the influence weight distribution table, the boundary condition parameters of the simulation model (such as valve opening and closing amplitude, closing time) are gradient adjusted (such as the parameter with weight 0.6 is increased by adjustment). Secondly, the coupling relationship between the valve membrane movement and the blood flow velocity is balanced through a constrained optimization algorithm (such as particle swarm optimization) (such as the need to synchronously adjust the flow velocity upper limit after the opening and closing amplitude increases). Finally, a boundary condition correction amount (such as "opening and closing angle + 12%") is generated and encapsulated as a configuration file that can be recognized by the simulation model for model parameter updates.

[0074] 405. Inject the boundary condition correction amount into the simulation model to correct the simulation model, and verify the spatiotemporal consistency of the valve motion and blood flow characteristic parameters of the corrected simulation model to generate a simulation model that is synchronously mapped with the patient's physiological state.

[0075] In step 405, the spatiotemporal consistency verification refers to the consistency test between the modified simulation results and the actual data in the two dimensions of spatial distribution and time evolution. Synchronous mapping refers to the dynamic matching between the simulation model and the real-time physiological state of the patient.

[0076] In the embodiment of the present application, first, the boundary condition correction amount is injected into the simulation model (such as replacing the original configuration file), and the fluid dynamics solver is re-run to calculate the corrected valve motion and blood flow parameters. Secondly, the matching degree of the correction result with the patient's actual data is compared through a spatiotemporal consistency verification algorithm (such as dynamic time warping or root mean square error calculation) (such as the peak flow rate error is reduced from 15% to 5%). Finally, the spatiotemporal consistency report of the corrected model is output. If the verification passes, a simulation model that is synchronously mapped with the patient's physiological state is generated; if it fails, it is iterated and corrected until it meets the standard.

[0077] Here is a specific example: In the preoperative planning scenario of a patient with complex bicuspid aortic valve malformation, a cardiovascular center used a hemodynamic simulation system to optimize the surgical plan. The three-dimensional model constructed based on the patient's enhanced CT showed that there was an abnormal high-pressure area in the left coronary sinus region when the valve was opened during systole. The visualization interface extracted the characteristic parameters of peak blood flow velocity of 5.2m / s and transvalvular pressure difference of 68mmHg, but the simulation model only generated a flow velocity of 4.8m / s and a pressure difference of 58mmHg under the default boundary conditions. The system identified that there was a 10mmHg deviation between the actual blood flow characteristics and the model prediction. Tracing back to the source, it was found that the 85% leaflet opening amplitude set by the model did not match the 72% opening and closing degree measured by the patient's CT. Through correlation analysis, it was determined that the weight of the influence of the valve motion amplitude on the transvalvular pressure difference was 0.78. Based on this, the leaflet opening and closing parameters in the simulation boundary conditions were adjusted from the free motion mode to the restricted mode, constraining the opening amplitude to 70% and enhancing the radial stiffness of the valve ring. The corrected model generated a flow rate of 5.1m / s and a pressure difference of 65mmHg at a time point of 0.25 seconds, with the error from the actual value measured by ultrasound Doppler reduced to 3%, and simultaneously reproduced the turbulent core area with a diameter of 4mm behind the left coronary sinus. This precise model successfully predicted the abnormal paravalvular flow velocity that may occur after the implantation of a biological valve, guided the selection of a 27mm valve and adjusted the suture angle during surgery, and the postoperative review showed that the actual transvalvular pressure difference of 42mmHg was highly consistent with the simulation prediction of 45mmHg, which verified the effectiveness of the model correction strategy and provided quantitative decision support for the personalized treatment of complex valve deformities.

[0078] In summary, steps 401 to 405 implement a dynamic optimization mechanism for the simulation model driven by digital twin data. Through the hemodynamic spatiotemporal characteristics extracted from the visual interface, the matching deviation between the boundary conditions of the simulation model and the real physiological state is reversely analyzed, and a boundary condition correction model based on the weight of the valve motion influence is constructed. This technology enables the simulation model to adaptively adjust the deformation parameters of the fluid domain by constraining the coupling relationship between the valve opening and closing amplitude and the blood flow velocity, and forms a closed-loop optimization through spatiotemporal consistency verification. This method fundamentally solves the problem of model distortion caused by the lack of physiological data in traditional simulations, and significantly improves the biomechanical consistency between the virtual valve motion and the individualized blood flow characteristics of the patient.

[0079] In some embodiments, as described in step 103, embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor to generate a mechanical signal reflecting the opening and closing state of the valve includes: 501. Determine the node distribution positions of the optical fiber virtual sensor in the curvature change area and the stress concentration area of ​​the valve surface topological structure; In step 501, the curvature change area refers to the area where the geometric shape of the valve surface changes suddenly, such as the leaflet joint or the calcified protrusion. The stress concentration area refers to the part where the valve is subjected to the maximum mechanical stress under the impact of blood flow. The optical fiber virtual sensor refers to the bionic sensor network established by mathematical modeling, and its node distribution position refers to the deployment coordinates of the simulated sensor on the valve surface. In an embodiment of the present application, first, based on the finite element analysis results of the valve surface topology, the coordinate distribution of the curvature change area (such as the bend of the leaflet edge) and the stress concentration area (such as near the valve hinge point) is extracted. The surface curvature mutation points are identified by a geometric curvature calculation algorithm (such as Gaussian curvature), and the grid nodes in the stress concentration area are screened out in combination with the stress cloud map. Secondly, the candidate node positions are marked in the double high curvature and stress areas (such as the junction of the leaflet center and the hinge point) to ensure that the fiber optic virtual sensor covers the mechanically sensitive area. Finally, a list of node distribution positions is generated, and the local geometric coordinates of the valve three-dimensional model are associated to provide input for the deployment of pressure sensing nodes.

