Coronary artery fractional flow reserve noninvasive evaluation method fused with hemodynamics

Abstract

The invention relates to a coronary artery fractional flow reserve non-invasive evaluation method fused with hemodynamics, and belongs to the technical field of medical detection, and the method comprises the following steps: obtaining coronary artery CT image data, carrying out preprocessing, and extracting a coronary artery three-dimensional geometric model; constructing a neural network fused with hemodynamics, and adding a residual term of a Navier-Stokes equation and a residual term of a Murray law into a loss function; collecting a real training sample, and generating numerical simulation data as a training data set; s4, training and optimizing the neural network fused with hemodynamics to minimize an error between an output result of the network and an FFR training value, and satisfying constraint conditions of a Navier-Stokes equation and a Murray law at the same time; and S5, processing the coronary artery CT image data by using the trained network model fused with hemodynamics to obtain a predicted FFR value.

Description

Technical Field

[0001] The present invention belongs to the field of medical detection technology and relates to a non-invasive assessment method of coronary blood flow reserve fraction integrated with hemodynamics. Background Art

[0002] Coronary artery fractional flow reserve (FFR) is an important indicator for assessing the physiological function of coronary artery stenosis. It reflects the ratio of blood flow to the myocardial region supplied by the distal coronary artery during maximal hyperemia to the maximum blood flow that can be achieved in that region under normal circumstances. Traditional FFR measurement requires invasive coronary angiography, where a pressure guidewire is inserted into the coronary artery for measurement. This procedure is not only painful and risky for the patient, but also carries a high rate of complications and is costly.

[0003] To avoid the drawbacks of invasive measurement, non-invasive FFR assessment technology based on CT images (computed tomography derived fractional flow reserve, CT-FFR) has emerged. However, existing CT-FFR assessment methods mainly use computational fluid dynamics to numerically solve the Navier-Stokes equations. Due to the high computational complexity, they require a large amount of computing resources and time, making them difficult to be widely used in clinical practice. In recent years, the rapid development of deep learning has brought new opportunities for FFR assessment. Using deep learning technology to learn the training set generated by computational fluid dynamics and then automatically generate the FFR of the coronary arteries can effectively reduce computing time and manual operations. However, the training of deep learning models requires large-scale data support, and the model's interpretability is poor, mainly due to the lack of application of the hidden physical laws in the hemodynamic system.

[0004] Therefore, by combining the advantages of deep learning with the basic laws of hemodynamics, an interpretable deep learning algorithm is constructed to achieve non-invasive, rapid and accurate assessment of coronary artery FFR, thereby providing a reliable functional basis for early screening of coronary artery stenosis lesions, optimization of revascularization strategies and prognosis evaluation, which has significant clinical application value and socioeconomic benefits. Summary of the Invention

[0005] In view of this, the object of the present invention is to provide a non-invasive assessment method of coronary artery blood flow reserve fraction integrated with hemodynamics.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A non-invasive method for assessing coronary artery blood flow reserve fraction integrated with hemodynamics comprises the following steps:

[0008] S1: Acquire coronary artery CT image data and perform preprocessing to extract the three-dimensional geometric model of the coronary artery;

[0009] S2: Construct a neural network that integrates hemodynamics and adds the residual terms of the Navier-Stokes equations and Murray's law to the network's loss function;

[0010] S3: Collect coronary artery CT image data and corresponding invasive FFR measurements as training samples, and use computational fluid dynamics to generate CT-FFR numerical simulation data as a training dataset;

[0011] S4: using the training samples and training data set to train and optimize the neural network integrated with hemodynamics, so as to minimize the error between the output result of the network and the FFR training value, while satisfying the constraints of the Navier-Stokes equation and Murray's law;

[0012] S5: Use the trained hemodynamic fusion network model to process the coronary artery CT image data to obtain the predicted FFR value.

[0013] Furthermore, the step S1 specifically includes:

[0014] The patient's coronary arteries are scanned using a CT scanner to obtain a CT image sequence containing coronary artery morphology and structural information; the CT images are preprocessed, including using a Gaussian filtering algorithm to reduce noise; a histogram equalization algorithm is used to enhance image contrast and make the boundaries of the coronary arteries clear; and a method based on threshold segmentation and morphological operations is used to segment the coronary arteries and extract a three-dimensional geometric model of the coronary arteries.

