Coronary artery plaque mechanical state evaluation method fused with physical and mechanical information

Abstract

The invention relates to a coronary artery plaque mechanical state evaluation method fusing physical and mechanical information, and belongs to the technical field of medical engineering, and the method comprises the following steps: S1, collecting medical image data and clinical information; s2, constructing a neural network fused with physical mechanics information, designing a loss item based on physical constraint by combining Hooke's law of elastic mechanics and an equilibrium equation, and combining the loss item into a loss function of the neural network; s3, solving the mechanical state of the coronary artery plaque by using a traditional numerical method, obtaining stress-strain related information, pairing the stress-strain related information with the corresponding medical image data and clinical information, and constructing a data set; s4, training, optimizing and testing the neural network fused with the physical and mechanical information by using a data set; s5, coronary artery medical image data to be evaluated and clinical information are input into the trained physical information neural network, and mechanical state evaluation results, including stress distribution and strain characteristics, of the plaques are output.

Description

Technical Field

[0001] The present invention belongs to the field of medical engineering technology and relates to a method for evaluating the mechanical state of coronary artery plaques by integrating physical and mechanical information. Background Art

[0002] Coronary plaque rupture is the main cause of acute cardiovascular events. Accurately assessing the mechanical state of coronary plaques is important for predicting plaque rupture risk and developing personalized treatment plans.

[0003] At present, traditional methods for assessing the mechanical state of coronary plaques are mainly based on numerical simulation techniques such as finite element analysis. Although numerical simulation methods can reflect the mechanical properties of plaques to a certain extent, they have many problems. On the one hand, the computational efficiency is low, and for complex coronary artery structures and large amounts of data processing, a lot of time and computing resources are required. On the other hand, there is a strong dependence on model parameters, and slight changes in parameters may lead to large deviations in the evaluation results. In addition, finite element analysis requires precise modeling of the geometric model of the coronary artery, which is often difficult to achieve in practical applications, especially for complex coronary artery structures and irregular plaque morphology. Accurate modeling is extremely difficult, thus limiting its widespread application in clinical practice.

[0004] In recent years, artificial intelligence technologies such as neural networks have been widely used in the biomedical field. However, most existing coronary plaque assessment methods only consider the geometric characteristics and image information of the plaque, ignoring the physical and mechanical information of the plaque's mechanical state. As a result, the accuracy and reliability of the assessment results need to be improved. Therefore, the development of a coronary plaque mechanical state assessment algorithm that can effectively integrate physical and mechanical information to improve assessment accuracy and efficiency is of great clinical significance. Summary of the Invention

[0005] In view of this, the object of the present invention is to provide a method for evaluating the mechanical state of coronary plaques by integrating physical and mechanical information.

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

[0007] A method for evaluating the mechanical state of coronary plaques by integrating physical and mechanical information comprises the following steps:

[0008] S1: Collect medical imaging data and clinical information;

[0009] S2: Build a neural network that integrates physical and mechanical information. Combine Hooke's law and equilibrium equations of elasticity to design a loss term based on physical constraints, and incorporate it into the loss function of the neural network.

[0010] S3: Use traditional numerical methods to solve the mechanical state of coronary artery plaques, obtain stress-strain related information, and pair it with corresponding medical imaging data and clinical information to construct a data set;

[0011] S4: using the data set to train, optimize and test the neural network integrating the physical and mechanical information;

[0012] S5: Input the coronary artery medical imaging data and clinical information to be evaluated into the trained physical information neural network, and output the mechanical state evaluation results of the plaque, including stress distribution and strain characteristics.

[0013] Furthermore, step S1 specifically includes:

[0014] Collect coronary artery CT angiography and magnetic resonance angiography from the image database; segment the coronary artery plaque, extract the geometric information of the plaque, perform normalization, and map its pixel value range to [0, 1];

[0015] The patients' clinical information, including age, gender, systolic blood pressure, diastolic blood pressure, total cholesterol, and triglycerides, was collected and standardized to have a mean of 0 and a standard deviation of 1.

[0016] Furthermore, the neural network integrating physical and mechanical information includes an input layer, a hidden layer and an output layer;

[0017] The input layer receives pre-processed medical imaging data and clinical information;

[0018] The hidden layer uses a fully connected neural network to perform feature extraction and nonlinear transformation on the input information;

[0019] The output layer uses a tanh activation function to output stress and strain information of coronary artery plaques;

[0020] The basic equations of continuum mechanics are used as loss functions or constraints, so that the neural network must not only minimize data errors but also satisfy the laws of physics during training.