[0080] 502. Setting a plurality of pressure sensing nodes in the node distribution position, and establishing a deformation constraint condition of each pressure sensing node under the action of blood flow by binding each pressure sensing node to the local geometric coordinates of the valve surface topological structure; In step 502, the pressure sensing node refers to a computing unit in the virtual sensor network that is responsible for collecting mechanical signals. The deformation constraint condition refers to the physical correlation equation between the node displacement and the elastic deformation of the valve surface material. In the embodiment of the present application, first, the specific coordinates of the pressure sensing node are selected in the node distribution position of step one, and each node is bound to the local geometric coordinates of the valve surface topological structure through a grid mapping algorithm (such as node A is bound to the coordinates [x1, y1, z1]). Secondly, according to the elastic modulus of the valve material and the blood flow pressure data, the deformation constraint conditions of each node are set (such as the maximum displacement of the node under blood flow pressure does not exceed 2mm). Finally, the displacement boundary conditions are defined for each node through a finite element solver (such as the deformation freedom limit in the X / Y / Z direction), and a node constraint parameter table is generated to provide rules for displacement calculation.

[0081] 503. Calculate the displacement of each pressure sensing node according to the deformation constraint condition, and convert the displacement into a three-dimensional space coordinate change sequence of the pressure sensing node; In step 503, the displacement refers to the displacement of the three-dimensional spatial position of the pressure sensing node under the impact of blood flow. The coordinate change sequence refers to the trajectory data set of the node position evolving over time. In the embodiment of the present application, first, the blood flow pressure data is loaded into the finite element simulation model, and the displacement of each pressure sensing node under the action of blood flow is calculated based on the deformation constraint condition (such as the displacement of node A during the systolic period is 1.5 mm). Secondly, the displacement is converted into a three-dimensional spatial coordinate change sequence of the node (such as the coordinates [x1+0.3, y1+0.2, z1+0.1] at t=0.1 seconds) through a coordinate transformation algorithm (such as a rigid body transformation or an affine transformation). Finally, the coordinate change data of all nodes are stored in a time series to form a "time-coordinate" mapping table for deformation feature extraction.

[0082] 504. Extract displacement change data of each pressure sensing node according to the three-dimensional space coordinate change sequence, and construct a deformation characteristic curve of the node based on the displacement change data; In step 504, the deformation characteristic curve refers to a time-amplitude relationship graph reflecting the node displacement mode. The displacement change data refers to a multivariate time series including displacement, velocity, and acceleration. In the embodiment of the present application, first, the time series displacement change data of each pressure sensing node is extracted from the "time-coordinate" mapping table (such as the displacement of node A increases from 0 to 1.5 mm in 0-0.5 seconds). Secondly, the deformation characteristic curve of the node is constructed by cubic spline interpolation or polynomial fitting algorithm (such as the displacement-time curve shows a trend of first accelerating and then decelerating). Finally, the curve parameters (such as peak displacement, slope change point) are associated with the valve opening and closing state (such as the displacement peak corresponding to the fully closed period) to generate a deformation characteristic curve library with time tags.

[0083] 505. Convert the deformation characteristic curve into a mechanical signal waveform, characterize the mechanical signal waveform through the displacement amplitude, velocity change rate and acceleration change rate of the pressure sensing node, and generate a mechanical signal reflecting the opening and closing state of the valve.

[0084] In step 505, the mechanical signal waveform refers to the process of converting the deformation characteristics into a quantifiable biomechanical index curve. Feature extraction refers to the process of identifying key parameters from the waveform. In an embodiment of the present application, first, the displacement amplitude (such as 1.5mm), velocity change rate (such as the displacement from 0 to 1.5mm takes 0.3 seconds corresponding to a velocity of 5mm / s) and acceleration change rate (such as the velocity from 0 to 5mm / s takes 0.1 seconds corresponding to an acceleration of 50mm / s²) in the deformation characteristic curve are used as input, and a mechanical signal waveform is generated through a signal conversion model (such as a mechanical transfer function) (such as the displacement amplitude is mapped to the voltage signal amplitude, and the acceleration is mapped to frequency modulation). Secondly, wavelet transform or Fourier analysis is used to extract the key features of the waveform (such as a high-frequency pulse signal corresponding to the closure period). Finally, the characteristic parameters are bound to the valve motion state, and a mechanical signal reflecting the opening and closing state is output (such as a sudden increase in the signal amplitude when fully closed) for real-time mapping and visualization of the digital twin system.