[0015] Furthermore, the step S2 specifically includes:

[0016] Using multilayer perceptron as the basic network structure, a fully connected neural network consisting of an input layer, multiple hidden layers and an output layer is constructed;

[0017] The input layer receives geometric parameters and boundary conditions of the coronary arteries;

[0018] The hidden layer uses a plurality of neurons and a tanh activation function to perform nonlinear transformation;

[0019] The output layer outputs blood flow velocity and pressure in the coronary artery;

[0020] The residual terms of the Navier-Stokes equation and Murray's law are added to the network's loss function.

[0021] Furthermore, the loss function of the neural network model integrating hemodynamics is composed of a data-driven term and a physical consistency loss term.

[0022] The data-driven item Measures the difference between model output and observed data:

[0023]

[0024] Among them, u obs and p obs Represent the true values of measured flow rate and pressure respectively;

[0025] The physical consistency loss term ensures that the model output satisfies the basic hemodynamic equations, including the Navier-Stokes equations, the continuity equation, Murray's law, initial conditions, and boundary conditions;

[0026] The loss function of a neural network is expressed as the sum of a data-driven term and a physical consistency term:

[0027]

[0028] Among them, {λ pde ,λ Mu ,λ data ,λ BC ,λ IC} is the weight coefficient of the corresponding loss term, which is used to adjust the relative importance of the loss term.

[0029] Furthermore, the blood flow in the coronary arteries satisfies the three-dimensional Navier-Stokes equations and the continuity equation:

[0030]

[0031] Where t represents time, u and p represent blood flow velocity and blood pressure respectively, ρ and μ represent blood density and viscosity respectively;

[0032] Defining deep neural networks in neural networks integrating hemodynamics To approximate the velocity u and pressure p, where x = (x, y, z) is the spatial coordinate and t is the time; substitute the output of the neural network into the Navier-Stokes equation to obtain the loss term of the hemodynamic equation

[0033]

[0034] in, and They represent the speed and pressure of the neural network output respectively, and the symbol ‖·‖ represents the L2 norm.

[0035] Furthermore, the boundary conditions include inlet boundary conditions, outlet boundary conditions and wall boundary conditions; the inlet boundary conditions give a velocity distribution, the outlet boundary conditions use a pressure boundary condition or a flow boundary condition, and the wall boundary conditions use a no-slip condition u=0; the boundary condition loss is defined as:

[0036]

[0037] Among them, u BC and p BC are the known velocity and pressure at the boundary.

[0038] Furthermore, the initial condition loss is:

[0039]

[0040] Where u0 represents the initial flow velocity in the blood vessel.

[0041] Furthermore, Murray's law states that at a bifurcation, the cube of the diameter of the main vessel, d0, is equal to the sum of the cubes of the branch vessel diameters, d1 and d2:

[0042]

[0043] Traffic distribution also follows the following rules:

[0044] Q1 / Q2=(d1 / d2) 7 / 3

[0045] Therefore, the loss term associated with Murray's law at the coronary bifurcation is Expressed as:

[0046]

[0047] Where Q1 and Q2 are the flow rates of branch vessels, respectively, which are calculated using the velocity and diameter output by the neural network;

[0048] Assuming that the number of bifurcations in the coronary artery tree is N, the total loss term of Murray's law is expressed as the sum of the loss terms at each bifurcation location:

[0049]

[0050] Furthermore, in step S4, a gradient-based adaptive weight adjustment method is used to dynamically adjust the weight of the loss term, specifically including:

[0051] First, calculate each loss term separately Gradient of the neural network parameter θ

[0052] The weight of each loss term is then adjusted according to the size of the gradient norm:

[0053]

[0054] in Represents the gradient The L2 norm of

[0055] Finally, the Adam algorithm is used to minimize the loss function To optimize the parameters of the neural network model:

[0056]

[0057] Using the transfer learning method, the pre-training stage uses simulated data to train the neural network model that integrates hemodynamics and learns the physical laws of hemodynamics; in the fine-tuning stage, the model is fine-tuned on a small amount of clinical data to adapt it to the anatomical structure and noise distribution of real scenarios.

