A method and system for generating aortic valve bicuspidization diagnosis and treatment recommendations

Abstract

The application discloses an aortic valve bicuspid malformation diagnosis and treatment suggestion generation method and system, relates to the field of intelligent medical treatment, and comprises the following steps: acquiring aortic computed tomography data and bicuspid aortic valve ultrasonic image data of a target patient; constructing an aortic root model according to the aortic computed tomography data and the bicuspid aortic valve ultrasonic image data; performing blood flow dynamics calculation based on the aortic root model by using a machine learning method to obtain blood flow dynamics parameters; and generating a diagnosis and treatment suggestion of the target patient by using a regression model and a clustering algorithm according to the blood flow dynamics parameters. The application improves the accuracy and generation efficiency of the diagnosis and treatment suggestion.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent medical treatment, and in particular to a method and system for generating diagnosis and treatment suggestions for aortic valve bicuspid malformation. BACKGROUND

[0002] Aortic valve bicuspid malformation (BAV) is a common type of congenital heart disease, which can lead to aortic valve stenosis or regurgitation, and further cause a series of serious circulatory system problems. Currently, the diagnosis of BAV and acute aortic dissection (AAD) mainly relies on imaging examinations such as computed tomography angiography (CTA) and echocardiography, and the selection of treatment plans needs to consider the specific condition of the patient, physiological and pathological characteristics, and hemodynamic parameters. However, the traditional diagnosis and treatment process is time-consuming and largely dependent on the subjective experience of doctors, which limits the accuracy of diagnosis and effectiveness of treatment.

[0003] Therefore, it is particularly important to develop a system that can quickly and accurately diagnose BAV based on multi-modal image data and assist in making treatment decisions. SUMMARY

[0004] The purpose of the present application is to provide a method and system for generating diagnosis and treatment suggestions for aortic valve bicuspid malformation, which can improve the efficiency and accuracy of generating diagnosis and treatment suggestions.

[0005] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0006] In a first aspect, the present application provides a method for generating diagnosis and treatment suggestions for aortic valve bicuspid malformation, comprising:

[0007] Obtaining aortic computed tomography data and bicuspid aortic valve ultrasound image data of a target patient;

[0008] Constructing an aortic root model according to the aortic computed tomography data and the bicuspid aortic valve ultrasound image data;

[0009] Based on the aortic root model, performing hemodynamic calculation using a machine learning method to obtain hemodynamic parameters;

[0010] According to the hemodynamic parameters, generating diagnosis and treatment suggestions for the target patient using a regression model and a clustering algorithm.

[0011] In a second aspect, the present application provides a system for generating diagnosis and treatment suggestions for aortic valve bicuspid malformation, comprising:

[0012] A computer tomography device is configured to acquire computer tomography data of a target patient's aorta;

[0013] An ultrasound device is configured to acquire ultrasound image data of a target patient's bicuspid aortic valve;

[0014] A processor is connected to the computer tomography device and the ultrasound device respectively, and is configured to construct an aortic root model according to the computer tomography data and the ultrasound image data; perform hemodynamic calculation based on the aortic root model by using a machine learning method to obtain hemodynamic parameters; and generate a diagnosis and treatment suggestion for the target patient by using a regression model and a clustering algorithm according to the hemodynamic parameters.

[0015] According to the specific embodiments provided in the present application, the following technical effects are achieved:

[0016] The present application provides a bicuspid aortic valve malformation diagnosis and treatment suggestion generation method and system, which constructs an accurate aortic root model according to the computer tomography data and the ultrasound image data of the patient's aorta, performs hemodynamic calculation quickly and accurately by using a machine learning method, generates a diagnosis and treatment suggestion for the target patient by using a regression model and a clustering algorithm, reduces subjective judgment errors, and improves the accuracy and generation efficiency of the diagnosis and treatment suggestion. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 A flowchart of a bicuspid aortic valve malformation diagnosis and treatment suggestion generation method provided by an embodiment of the present application;

[0019] Figure 2 A detailed flowchart of a bicuspid aortic valve malformation diagnosis and treatment suggestion generation method provided by an embodiment of the present application;

[0020] Figure 3 A schematic diagram of a risk decision level determination process in an embodiment of the present application;

[0021] Figure 4 A structural schematic diagram of a bicuspid aortic valve malformation diagnosis and treatment suggestion generation system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0022] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0023] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0024] In one exemplary embodiment, as shown in Figure 1 and Figure 2 , a method for generating aortic valve bicuspidization diagnosis and treatment suggestion is provided, comprising the following steps 101 to step 104.