[0085] In summary, steps 501 to 505 realize the distributed mechanical monitoring virtualization technology based on valve topological deformation. Through the intelligent distribution strategy of nodes in the curvature change and stress concentration area, combined with the local geometric coordinate binding algorithm, a spatial position optimization model of the optical fiber virtual sensor is constructed. Based on the displacement conversion and characteristic curve dynamic mapping technology of deformation constraints, the three-dimensional spatial coordinate sequence is converted into a mechanical signal waveform containing amplitude, velocity and acceleration changes, which can accurately reflect the mechanical conduction path of micro-deformation of the valve surface. This technology realizes distributed mechanical feedback equivalent to real sensor measurement in a virtual environment for the first time, providing a high-precision virtual detection method for the mechanical tracing of valve opening and closing dysfunction.

[0086] Here is a specific example: In the preoperative planning of mitral valve repair for patients with rheumatic heart disease, a cardiovascular center built a dynamic valve model based on the patient's MRI data. In view of the curvature mutation area (curvature radius drops sharply from 8.2mm to 3.5mm) and stress concentration area (peak stress reaches 0.85MPa) found in the middle section of the posterior leaflet, the system arranged 12 groups of fiber optic virtual sensor nodes on the surface of the three-dimensional model, among which the key node P07 is located at the connection between the posterior leaflet and the tendon (X45, Y28, Z15). By binding the geometric coordinates of this area, a deformation constraint is established: when the valve is closed, the Z-axis displacement is limited to no more than 2.3mm. During the simulation of the cardiac systole, the P07 node detected a composite deformation of 1.8mm positive displacement on the X-axis and 2.1mm compressive displacement on the Z-axis at 0.25 seconds, and its displacement speed surged from 4mm / s to 12mm / s between 0.18-0.22 seconds, forming a characteristic double-peak waveform. After the system converted the deformation data into mechanical signals, it showed an abnormal impact waveform with an amplitude of 5.6N and an acceleration of 15m / s² at 0.28 seconds, which was highly consistent with the spatial position of the chordae calcification points found by ultrasound angiography (X44, Y27, Z16). The clinical team adjusted the surgical plan accordingly, adding calcification resection and artificial chordae tendineae implantation on the basis of retaining the native valve. The postoperative simulation showed that the amplitude of the impact waveform of the P07 node dropped to 2.1N, verifying the optimization effect of the repair strategy on the valve stress distribution.

[0087] In some embodiments, as described in step 101, constructing a three-dimensional geometric model of the aortic valve based on the CT image data, combining the three-dimensional geometric model with the fluid mechanics control equations, and establishing a simulation model including valve motion conditions includes: 601. Extracting contour information of the aortic valve from the aortic CT image data, and constructing a preliminary three-dimensional geometric structure of the valve based on the contour information, wherein the structure includes morphological features and spatial position of the valve; In step 601, the contour information refers to a pixel set of the valve edge segmented from the CT image. The preliminary three-dimensional geometric structure refers to a triangular mesh model containing the basic shape of the valve. In the embodiment of the present application, firstly, the aortic CT image data is threshold segmented by medical image processing software (such as Mimics or 3D Slicer), the two-dimensional contour information of the aortic valve is extracted, and the noise interference is removed by morphological optimization (such as corrosion expansion algorithm). Secondly, the continuous two-dimensional contour layers are stacked, and the Marching Cubes algorithm is used to generate a preliminary three-dimensional geometric structure, retaining the morphological characteristics of the leaflets (such as leaflet thickness, curvature) and spatial position (such as the connection relationship with the aortic root). Finally, the model is aligned with the original CT coordinate system through spatial registration technology to ensure the accuracy of the anatomical position of the three-dimensional structure.

[0088] 602. Refine the preliminary three-dimensional geometric structure, supplement the detailed features of the valve, and form a complete three-dimensional geometric model of the aortic valve; In step 602, the detail features refer to the micro anatomical structures on the valve surface, such as calcified nodules, fibrotic stripes, etc. The complete three-dimensional geometric model refers to a refined digital model including physiological surface features. In the embodiment of the present application, first, the preliminary three-dimensional geometric structure is meshed and refined (e.g., based on the Laplacian smoothing algorithm) to repair contour jumps or hole defects (e.g., the jagged structure of the leaflet edge). Secondly, the contact surface details when the valve is closed (e.g., the microstructure of the leaflet coaptation area) are supplemented by manual or semi-automatic tools (e.g., the "hole filling" function of MeshLab), and key anatomical landmarks such as hinge points and valve rings are added. Finally, the geometric rationality of the model is verified in combination with clinical data (e.g., the normal valve opening and closing angle range), and a complete three-dimensional geometric model is output.