[0058] Furthermore, after training is completed, the sensitivity, specificity, and area under the curve are calculated based on the predicted FFR value and invasive blood flow reserve fraction value; the receiver operating characteristic curve is drawn to intuitively display the performance of the model; the sensitivity, specificity, and AUC indicators are analyzed to determine whether the model meets the usage requirements; false positive and false negative cases are analyzed to identify problems and deficiencies in the model; based on the analysis conclusions, the model is improved and optimized, including adjusting the model parameters and increasing training data; the model verification process is repeated until the model performance meets the target requirements.

[0059] The beneficial effects of the present invention are:

[0060] Non-invasiveness: This invention uses coronary artery CT image data to perform FFR assessment, avoiding the pain and risks of invasive measurement and improving patient acceptance.

[0061] Speed: The neural network integrated with hemodynamics has high computational efficiency and can complete CT-FFR assessment in a short time, meeting the needs of clinical practice.

[0062] Accuracy: Through the Navier-Stokes equation and Murray's law, the physical laws of hemodynamics in the coronary arteries and the blood flow distribution relationship at the vascular branches are fully considered, making the evaluation results more accurate and reliable.

[0063] Other advantages, objects, and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art upon examination of the following description or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings, in which:

[0065] Figure 1 Schematic diagram of the non-invasive assessment process of coronary artery FFR;

[0066] Figure 2 This is a neural network framework for non-invasive assessment of coronary FFR integrated with hemodynamics; (a) is the input variable of the neural network: the geometry of the coronary artery, (b) is the fully connected neural network, and (c) is the output variable of the neural network: flow velocity and pressure;

[0067] Figure 3 Schematic diagram of Murray's law at the location of coronary artery bifurcation. DETAILED DESCRIPTION

[0068] The following describes the embodiments of the present invention by means of specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0069] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0070] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0071] Example 1:

[0072] like Figure 1 As shown, the present invention provides a non-invasive assessment method of coronary artery blood flow reserve fraction integrated with hemodynamics, comprising the following steps:

[0073] (1) Obtain coronary artery CT image data: Scan the patient's coronary arteries using a CT scanner to obtain a CT image sequence containing coronary artery morphology and structural information. Preprocess the CT images, including image noise reduction, enhancement, and segmentation, to extract the three-dimensional geometric model of the coronary arteries.

[0074] (2) Constructing a neural network integrating hemodynamics: A multilayer perceptron is used as the basic network structure. The input layer receives the geometric parameters of the coronary arteries (such as vessel diameter and length) and boundary conditions (such as inlet pressure and outlet pressure). The hidden layer performs nonlinear transformations through multiple neurons. The output layer outputs physical quantities such as blood flow velocity and pressure in the coronary arteries. The residual terms of the Navier-Stokes equation and Murray's law are added to the network's loss function.

[0075] (3) Collect and generate training data sets: Collect coronary artery CT image data and corresponding invasive FFR measurement values as training samples, and use computational fluid dynamics to generate CT-FFR numerical simulation data as training data sets.

[0076] (4) Training the neural network integrated with hemodynamics: The network is trained using the stochastic gradient descent algorithm, combined with adaptive weight coefficients and transfer learning. By continuously adjusting the network parameters, the error between the network output and the FFR training value is minimized while satisfying the constraints of the Navier-Stokes equation and Murray's law.

[0077] (5) Non-invasive FFR assessment and verification: The coronary artery CT image data of the patient to be assessed is pre-processed and input into the trained neural network. The network outputs physical quantities such as blood flow velocity and pressure in the coronary artery. The FFR value is calculated based on these physical quantities and appropriately adjusted and optimized based on the model evaluation results.

[0078] Example 2:

[0079] This embodiment provides a detailed method for non-invasively assessing coronary artery flow reserve fraction integrated with hemodynamics, including the following steps:

[0080] (1) Data collection and preprocessing

[0081] Coronary CT angiography was performed using a multi-slice spiral CT scanner. Scanning parameters were adjusted based on the patient's specific condition and scanner performance. The scan range should cover the entire coronary artery tree, and the acquired image data were stored in DICOM format.

[0082] The CT image data was transferred to a computer and preprocessed using specialized image processing software. First, a Gaussian filter algorithm was used to reduce noise in the image. Then, a histogram equalization algorithm was used to enhance image contrast and clarify the boundaries of the coronary arteries. Finally, a threshold-based segmentation and morphological operation method was used to segment the coronary arteries and extract their 3D geometric model.