[0025] Step 101, obtaining aortic computed tomography data and bicuspid aortic valve ultrasound image data of a target patient.

[0026] The CTA device is used to scan the aortic root of the target patient to obtain the aortic computed tomography data.

[0027] Step 102, constructing an aortic root model according to the aortic computed tomography data and the bicuspid aortic valve ultrasound image data.

[0028] In one exemplary embodiment, step 102 comprises the following steps 201 to step 204.

[0029] Step 201, establishing an aortic root blood vessel model according to the aortic computed tomography data.

[0030] Specifically, a personalized aortic root blood vessel model is established by using a three-dimensional reconstruction method according to the aortic computed tomography data. Wherein, the convolutional neural network is used to realize the rapid segmentation and reconstruction of the patient's aortic root blood vessel CTA image.

[0031] Step 202, establishing a parameterized bicuspid aortic valve model according to the bicuspid aortic valve ultrasound image data.

[0032] Specifically, based on the geometric data measured by the bicuspid aortic valve ultrasound image data, a parameterized bicuspid aortic valve model is constructed by using computer aided design (Computer Aided Design, CAD) forward engineering technology. The valve parameters contained include BAV phenotype, annulus diameter, commissure height, geometric height, free edge length and sinus tube junction diameter.

[0033] Step 203, according to the two-leaflet aortic valve ultrasound image data, determine the valve offset angle and the valve offset height.

[0034] Step 204, according to the valve offset angle and the valve offset height, perform Boolean operation assembly on the aortic root blood vessel model and the parameterized two-leaflet aortic valve model to obtain an aortic root model.

[0035] In the assembly process, the valve offset angle and the valve offset height are used as assembly parameters. By constructing the complete aortic root geometry, an accurate geometric basis is provided for subsequent blood flow dynamics calculation.

[0036] Step 103, based on the aortic root model, perform blood flow dynamics calculation using a machine learning method to obtain blood flow dynamics parameters. Provide important physiological parameters for diagnosis and treatment decisions.

[0037] Based on the aortic root model, the related blood flow dynamics deep learning algorithm is used to realize the fast calculation of the aortic personalized blood flow dynamics parameters of the BAV patient.

[0038] In one exemplary embodiment, the machine learning method is a physically embedded neural network. The physical equation (Navier-Stokes equation, NS equation) is used as part of the loss function of the deep learning neural network to form a physically embedded neural network. The loss function after embedding the NS equation is:

[0039] Loss=loss data +ω*loss physics ;

[0040]

[0041] Where Loss is the total loss function value, loss data is the data fitting term, ω is the physical consistency coefficient, loss physics is the physical consistency term, N data is the number of training set samples, U true (X j ,t j ) is the sample point true value, U pred (X j ,t j ) is the sample point predicted value, X j is the spatial coordinate, t j is the time coordinate, e i (X j ,t j ) is the difference between the neural network predicted solution and the actual solution of the NS equation.

[0042] The physical embedding neural network is obtained by pre-training using a shape-force response database. The shape-force response database includes blood flow dynamics simulation data corresponding to different bicuspid deformity phenotypes.

[0043] In an exemplary embodiment, the blood flow dynamics parameters include preoperative blood flow dynamics parameters and postoperative blood flow dynamics parameters. Step 103 includes steps 301 to 303.

[0044] Step 301, based on the aortic root model, a virtual surgery is performed to obtain a post-virtual surgery aortic root model.

[0045] Step 302, based on the aortic root model, a blood flow dynamics calculation is performed using a machine learning method to obtain preoperative blood flow dynamics parameters.

[0046] Step 303, based on the post-virtual surgery aortic root model, a blood flow dynamics calculation is performed using a machine learning method to obtain postoperative blood flow dynamics parameters.

[0047] The present application further integrates a clinical physiological and pathological database, a shape-force response database and a virtual surgery database to provide comprehensive information support for blood flow dynamics calculation.