[0089] 603. Extracting geometric characteristic parameters of the valve surface based on the three-dimensional geometric model, and defining motion constraints during the valve opening and closing process in combination with the physiological motion law of the valve in the three-dimensional geometric model; In step 603, the geometric characteristic parameters include morphological indicators such as leaflet thickness, radius of curvature, surface area, etc. The motion constraint condition refers to the physical motion rules restricted by the anatomical structure during the opening and closing of the valve. In the embodiment of the present application, first, the geometric characteristic parameters such as the curvature, thickness, and area of ​​the valve surface (such as the leaflet center curvature radius of 2 mm) are extracted from the three-dimensional geometric model, and the physiological motion law of the valve (such as the diastolic leaflet closing speed) is simulated based on finite element analysis (such as ABAQUS). Secondly, the constraint conditions are defined according to the motion law: the valve opening and closing angle limit (such as the maximum opening angle of 85°), the closing phase delay time (such as 0.1 seconds) and other parameters are set through motion capture data or clinical observation. Finally, a structured constraint parameter table containing geometric features and motion rules is generated.

[0090] 604. Combining the three-dimensional geometric model with the fluid mechanics control equations, and constructing a dynamic model of the interaction between blood flow and valve according to the valve motion constraint conditions, wherein the dynamic model includes the pressure effect of blood flow on the valve and the response of the valve to the blood flow; In step 604, the fluid dynamics governing equations include the Navier-Stokes equations and the continuity equation. The dynamics model refers to a mathematical expression system of fluid-solid interaction. In the embodiment of the present application, first, the three-dimensional geometric model is imported into the fluid mechanics simulation software (such as ANSYS or OpenFOAM), and the body grid is divided (such as the dynamic layer grid adapts to the valve motion). Secondly, the Navier-Stokes equation is combined with the valve motion constraint conditions to set the blood flow boundary parameters (such as inlet flow velocity, outlet pressure) and valve dynamic response parameters (such as elastic modulus, damping coefficient). Then, the fluid-solid coupling equation is solved by the finite volume method to calculate the pressure effect of blood flow on the valve (such as the pressure distribution on the leaflet surface) and the reaction of valve deformation to blood flow (such as flow field disturbance). Finally, a dynamic model containing bidirectional action parameters is generated.

[0091] 605. Bidirectionally couple the pressure effect parameters of the blood flow on the valve and the response parameters of the valve to the blood flow in the dynamic model, and construct a simulation model including valve motion conditions by combining the motion constraint conditions with the morphological characteristics of the three-dimensional geometric model. In step 605, bidirectional coupling refers to the real-time interactive calculation of fluid pressure and structural deformation. The simulation model refers to a digital twin system that can reproduce the dynamic behavior of the valve under physiological conditions. In the embodiment of the present application, first, blood pressure parameters (such as systolic peak pressure of 120Pa) and valve response parameters (such as leaflet displacement of 2mm) are extracted from the dynamic model, and a real-time feedback relationship between the two is established through a bidirectional coupling iterative algorithm (such as a strongly coupled FSI). Secondly, the grid deformation parameters (such as the maximum displacement threshold) are adjusted according to the motion constraints (such as an opening angle limit of 85°) to ensure that the valve movement conforms to physiological laws. Finally, the convergence of the model is verified through multi-physics field simulation (such as a residual less than 1e-5), and a simulation model containing dynamic opening and closing of the valve, blood pressure fluctuations, and flow field vortex characteristics is output to support pathological analysis and surgical planning.

[0092] Here is a specific example: In the preoperative planning of transcatheter valve replacement (TAVR) for a patient with severe aortic stenosis, a heart center conducted hemodynamic simulation based on the patient's enhanced CT imaging data. The system first extracted the contour information of the valve calcification area from the CT sequence and reconstructed the initial three-dimensional structure, showing that the valve leaflets were thickened and fused, and the opening area was only 0.8 cm². Through refinement to supplement the spatial distribution details of the calcification foci, a calcified nodule with a diameter of 6.3 mm can be seen at the root of the right coronary valve leaflet in the complete model, resulting in the maximum opening and closing angle of the valve leaflet being limited to 32° (normal>60°) when opening and closing. Combined with the law of cardiac cycle motion, the constraint condition that the Z-axis displacement of the calcified nodule area does not exceed 1.2 mm when the valve leaflet is closed during diastole is set. After coupling the model with computational fluid dynamics, the simulation shows that the systolic blood flow forms a jet of 8.5 m / s at the calcified nodule, and the local pressure gradient reaches 78 mmHg. When the bidirectional coupling parameters were iterated for the fifth time, the model successfully reproduced the abnormal vibration mode of the valve leaflet under the constraint of calcification: the right coronary valve leaflet had a 3.7mm hysteresis displacement at the time point of 0.22 seconds, which was 93% consistent with the asynchronous characteristics of the valve leaflet motion measured by echocardiography. Based on this model, the surgical team selected a 26mm balloon-expandable valve, and the simulation predicted that the valve orifice flow velocity dropped to 2.3m / s after implantation, which was only 0.2m / s different from the actual postoperative review data, accurately avoiding the risk of paravalvular leakage and verifying the clinical value of the whole process from imaging to simulation. In summary, steps 601 to 605 realize the aortic valve dynamic coupling modeling method based on CT images. Through contour information extraction and geometric refinement, a three-dimensional valve model with both anatomical details and kinematic characteristics was established. Combining the fluid mechanics control equations with the valve physiological motion constraints, a bidirectional coupling model of blood flow pressure and valve mechanical response was constructed, realizing the synchronous dynamic solution of valve deformation and hemodynamics. This method solves the problem of accuracy loss in simplified boundary conditions in traditional one-way fluid-solid interaction simulations, forming a high-fidelity simulation framework from static structure to dynamic blood flow interaction, laying a precise numerical experimental foundation for the hemodynamic study of valve pathological mechanisms.