[0083] (2) Construction of a neural network integrating hemodynamics

[0084] A fully connected neural network was constructed, consisting of an input layer, multiple hidden layers, and an output layer. The input layer had 10 neurons, which received geometric parameters and boundary conditions such as coronary artery diameter, length, inlet pressure, and outlet pressure. Each hidden layer had 50 neurons, using the tanh activation function for nonlinear transformation. The output layer had two neurons, which output the blood flow velocity and pressure within the coronary arteries.

[0085] In hemodynamics, the Navier-Stokes equations describe the motion of fluids. Combining these with the geometrical information of coronary arteries allows for the construction of a PINN framework. Blood is assumed to be an incompressible laminar Newtonian fluid, and the vessel walls are simplified to linear elastic materials. Therefore, coronary blood flow satisfies the three-dimensional Navier-Stokes equations and the continuity equation:

[0086]

[0087] Where t represents time, u and p represent blood flow velocity and blood pressure, respectively, and ρ and μ represent blood density and viscosity, respectively.

[0088] like Figure 2 As shown in (a)-(c), in the neural network integrating hemodynamics, a deep neural network is defined To approximate the velocity u and pressure p, where x = (x, y, z) is the spatial coordinate and t is the time. Substituting the output of the neural network into the Navier-Stokes equation, we get the loss term of the hemodynamic equation

[0089]

[0090] in, and They represent the speed and pressure of the neural network output respectively, and the symbol ‖·‖ represents the L2 norm.

[0091] Boundary conditions include inlet boundary conditions, outlet boundary conditions, and wall boundary conditions. Inlet boundary conditions usually give a velocity distribution, outlet boundary conditions can use pressure boundary conditions or flow boundary conditions, and wall boundary conditions usually use a no-slip condition (u=0). Boundary condition loss can be defined as:

[0092]

[0093] Among them, u BC and p BC are the known velocity and pressure at the boundary.

[0094] Considering the initial conditions, such as the initial velocity distribution, the initial condition loss can be defined as:

[0095]

[0096] Where u0 represents the initial flow velocity in the blood vessel.

[0097] A Murray's law fusion mechanism is designed and embedded into the neural network model to improve the model's calculation accuracy for complex anatomical structures (such as bifurcations and side branches).

[0098] like Figure 3 As shown in Figure 2, Murray's law states that at a bifurcation, the cube of the main vessel diameter d0 is equal to the sum of the cubes of the branch vessel diameters d1 and d2:

[0099]

[0100] At the same time, traffic distribution also follows the following rules:

[0101] Q1 / Q2=(d1 / d2) 7 / 3

[0102] Therefore, the loss term associated with Murray's law at the coronary bifurcation is Expressed as:

[0103]

[0104] Where Q1 and Q2 are the flow rates of branch vessels, respectively, which can be calculated using the speed and diameter output by the neural network.

[0105] Assuming that the number of bifurcations in the coronary artery tree is N, the total loss term of Murray's law is expressed as the sum of the loss terms at each bifurcation location:

[0106]

[0107] The loss function of the neural network model integrating hemodynamics consists of a data-driven term and a physical consistency loss term. Measure the difference between model output and observed data (such as CFD blood flow velocity field and pressure field):

[0108]

[0109] Among them, u obs and p obs Represent the true values of measured flow rate and pressure respectively.

[0110] The physical consistency loss term ensures that the model output satisfies the basic hemodynamic equations, including the Navier-Stokes equations, the continuity equation, Murray's law, initial conditions, and boundary conditions. Therefore, the loss function of the neural network is expressed as the sum of the data-driven term and the physical consistency term:

[0111]

[0112] Among them, {λ pde ,λ Mu ,λ data ,λ BC ,λ IC} is the weight coefficient of the corresponding loss term, which is used to adjust the relative importance of the loss term.

[0113] (3) Collect and generate training datasets

[0114] Coronary artery CT image data and corresponding invasive FFR measurements were collected as training samples, and numerical simulation data of CT-FFR were generated by computational fluid dynamics as a training dataset.

[0115] (4) Training a neural network that integrates hemodynamics

[0116] Since different loss terms (such as boundary conditions, control equations, etc.) usually have different scales and importance, a gradient-based adaptive weight adjustment method is used to dynamically adjust the weights of these loss terms to improve the training effect and convergence speed of the model.