[0048] The clinical physiological and pathological database and the shape-force response database are constructed by using existing clinical BAV case data and patient aortic computational fluid dynamics results. The shape-force response database is constructed using computational fluid dynamics technology and aortic root models. The virtual surgery database is established by using clinical BAV case data and CAD technology for virtual surgery, providing data set support for machine learning algorithms.

[0049] The shape-force response database includes different BAV phenotype classification data obtained based on analysis of a large number of historical cases, which classifies BAV phenotypes into fusion type, symmetric type and partial fusion type. The aortic blood flow dynamics simulation data of different BAV phenotypes is classified to construct the shape-force response database. Different shape-force response databases are established for different phenotype types, and different types of databases are directly corresponded to patients with different BAV phenotypes in clinical application.

[0050] The clinical physiological and pathological database is collected through patient visits, mainly including age, gender, medical history, surgical history, aortic valve auscultation area murmur, nature and degree of heart murmur, chest pain, dyspnea, syncope, blood pressure, heart rate, heart rhythm, etc.

[0051] In constructing the virtual surgery database, CAD technology is adopted to create highly precise three-dimensional models of the aortic root and valve. These models not only provide a solid foundation for virtual surgery, but also allow doctors to simulate and evaluate different surgical plans, such as valve repair or replacement, before surgery, thereby optimizing treatment plans and improving surgical success rates. In addition, the virtual surgery database achieves continuous optimization and updating of surgical plans by accumulating actual surgical case data and analyzing them using machine learning algorithms.

[0052] Step 104, according to the hemodynamic parameters, using regression model and clustering algorithm to generate the diagnosis and treatment suggestion of the target patient.

[0053] In an exemplary embodiment, first, according to the preoperative hemodynamic parameters and the postoperative hemodynamic parameters, the influence coefficients of different parts are determined by using the regression model. Then, according to the influence coefficients of different parts and the clinical information of the target patient, the diagnosis and treatment suggestion of the target patient is determined by using the clustering algorithm. The diagnosis and treatment suggestion includes reference treatment plan and reference surgical strategy.

[0054] As shown in Figure 3 , first, the hemodynamic parameters of the patient's aortic arch, ascending aorta, sinus, left ventricle, and coronary artery before and after virtual surgery are used as input to perform principal component analysis and correlation coefficient feature extraction using the regression model, obtaining the influence coefficients of different parts. These influence coefficients and the patient's clinical information (gender, age, medical history, BAV type, etc.) are used as input features of the clustering algorithm (or decision tree), and the risk decision level (such as first, second, and third) is given by the clustering algorithm (or decision tree), thereby improving the accuracy and stability of the diagnosis and treatment evaluation decision reasoning results in the treatment of BAV-AAD patients.

[0055] Further, the aortic valve bicuspid malformation diagnosis and treatment suggestion generation method further comprises step 105.

[0056] Step 105, visualizing the aortic computed tomography data, the bicuspid aortic valve ultrasound image data, the aortic root model, the hemodynamic parameters, and the diagnosis and treatment suggestion.

[0057] The application integrates multi-modal image data processing and three-dimensional reconstruction technology to construct an accurate aortic root model, combines with a clinical physiological and pathological database, a shape-force response database and a virtual surgery database, and can quickly calculate hemodynamic parameters through a machine learning method, generates a diagnosis and treatment suggestion for a target patient by using a regression model and a clustering algorithm, improves the accuracy and efficiency of the diagnosis and treatment suggestion, reduces the subjective judgment error of doctors, greatly shortens the time required in a traditional diagnosis process, and is helpful for reasonable allocation of medical resources, especially in resource-limited areas. In addition, the decision diagnosis opinion is provided according to the personalized clinical information of the patient, and the accurate diagnosis and personalized treatment suggestion are helpful for improving the treatment experience and satisfaction of the patient, and provide an efficient and accurate auxiliary tool for diagnosis and treatment of BAV-AAD patients.