[0093] In some embodiments, as described in step 604, combining the three-dimensional geometric model with the fluid mechanics control equations to construct a dynamic model of the interaction between blood flow and valve according to the valve motion constraint conditions includes: 701. Extracting the valve surface topological structure of the three-dimensional geometric model, obtaining the curvature distribution characteristics and node connection relationship in the valve surface topological structure, and generating a surface mesh for coupling fluid and structure; In step 701, the surface topology refers to the meshed mathematical expression of the valve surface geometry. The curvature distribution feature refers to the quantitative index of the curvature degree of each area of ​​the surface. The surface mesh refers to the hybrid unit mesh used for fluid-structure interaction analysis.

[0094] In the embodiment of the present application, first, the mesh data of the valve surface topological structure (such as vertex coordinates, facet connection relationship) is extracted from the three-dimensional geometric model, and the surface curvature distribution characteristics (such as Gaussian curvature and mean curvature) are obtained through the curvature calculation tool (such as the curvature analysis module of MeshLab). Secondly, the surface mesh of fluid and structure coupling is constructed based on the node connection relationship (such as the vertex sharing rule of the triangular facet) to ensure that the mesh nodes are aligned with the valve deformation area. Then, the jagged edges are eliminated through the mesh optimization algorithm (such as Laplacian smoothing), and the surface mesh adapted for fluid mechanics calculation is generated, and it is exported to a format that can be recognized by finite element analysis software (such as ANSYS) to provide input for boundary condition matching.

[0095] 702. Match the boundary conditions of the surface grid with the fluid mechanics control equation, analyze the relationship between the fluid pressure gradient pulsation characteristics in the fluid mechanics control equation and the deformation response of the valve surface topological structure, and generate a grid topology for coupling the fluid and the structure; In step 702, boundary condition matching refers to mathematically associating fluid pressure pulsation with structural deformation. Grid topology refers to the data transmission relationship network at the fluid-solid interface. In the embodiment of the present application, first, the surface mesh is imported into a fluid mechanics simulation environment (such as OpenFOAM), and the inlet, outlet and wall boundary conditions of the fluid domain are set (such as the valve surface is set as a fluid-solid coupling boundary). Secondly, the pressure gradient pulsation characteristics in the Navier-Stokes equation are analyzed (such as the high-pressure area during systole and the low-pressure area during diastole), and the deformation response of the valve surface mesh under fluid pressure is calculated by finite element analysis (such as a deformation of 0.1 mm for every 10 Pa increase in pressure). Then, the fluid pressure gradient is associated with the structural deformation data to generate a dynamic fluid-solid coupling mesh topology (such as the deformation mesh is updated in real time with the pressure) to ensure bidirectional synchronization of the data in the fluid domain and the structural domain.

[0096] 703. Extracting valve opening and closing phase and motion trajectory feature points in the valve motion constraint conditions, matching the valve opening and closing phase and motion trajectory feature points with the mesh topology of the fluid-structure coupling, and generating fluid domain deformation parameters and valve motion boundary conditions; In step 703, the motion trajectory feature points refer to the key coordinate points of the landmark positions during the valve opening and closing process. The fluid domain deformation parameter refers to the quantitative index of the deformation of the blood flow area with the movement of the valve. In the embodiment of the present application, first, the opening and closing phase (such as 0-0.3 seconds for systole and 0.3-0.6 seconds for diastole) and the motion trajectory feature points (such as the displacement path of the leaflet hinge point) are extracted from the valve motion constraint conditions. Secondly, the phase and trajectory points are mapped to the mesh topology of fluid and structure coupling: by aligning the time series (such as the high-pressure fluid boundary corresponding to the systole), the displacement constraints of the trajectory feature points are marked at the key nodes of the mesh (such as the leaflet edge) (such as the displacement in the X direction does not exceed 2mm). Then, the fluid domain deformation parameters (such as the mesh stretching ratio) and the valve motion boundary conditions (such as the leaflet opening and closing speed limit) are generated according to the displacement constraints to form a dynamically updated mesh deformation rule table.

[0097] 704. Input the fluid domain deformation parameter and the valve motion boundary condition into the fluid mechanics control equation, and simultaneously solve the fluid mechanics control equation and the valve kinematic constraint to generate the pressure effect of the blood flow on the valve and the response data of the valve to the blood flow; In step 704, the synchronous solution refers to the real-time joint calculation of the fluid equation and the structural equation. The response data refers to the bidirectional mechanical parameters of the fluid-structure coupling effect. In the embodiment of the present application, first, the deformation parameters of the fluid domain (such as the grid stretching ratio of 0.1) and the boundary conditions of the valve motion (such as the maximum opening angle of the leaflet of 85°) are input into the fluid mechanics control equation (such as the transient Navier-Stokes equation), and the fluid pressure and structural deformation are solved synchronously by a strong coupling solver (such as the FSI module). Secondly, the pressure effect of blood flow on the valve (such as the pressure distribution on the leaflet surface) and the reaction of valve deformation to the flow field (such as the flow velocity fluctuation caused by the change in the cross-sectional area of ​​the flow channel) are iteratively calculated in each time step. Finally, a two-way response data set containing the peak pressure effect, valve displacement and flow field vortex characteristics is output for model correction.