[0021] Furthermore, the loss function of the neural network includes the data residual term Physical residual term Boundary condition residual term and the initial condition residual

[0022]

[0023] Among them, λ1, λ2, λ3, λ4 represent the weight coefficients of different loss terms;

[0024]

[0025] Among them, N d is the number of sampling points in the measurement data, u i and are the predicted and measured displacement values at sampling point i:

[0026]

[0027] Among them, N p is the number of sampling points in the plaque mechanics equation;

[0028]

[0029] Among them, N b is the number of boundary condition sampling points, u bnd is the reference value of the boundary condition;

[0030]

[0031] Among them, N i is the number of initial condition sampling points, u int is the reference value of the initial condition.

[0032] Furthermore, in step S3, the finite element method is used to numerically solve the physical model of the plaque. During the solution process, the dynamic characteristics of blood flow, the material properties of the blood vessel wall, and the geometric shape of the plaque are taken into account to calculate the stress and strain information around the coronary artery plaque. Assuming that the mechanical behavior of all vascular components of the coronary artery plaque conforms to that of linear elastic materials, the stress and strain information is expressed as follows:

[0033] σ ij =λδ ij ε kk +2Gε ij

[0034] Among them, σ ij and ε ij are the stress tensor and strain tensor, λ and G are the material parameters of the patch;

[0035] The equation of motion of the plaque is expressed as:

[0036]

[0037] Where t represents time, ρ s and u are the density and displacement of the plaque material, respectively, and b represents the body force.

[0038] Further, in step S4, the collected data set is divided into a training set, a validation set, and a test set;

[0039] Train the neural network using the training set, setting the learning rate, batch size, and number of training rounds;

[0040] The stochastic gradient descent algorithm is used to minimize the loss function to optimize the model parameters:

[0041]

[0042] Here, θ represents all the parameters of the neural network, including weights and biases;

[0043] During the training process, the validation set is used to evaluate and optimize the model. The network structure and learning rate are adjusted, and an early stopping strategy is used to avoid overfitting.

[0044] Use the test set to test the trained neural network.

[0045] Furthermore, the coronary artery medical imaging data and clinical information to be evaluated are input into a trained neural network that integrates physical and mechanical information, and the mechanical state assessment results of the plaque, including stress distribution and strain characteristics, are output; the sensitivity, specificity, and area under the curve (AUC) are calculated based on the predicted values and true measurement values, and the receiver operating characteristic curve is plotted; the performance indicators and false positive and false negative cases are analyzed to identify the deficiencies of the model and make improvements and optimizations until the performance meets the standards.

[0046] The present invention significantly improves the accuracy of coronary plaque mechanical state assessment by incorporating solid mechanics principles of coronary plaque mechanics into the neural network's loss function and utilizing physical laws to constrain network convergence, making the assessment results more consistent with physical reality. Compared to traditional numerical simulation methods, the neural network algorithm of the present invention offers faster computational speeds and can output assessment results in a shorter time, meeting the clinical need for rapid diagnosis.

[0047] 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

[0048] 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:

[0049] Figure 1 Flowchart of the coronary plaque mechanical status assessment method integrating physical and mechanical information;

[0050] Figure 2Figure 3 is a diagram of the neural network structure that integrates physical and mechanical information, where (a) is the neural network input parameter: coronary artery geometry and clinical information, (b) is the physical and mechanical information of the plaque integrated by encoding the basic mechanical equations, and (c) is the neural network output parameter: the stress and strain of the plaque. DETAILED DESCRIPTION

[0051] 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.

[0052] 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.

[0053] 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.

[0054] Example 1:

[0055] like Figure 1 As shown, the present invention provides a method for evaluating the mechanical state of coronary plaques by integrating physical and mechanical information, comprising the following steps:

[0056] Step 1: Medical imaging data acquisition: Collect medical imaging data of the coronary arteries, including CT angiography (CTA), magnetic resonance angiography (MRA), and optical coherence tomography (OCT), covering the three-dimensional geometric structure of the coronary arteries and their plaques. Collect clinical information of the patients, including age, gender, blood pressure, blood lipids, blood glucose, and other physiological parameters. Standardize the medical imaging data, including normalization and denoising. Use image segmentation algorithms to extract geometric information of coronary artery plaques, including plaque location, size, shape, and fibrous cap thickness. Standardize the clinical information to a mean of 0 and a standard deviation of 1 to improve the stability and generalization ability of the model.

[0057] Step 2: Constructing a neural network that integrates physical and mechanical information: A deep neural network architecture is designed, consisting of an input layer, hidden layers, and an output layer. The input layer receives preprocessed medical imaging data and clinical information. The hidden layer utilizes a fully connected neural network to extract features and perform nonlinear transformations on the input information. The output layer outputs mechanical state assessment results of the coronary plaque, including key mechanical information such as stress and strain. The solid mechanics principles of coronary plaques are cleverly embedded in the neural network's loss function. Specifically, a loss term based on physical constraints is designed, combining physical laws such as Hooke's law and equilibrium equations from elasticity mechanics. This loss term is then combined with the traditional data loss term to form a composite loss function.