[0117] First, calculate each loss term separately Gradient with respect to the neural network parameters θ (including weights and biases) Then, the weight of each loss term is adjusted according to the size of the gradient norm:

[0118]

[0119] in Represents the gradient The L2 norm of .

[0120] Finally, the Adam algorithm is used to minimize the loss function To optimize the parameters of the neural network model:

[0121]

[0122] Transfer learning methods are used to enhance the generalization performance of the model.

[0123] Clinical data (paired CT-FFR data) is typically sparse and expensive to obtain, while simulated data (hemodynamic simulation based on computational fluid dynamics) can be generated on a large scale. The core concept of transfer learning is that, during the pre-training phase, rich simulated data is used to train a neural network model that integrates hemodynamics and learns the physical laws of hemodynamics. During the fine-tuning phase, the model is fine-tuned on a small amount of clinical data to adapt it to the anatomical structure and noise distribution of real-world scenarios. Through a physical data collaborative optimization strategy and transfer learning methods, high-precision FFR assessment can be achieved with sparse clinical data, while ensuring that the model output conforms to the physical laws of hemodynamics.

[0124] (5) Non-invasive FFR evaluation and verification

[0125] Based on the predicted FFR value and invasive flow reserve fraction value, calculate performance indicators such as sensitivity, specificity, and area under the curve. Draw the receiver operating characteristic curve to intuitively display the performance of the model. Analyze performance indicators such as sensitivity, specificity, and AUC to determine whether the model meets the requirements. Analyze false-positive and false-negative cases to identify problems and deficiencies in the model. Based on the conclusions of the result analysis, improve and optimize the model, including adjusting the model parameters and increasing training data. Repeat the model validation process until the model performance meets the target requirements. Use the trained fusion hemodynamic network model to predict the preprocessed image data to obtain the predicted FFR value for each patient.

[0126] In the above embodiments, references to "this embodiment" in the specification indicate that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment" do not necessarily refer to the same embodiment.

[0127] In the above embodiments, although the invention has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those skilled in the art based on the foregoing description. For example, other memory structures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed. The embodiments of the present invention are intended to encompass all such alternatives, modifications, and variations that fall within the broad scope of the appended claims.

[0128] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented.

[0129] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0130] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0131] Regarding the computer-readable storage medium in this embodiment, those skilled in the art will appreciate that all or part of the steps in the aforementioned method embodiments can be implemented using hardware associated with the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps in the aforementioned method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0132] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used for communication, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes the various steps of the above method.

[0133] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0134] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0135] The present invention can be used in a wide variety of general-purpose or special-purpose computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments that include any of the above.

[0136] The present invention may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions, which should all be included in the scope of the claims of the present invention.

Claims

1. A non-invasive method for assessing coronary artery flow reserve fractional volume integrated with hemodynamics, characterized by: The following steps are involved: S1: Acquire coronary artery CT image data and perform preprocessing to extract the three-dimensional geometric model of the coronary artery; S2: Construct a neural network that integrates hemodynamics and adds the residual terms of the Navier-Stokes equations and Murray's law to the network's loss function; S3: Collect coronary artery CT image data and corresponding invasive FFR measurements as training samples, and use computational fluid dynamics to generate CT-FFR numerical simulation data as a training dataset; S4: using the training samples and training data set to train and optimize the neural network integrated with hemodynamics, so as to minimize the error between the output result of the network and the FFR training value, while satisfying the constraints of the Navier-Stokes equation and Murray's law; S5: Use the trained hemodynamic fusion network model to process the coronary artery CT image data to obtain the predicted FFR value.

2. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 1, characterized in that: The step S1 specifically includes: The patient's coronary arteries are scanned using a CT scanner to obtain a CT image sequence containing coronary artery morphology and structural information; the CT images are preprocessed, including using a Gaussian filtering algorithm to reduce noise; a histogram equalization algorithm is used to enhance image contrast and make the boundaries of the coronary arteries clear; and a method based on threshold segmentation and morphological operations is used to segment the coronary arteries and extract a three-dimensional geometric model of the coronary arteries.

3. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 1, characterized in that: The step S2 specifically includes: Using multilayer perceptron as the basic network structure, a fully connected neural network consisting of an input layer, multiple hidden layers and an output layer is constructed; The input layer receives geometric parameters and boundary conditions of the coronary arteries; The hidden layer uses a plurality of neurons and a tanh activation function to perform nonlinear transformation; The output layer outputs blood flow velocity and pressure in the coronary artery; The residual terms of the Navier-Stokes equation and Murray's law are added to the loss function of the network.

4. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 3, characterized in that: The loss function of the neural network model integrating hemodynamics is composed of a data-driven term and a physical consistency loss term. The data-driven item Measures the difference between model output and observed data: Among them, u obs and p obs Represent the true values of measured flow rate and pressure respectively; The physical consistency loss term ensures that the model output satisfies the basic hemodynamic equations, including the Navier-Stokes equations, the continuity equation, Murray's law, initial conditions, and boundary conditions; The loss function of a neural network is expressed as the sum of a data-driven term and a physical consistency term: Among them, {λ pde ,λ Mu ,λ data ,λ BC ,λ IC } is the weight coefficient of the corresponding loss term, which is used to adjust the relative importance of the loss term.

5. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 4, characterized in that: The blood flow in the coronary arteries satisfies the three-dimensional Navier-Stokes equations and the continuity equation: Where t represents time, u and p represent blood flow velocity and blood pressure respectively, ρ and μ represent blood density and viscosity respectively; Defining deep neural networks in neural networks integrating hemodynamics To approximate the velocity u and pressure p, where x = (x, y, z) is the spatial coordinate and t is the time; substitute the output of the neural network into the Navier-Stokes equation to obtain the loss term of the hemodynamic equation in, and They represent the speed and pressure of the neural network output respectively, and the symbol ‖·‖ represents the L2 norm.

6. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 4, characterized in that: The boundary conditions include inlet boundary conditions, outlet boundary conditions and wall boundary conditions; the inlet boundary conditions give a velocity distribution, the outlet boundary conditions use pressure boundary conditions or flow boundary conditions, and the wall boundary conditions use a no-slip condition u=0; the boundary condition loss is defined as: Among them, u BC and p BC are the known velocity and pressure at the boundary.

7. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 4, characterized in that: The initial condition loss is: Where u0 represents the initial flow velocity in the blood vessel.

8. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 4, characterized in that: Murray's law states that at a bifurcation, the cube of the main vessel diameter d0 is equal to the sum of the cubes of the branch vessel diameters d1 and d2: Traffic distribution also follows the following rules: <h2 style=";text-align:left;direction:ltr">Q1 / Q2 = (d1 / d2)<h2 style=";text-align:left;direction:ltr"> 7 / 3 Therefore, the loss term associated with Murray's law at the coronary bifurcation is Expressed as: Where Q1 and Q2 are the flow rates of branch vessels, respectively, which are calculated using the velocity and diameter output by the neural network; Assuming that the number of bifurcations in the coronary artery tree is N, the total loss term of Murray's law is expressed as the sum of the loss terms at each bifurcation location:

9. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 1, characterized in that: In step S4, the weight of the loss term is dynamically adjusted using a gradient-based adaptive weight adjustment method, specifically including: First, calculate each loss term separately Gradient of the neural network parameter θ The weight of each loss term is then adjusted according to the size of the gradient norm: in Represents the gradient The L2 norm of Finally, the Adam algorithm is used to minimize the loss function To optimize the parameters of the neural network model: Using the transfer learning method, the pre-training stage uses simulated data to train the neural network model that integrates hemodynamics and learns the physical laws of hemodynamics; in the fine-tuning stage, the model is fine-tuned on a small amount of clinical data to adapt it to the anatomical structure and noise distribution of real scenarios.

10. The non-invasive assessment method of coronary artery fractional flow reserve integrated with hemodynamics according to claim 1, characterized in that: After training, the sensitivity, specificity, and area under the curve were calculated based on the predicted FFR and invasive flow reserve values; Draw the receiver operating characteristic curve to visually demonstrate the model's performance; analyze sensitivity, specificity, and AUC indicators to determine whether the model meets the requirements; analyze false positive and false negative cases to identify problems and deficiencies in the model; based on the analysis conclusions, improve and optimize the model, including adjusting model parameters and increasing training data; The model validation process is repeated until the model's performance meets the target requirements.

Images