[0098] 705. Map the response data to the valve surface topology of the three-dimensional geometric model, correct the position offset of the valve surface topology according to the response data, and generate a dynamic model including the interaction between blood flow and valve by synchronous iterative updating of the position offset and pressure.

[0099] In step 705, the position offset refers to the displacement of the structural grid node relative to the initial position. Synchronous iterative update refers to the real-time feedback correction of fluid pressure and structural deformation. In the embodiment of the present application, first, the pressure effect and valve displacement in the response data are mapped to the surface topological nodes of the three-dimensional geometric model (such as a pressure of 120Pa at node A corresponds to a displacement of 1.2mm), and the node position offset is corrected by a coordinate transformation algorithm (such as an affine transformation). Secondly, the fluid domain mesh deformation is recalculated according to the corrected position (such as an increase of 0.5mm in the flow channel width), and the model convergence is verified by iterative updates (such as synchronization every 0.01 seconds) (such as a residual less than 1e-5). Finally, a dynamic model containing dynamic blood flow pressure, valve deformation, and real-time updates of the mesh topology is output to support pathological simulation and surgical plan evaluation.

[0100] Here is a specific example: In the interventional treatment planning of a patient with congenital bicuspid aortic valve malformation, the medical team constructed a three-dimensional model of the abnormal valve based on high-resolution CT data. For the dumbbell-shaped opening formed by the fusion of the left and right coronary valves, the system extracted the valve surface topology and found that the radius of curvature of the anterior commissure area dropped sharply from 5.2mm of the normal leaflet to 1.8mm, and a high-density coupling grid was laid out in this area. When the surface grid was matched with the fluid mechanics equation, the model captured that when the leaflet opened to the maximum angle of 32° during systole (about 60° for a normal tricuspid valve), the grid nodes in the anterior commissure area produced a local pressure gradient of 14kPa due to the sudden change in curvature, resulting in an abnormal backward displacement of 3.2mm in the area at 0.22 seconds. By matching the valve opening and closing phase data obtained by echocardiography, the model linked the characteristic points of the leaflet closure trajectory with the deformation parameters of the fluid domain in the early diastole (0.15 seconds), reproducing the eccentric regurgitant jet caused by poor leaflet coaptation, with a peak flow velocity of 4.8m / s. After three iterations of bidirectional coupling, the model showed that abnormal displacement caused a continuous vortex with a diameter of 6 mm in the left coronary sinus, which matched the actual regurgitation area detected by cardiac MRI four-dimensional blood flow imaging by 91%. Based on this, the surgical team adjusted the original intervention plan and selected a 26mm artificial valve with an external skirt. The postoperative simulation showed that the regurgitation velocity dropped to 1.2m / s, which was within the clinically acceptable range compared with the 1.3m / s measured by the actual catheter, successfully avoiding the risk of paravalvular leakage that may be missed by traditional imaging assessment.

[0101] In summary, steps 701 to 705 implement an efficient dynamic simulation algorithm for the fluid-structure coupling field. Through the curvature-driven fluid-structure grid topology matching technology, the valve opening and closing phase and motion trajectory feature points are embedded in the solution process of the fluid domain deformation parameters. The valve kinematic constraints are synchronously integrated in the fluid mechanics equations, and the real-time synchronous solution of bidirectional fluid-solid coupling is achieved through the iterative correction mechanism of pressure action and displacement response. This technology significantly improves the modeling accuracy of the dynamic response of valve deformation to blood flow pressure gradient, solves the problem of energy transfer error caused by timing mismatch in traditional step-by-step solution methods, and provides a highly reliable simulation platform for the prediction of hemodynamic effects of complex valvular lesions.

[0102] Figure 2 A structural schematic diagram of a real-time monitoring system for the physiological state of a virtual patient based on digital twins is provided for an embodiment of the present application, such as Figure 2 As shown, the system includes: A construction module 21 is used to obtain aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, and combine the three-dimensional geometric model with fluid mechanics control equations to establish a simulation model including valve motion conditions; A calculation module 22, for calculating the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve in real time based on the simulation model; A generating module 23, used for embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model, so as to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor, and generate a mechanical signal reflecting the opening and closing state of the valve; An input module 24, used to input the pressure difference, eddy current parameters and mechanical signals into a rendering engine in the digital twin environment, and generate a visualization interface by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; The correction module 25 is used to obtain the spatiotemporal relationship data between the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

[0103] Figure 2 The real-time monitoring system of virtual patient physiological status based on digital twin can perform Figure 1 The implementation principle and technical effect of the method for real-time monitoring of the physiological state of a virtual patient based on digital twins described in the illustrated embodiment will not be repeated. The specific manner in which each module and unit performs operations in the real-time monitoring system for the physiological state of a virtual patient based on digital twins in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0104] In one possible design, Figure 2 A real-time monitoring system for the physiological state of a virtual patient based on digital twins of the embodiment shown 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 called and executed by the processing component 32 .