[0058] Step 3: Construction of a training dataset: Traditional numerical methods, such as finite element analysis (FEA), are used to accurately determine the mechanical state of coronary plaques. This process fully considers factors such as the dynamics of blood flow, the material properties of the vessel wall, and the geometry of the plaque to obtain accurate stress and strain information. This mechanical information, derived through traditional methods, is paired with corresponding medical imaging data and clinical information to construct a high-quality training dataset.

[0059] Step 4: Model Training and Optimization: The collected dataset was divided into a training set, a validation set, and a test set, with a ratio of 70%, 15%, and 15%, respectively. The physical information neural network was trained using the training set, and the network parameters were adjusted using the stochastic gradient descent optimization algorithm to minimize the loss function. During training, the validation set was used to evaluate and optimize the model. Hyperparameters such as the network structure and learning rate were adjusted to improve the model's performance and generalization capabilities.

[0060] Step 5: Coronary Plaque Mechanical Status Assessment and Validation: The coronary artery medical imaging data and clinical information to be assessed are input into the trained physical information neural network. The network outputs the mechanical status assessment results of the plaque, including detailed information such as stress distribution and strain characteristics. Performance indicators such as sensitivity, specificity, and area under the curve (AUC) are calculated based on the predicted values and actual measured values, and the receiver operating characteristic curve is plotted. Performance indicators and false positive and false negative cases are analyzed to identify model deficiencies and implement improvements and optimization until performance meets the target.

[0061] Example 2:

[0062] like Figure 2 As shown in (a)-(c), in a specific embodiment of the present invention, the method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information specifically includes the following steps:

[0063] Step 1: Data Acquisition and Preprocessing: Coronary artery CT angiography, magnetic resonance angiography, and other imaging data were collected from the hospital's imaging database. Patient clinical information, including age, gender, systolic and diastolic blood pressure, total cholesterol, and triglycerides, was also collected. Coronary artery plaques were segmented and their geometric information extracted. Medical imaging data was normalized to map pixel values to the range [0, 1]. Clinical information was standardized to a mean of 0 and a standard deviation of 1.

[0064] Step 2: Design a neural network architecture based on the integration of physical and mechanical information: The input layer receives medical imaging data and clinical information vectors. The hidden layer consists of eight fully connected layers, which are used to integrate and classify features. The output layer uses a tanh activation function to output plaque stress and strain information. Based on the continuum mechanics equations, a physical information neural network model of coronary plaque mechanics is constructed. The key to building a physical information deep learning model lies in using basic mechanical equations as the neural network's loss function or constraints, ensuring that the neural network not only minimizes data errors during training but also complies with the laws of physics.

[0065] The loss function of the neural network includes the data residual term Physical residual term Boundary condition residual term and the initial condition residual composition:

[0066]

[0067] Among them, λ1, λ2, λ3, and λ4 represent the weight coefficients of different loss terms.

[0068]

[0069] Among them, N d is the number of sampling points in the measurement data, u i and are the predicted and measured displacement values at sampling point i, respectively.

[0070]

[0071] Among them, N p is the number of sampling points in the plaque mechanics equation.

[0072]

[0073] Among them, N p is the number of boundary condition sampling points, u bnd is the reference value of the boundary condition.

[0074]

[0075] Among them, N i is the number of initial condition sampling points, u int is the reference value of the initial condition.

[0076] Step 3: Use the finite element method to numerically solve the physical model of the plaque and calculate mechanical information such as stress and strain around the coronary plaque. Assume that the mechanical behavior of all vascular components of the coronary plaque (including the fibrous cap, lipid core, and vascular wall) conforms to that of linear elastic materials, expressed as follows:

[0077] σ ij =λδ ij ε kk +2Gε ij

[0078] Among them, σ ij and ε ij are the stress tensor and strain tensor, λ and G are the material parameters of the patch.

[0079] Therefore, the equation of motion of the plaque can be expressed as:

[0080]

[0081] Where t represents time, ρ s and u are the density and displacement of the plaque material, respectively, and b represents the body force.

[0082] Step 4: Divide the collected data into training, validation, and test sets in a ratio of 7:1.5:1.5. Use the training set to train the physical information neural network, setting the learning rate to 0.001, the batch size to 64, and the number of training rounds to 10,000. Finally, use the stochastic gradient descent algorithm to minimize the loss function to optimize the model parameters:

[0083]

[0084] Here, θ represents all the parameters of the neural network, including weights and biases.

[0085] During the training process, the learning rate and network structure are adjusted according to the evaluation results of the validation set, and an early stopping strategy is used to avoid overfitting.