[0105] The processing component 32 is used for the above Figure 1 The embodiment provides a real-time monitoring method for the physiological state of a virtual patient based on digital twins.

[0106] 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 method. Of course, the processing component may also be implemented by 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 method.

[0107] The 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 memory, flash memory, magnetic disk or optical disk.

[0108] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0109] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0110] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0111] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0112] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A method for real-time monitoring of the physiological state of a virtual patient based on digital twins is shown in the embodiment.

[0113] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0114] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0115] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A real-time monitoring method for the physiological state of a virtual patient based on digital twins, characterized in that: include: Acquire aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, combine the three-dimensional geometric model with fluid mechanics control equations, and establish a simulation model including valve motion conditions; Based on the simulation model, the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve are calculated in real time; Embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor to generate a mechanical signal reflecting the opening and closing state of the valve; The pressure difference, eddy current parameters and mechanical signals are input into a rendering engine in the digital twin environment, and a visualization interface is generated by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; The spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface is obtained to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

2. The method according to claim 1, characterized in that Based on the simulation model, the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve are calculated in real time, including: Extracting geometric deformation data of the aortic valve during opening and closing from the simulation model, and dividing a dynamic region synchronized with the opening and closing state of the valve in the three-dimensional geometric model based on the movement trajectory of the valve apex in the geometric deformation data; In the dynamic region, according to the difference in the movement trajectory of the valve apex, the pressure difference between adjacent valve apex is calculated, and the pressure distribution path of the valve surface is constructed through the pressure difference; Based on the sudden change position of the pressure difference in the pressure distribution path on the valve surface, marking the area where the blood flow direction suddenly changes as the vortex initial area, and counting the geometric center of the vortex initial area; Matching the geometric center of the vortex initial area to the corresponding time point according to the time series of the valve opening and closing state, calculating the displacement difference value of the geometric center between adjacent time points to determine the movement distance corresponding to each time point, and calculating the diffusion intensity based on the coverage change of the vortex initial area at different time points to generate vortex diffusion parameters including the movement distance and the diffusion intensity; The weighted superposition result between the intensity change of the pressure distribution path on the valve surface and the moving distance and diffusion intensity in the eddy current diffusion parameter is used as a pressure difference parameter; The ratio of the moving distance to the diffusion intensity is taken as the eddy current parameter.

3. The method according to claim 2, characterized in that The geometric center of the vortex initial area is matched to the corresponding time point according to the time series of the valve opening and closing state, and the displacement difference value of the geometric center between adjacent time points is calculated to determine the movement distance corresponding to each time point, and the diffusion intensity is calculated based on the coverage change of the vortex initial area at different time points to generate the vortex diffusion parameters including the movement distance and the diffusion intensity, including: Extracting continuous time points from the time series of the valve opening and closing states, and marking the geometric center position of the vortex initial area at each time point as the corresponding spatial coordinate; Arrange the spatial coordinates of adjacent time points in chronological order, calculate the straight-line distance difference of the spatial coordinates of the adjacent time points in three-dimensional space, and generate the displacement difference value of the geometric center between the adjacent time points; Based on the number of covered grids in the eddy initial region at different time points, the ratio of the number of covered grids between adjacent time points is calculated as the diffusion rate, wherein the number of covered grids is obtained by counting the total number of boundary grid units in the eddy initial region; The displacement difference value and the diffusion rate are respectively superimposed in time series to generate the movement distance and diffusion intensity corresponding to each time point; The result of the multiplication relationship between the moving distance and the diffusion intensity is used as the eddy current diffusion parameter including the moving distance and the diffusion intensity.

4. The method according to claim 1, characterized in that: Acquiring the spatiotemporal relationship data of the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient, including: Extracting the spatiotemporal relationship data of blood flow velocity and pressure gradient from the visualization interface, decomposing the hemodynamic characteristics in the spatiotemporal relationship data, and generating blood flow characteristic parameters reflecting the actual physiological state of the patient; Comparing the blood flow characteristic parameters with the boundary conditions of the current valve motion in the simulation model, identifying the matching deviation between the blood flow characteristic parameters and the boundary conditions, and generating boundary condition correction requirements; Based on the boundary condition correction requirement, the correlation between the valve motion trajectory and the blood flow characteristic parameters in the three-dimensional geometric model is analyzed to determine the influence weight of the valve motion on the blood flow characteristic parameters; Adjusting the boundary conditions of the simulation model according to the influence weights, and generating boundary condition corrections by constraining the coupling relationship between the amplitude of the valve opening and closing state and the blood flow velocity; The boundary condition correction amount is injected into the simulation model to correct the simulation model, and the temporal and spatial consistency of the valve motion and blood flow characteristic parameters of the corrected simulation model is verified to generate a simulation model that is synchronously mapped with the patient's physiological state.