[0086] Step 5: The coronary artery medical imaging data and clinical information to be evaluated are input into the trained physical information neural network. The network outputs the mechanical state assessment results of the plaque, including detailed information such as stress distribution and strain characteristics. Performance indicators such as sensitivity, specificity, and area under the curve (AUC) are calculated based on the predicted values and the actual measured values, and the receiver operating characteristic curve (ROC curve) is plotted. Performance indicators and false positive and false negative cases are analyzed to identify model deficiencies and make improvements and optimizations until performance meets the target. Finally, the coronary artery medical imaging data and clinical information to be evaluated are input into the trained model, which outputs the mechanical state assessment results of the plaque.

[0087] 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.

[0088] 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.

[0089] 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.

[0090] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0091] 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.

[0092] 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.

[0093] 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.

[0094] 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.

[0095] 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.

[0096] 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.

[0097] 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.

[0098] 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 method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information, characterized by: The following steps are involved: S1: Collect medical imaging data and clinical information; S2: Build a neural network that integrates physical and mechanical information. Combine Hooke's law and equilibrium equations of elasticity to design a loss term based on physical constraints, and incorporate it into the loss function of the neural network. S3: Use traditional numerical methods to solve the mechanical state of coronary artery plaques, obtain stress-strain related information, and pair it with corresponding medical imaging data and clinical information to construct a data set; S4: using the data set to train, optimize and test the neural network integrating the physical and mechanical information; S5: Input the coronary artery medical imaging data and clinical information to be evaluated into the trained physical information neural network, and output the mechanical state evaluation results of the plaque, including stress distribution and strain characteristics.

2. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 1, characterized in that: Step S1 specifically includes: Collect coronary artery CT angiography and magnetic resonance angiography from the image database; segment the coronary artery plaque, extract the geometric information of the plaque, perform normalization, and map its pixel value range to [0, 1]; The patients' clinical information, including age, gender, systolic blood pressure, diastolic blood pressure, total cholesterol, and triglycerides, was collected and standardized to have a mean of 0 and a standard deviation of 1.

3. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 1, characterized in that: The neural network integrating physical and mechanical information includes an input layer, a hidden layer and an output layer; The input layer receives pre-processed medical imaging data and clinical information; The hidden layer uses a fully connected neural network to perform feature extraction and nonlinear transformation on the input information; The output layer uses a tanh activation function to output stress and strain information of coronary artery plaques; The basic equations of continuum mechanics are used as loss functions or constraints, so that the neural network must not only minimize data errors but also satisfy the laws of physics during training.

4. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 3, characterized in that: The loss function of the neural network includes the data residual term Physical residual term Boundary condition residual term and the initial condition residual Among them, λ1, λ2, λ3, λ4 represent the weight coefficients of different loss terms; Among them, N d is the number of sampling points in the measurement data, u i and are the predicted and measured displacement values at sampling point i: Among them, N p is the number of sampling points in the plaque mechanics equation; Among them, N b is the number of boundary condition sampling points, u bnd is the reference value of the boundary condition; Among them, N i is the number of initial condition sampling points, u int is the reference value of the initial condition.

5. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 1, characterized in that: In step S3, the finite element method is used to numerically solve the physical model of the plaque. During the solution process, the dynamic characteristics of blood flow, the material properties of the blood vessel wall, and the geometric shape of the plaque are taken into account to calculate the stress and strain information around the coronary artery plaque. Assuming that the mechanical behavior of all vascular components of the coronary artery plaque conforms to that of linear elastic materials, the stress and strain information is expressed as follows: s ij =λδ ij e kk +2Gε ij Among them, σ ij and ε ij are the stress tensor and strain tensor, λ and G are the material parameters of the patch; The equation of motion of the plaque is expressed as: Where t represents time, ρ s and u are the density and displacement of the plaque material, respectively, and b represents the body force.

6. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 1, characterized in that: In step S4, the collected data set is divided into a training set, a validation set, and a test set; Train the neural network using the training set, setting the learning rate, batch size, and number of training rounds; The stochastic gradient descent algorithm is used to minimize the loss function to optimize the model parameters: Here, θ represents all the parameters of the neural network, including weights and biases; During the training process, the validation set is used to evaluate and optimize the model. The network structure and learning rate are adjusted, and an early stopping strategy is used to avoid overfitting. Use the test set to test the trained neural network.

7. The method for assessing the mechanical state of coronary plaques by integrating physical and mechanical information according to claim 1, characterized in that: The coronary artery medical imaging data and clinical information to be evaluated are input into a trained neural network that integrates physical and mechanical information, and the mechanical state assessment results of the plaque, including stress distribution and strain characteristics, are output. The sensitivity, specificity, and area under the curve (AUC) are calculated based on the predicted values and actual measured values, and the receiver operating characteristic curve is plotted. The performance indicators and false positive and false negative cases are analyzed to identify the deficiencies of the model and make improvements and optimizations until the performance meets the standards.

Images