5. The method according to claim 1, characterized in that Embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor to generate a mechanical signal reflecting the opening and closing state of the valve, including: Determining the node distribution positions of the optical fiber virtual sensor in the curvature change area and the stress concentration area of ​​the valve surface topological structure; A plurality of pressure sensing nodes are set in the node distribution position, and a deformation constraint condition of each pressure sensing node under the action of blood flow is established by binding each pressure sensing node with the local geometric coordinates of the valve surface topological structure; Calculating the displacement of each pressure sensing node by using the deformation constraint condition, and converting the displacement into a three-dimensional space coordinate change sequence of the pressure sensing node; Extracting displacement change data of each pressure sensing node according to the three-dimensional space coordinate change sequence, and constructing a deformation characteristic curve of the node based on the displacement change data; The deformation characteristic curve is converted into a mechanical signal waveform, and the mechanical signal waveform is characterized by the displacement amplitude, velocity change rate and acceleration change rate of the pressure sensing node to generate a mechanical signal reflecting the opening and closing state of the valve.

6. The method according to claim 1, characterized in that A three-dimensional geometric model of the aortic valve is constructed based on the CT image data, and the three-dimensional geometric model is combined with a fluid mechanics control equation to establish a simulation model including valve motion conditions, including: Extracting contour information of the aortic valve from the aortic CT image data, and constructing a preliminary three-dimensional geometric structure of the valve based on the contour information, wherein the structure includes morphological characteristics and spatial position of the valve; Refining the preliminary three-dimensional geometric structure, supplementing the detailed features of the valve, and forming a complete three-dimensional geometric model of the aortic valve; Based on the three-dimensional geometric model, the geometric characteristic parameters of the valve surface are extracted, and the motion constraint conditions during the opening and closing process of the valve are defined in combination with the physiological motion law of the valve in the three-dimensional geometric model; Combining the three-dimensional geometric model with the fluid mechanics control equations, and constructing a dynamic model of the interaction between blood flow and valve according to the valve motion constraint conditions, wherein the dynamic model includes the pressure effect of blood flow on the valve and the response of the valve to the blood flow; The pressure effect parameters of blood flow on the valve and the response parameters of the valve to blood flow in the dynamic model are bidirectionally coupled, and the motion constraint conditions are combined with the morphological characteristics of the three-dimensional geometric model to construct a simulation model including valve motion conditions.

7. The method according to claim 6, characterized in that The three-dimensional geometric model is combined with the fluid mechanics control equations, and a dynamic model of the interaction between blood flow and valve is constructed according to the valve motion constraint conditions, including: Extracting the valve surface topology of the three-dimensional geometric model, obtaining the curvature distribution characteristics and node connection relationship in the valve surface topology, and generating a surface mesh for coupling fluid and structure; Matching the boundary conditions of the surface grid with the fluid mechanics control equation, analyzing the relationship between the fluid pressure gradient pulsation characteristics in the fluid mechanics control equation and the deformation response of the valve surface topological structure, and generating a grid topology for coupling the fluid and the structure; Extracting the valve opening and closing phase and motion trajectory feature points in the valve motion constraint conditions, matching the valve opening and closing phase and motion trajectory feature points with the mesh topology of the fluid-structure coupling, and generating fluid domain deformation parameters and valve motion boundary conditions; Inputting the fluid domain deformation parameters and the valve motion boundary conditions into the fluid mechanics control equation, synchronously solving the fluid mechanics control equation and the valve kinematic constraints, and generating the pressure effect of the blood flow on the valve and the response data of the valve to the blood flow; The response data is mapped to the valve surface topology of the three-dimensional geometric model, the position offset of the valve surface topology is corrected according to the response data, and a dynamic model including the interaction between blood flow and valve is generated by synchronous iterative updating of the position offset and pressure action.

8. A real-time monitoring system for the physiological state of virtual patients based on digital twins, characterized in that: include: A construction module is used to obtain aortic CT image data of a virtual patient, construct a three-dimensional geometric model of the aortic valve based on the CT image data, and combine the three-dimensional geometric model with fluid mechanics control equations to establish a simulation model including valve motion conditions; A calculation module, used for calculating the pressure difference parameters and vortex parameters of blood flow during the opening and closing of the aortic valve in real time based on the simulation model; A generating module, used for embedding a fiber optic virtual sensor in the valve surface topological structure of the three-dimensional geometric model, so as to simulate the deformation of the pressure sensing node under the action of blood flow through the fiber optic virtual sensor, and generate a mechanical signal reflecting the opening and closing state of the valve; An input module, used to input the pressure difference, eddy current parameters and mechanical signals into a rendering engine in the digital twin environment, and generate a visualization interface by adjusting the optical refractive index distribution of blood flow particle trajectories and pressure gradient fields; The correction module is used to obtain the spatiotemporal relationship data between the blood flow velocity and the pressure gradient in the visualization interface to correct the boundary conditions of the simulation model so that the valve movement of the three-dimensional geometric model is synchronously mapped with the actual physiological state of the patient.

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 used to be called and executed by the processing component to implement a real-time monitoring method for the physiological state of a virtual patient based on digital twins as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a real-time monitoring method for the physiological state of a virtual patient based on digital twins as described in any one of claims 1 to 7 is implemented.

Images