[0058] Based on the same inventive concept, the application also provides an aortic valve bicuspid malformation diagnosis and treatment suggestion generation system for implementing the aortic valve bicuspid malformation diagnosis and treatment suggestion generation method. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme described in the above method, so the specific limitations in one or more aortic valve bicuspid malformation diagnosis and treatment suggestion generation system embodiments provided below can be referred to the limitations of the aortic valve bicuspid malformation diagnosis and treatment suggestion generation method described above, and will not be described here.

[0059] In an exemplary embodiment, as shown in Figure 4 An aortic valve bicuspid malformation diagnosis and treatment suggestion generation system is provided, which includes a computed tomography device 401, an ultrasound device 402 and a processor 403.

[0060] The computed tomography device 401 is used to collect aortic computed tomography data of a target patient.

[0061] The ultrasound device 402 is used to collect bicuspid aortic valve ultrasound image data of the target patient.

[0062] The processor 403 is connected with the computed tomography device 401 and the ultrasound device 402 respectively, and the processor 403 is used to construct an aortic root model according to the aortic computed tomography data and the bicuspid aortic valve ultrasound image data; perform hemodynamic calculation by using a machine learning method based on the aortic root model to obtain hemodynamic parameters; and generate a diagnosis and treatment suggestion for the target patient by using a regression model and a clustering algorithm according to the hemodynamic parameters.

[0063] Further, the aortic valve bicuspid malformation diagnosis and treatment suggestion generation system further includes a terminal 404.

[0064] The terminal 404 is connected with the processor 403, and the terminal 404 is used for displaying the aortic computer tomography data, the bicuspid aortic valve ultrasound image data, the aortic root model, the hemodynamic parameters and the diagnosis and treatment suggestions.

[0065] Further, the aortic valve bicuspid malformation diagnosis and treatment suggestion generation system further comprises a database 405. The processor 403 is connected with the database 405.

[0066] The processor 403 is further used for optimizing the aortic computer tomography data and the bicuspid aortic valve ultrasound image data. The processor 403 simulates the hemodynamics under different physiological conditions. The processor 403 is deployed on a cloud computing platform, so as to facilitate remote access and use.

[0067] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of the related data need to comply with the relevant regulations.

[0068] In the present application, all actions of obtaining signals, information or data are performed under the premise of complying with the corresponding data protection regulations and policies of the place, and under the premise of obtaining authorization from the owner of the corresponding device.

[0069] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0070] The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a blockchain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0071] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.

[0072] The principles and implementation modes of the present application are described by applying specific examples herein, and the above-mentioned embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. An aortic valve bicuspidization malformation diagnosis and treatment recommendation generation method, characterized by, The aortic valve bicuspid malformation diagnosis and treatment suggestion generation method comprises: Obtain aortic computed tomography data and bicuspid aortic valve ultrasound image data of a target patient; Construct an aortic root model according to the aortic computed tomography data and the bicuspid aortic valve ultrasound image data; Based on the aortic root model, blood flow dynamics calculation is performed by using a machine learning method to obtain blood flow dynamics parameters; the machine learning method is a physically embedded neural network; the physically embedded neural network is obtained by pre-training using a shape-force response database; the shape-force response database includes blood flow dynamics simulation data corresponding to different bicuspid malformation phenotypes; wherein the NS equation is used as part of the loss function of the deep learning neural network to form the physically embedded neural network; the loss function after embedding the NS equation is: Loss = loss data + ω * loss physics ; Where Loss is the total loss function value, loss data Here, ω represents the data fitting term, ω is the physical consistency coefficient, and loss is... physics For physical consistency terms, N data U represents the number of samples in the training set. true (X j ,t j ) represents the true value of the sample point, U pred (X j ,t j X represents the predicted value of the sample point. j Let t be the spatial coordinate. j e is the time coordinate. i (X j ,t j The difference between the solution predicted by the neural network and the actual solution of the Navier-Stokes equations is represented by . According to the blood flow dynamics parameters, a regression model and a clustering algorithm are used to generate a diagnosis and treatment suggestion for the target patient.

2. The method of claim 1, wherein the method further comprises: According to the aortic computed tomography data and the bicuspid aortic valve ultrasound image data, an aortic root model is constructed, specifically comprising: Establish an aortic root blood vessel model according to the aortic computed tomography data; According to the bicuspid aortic valve ultrasound image data, a parameterized bicuspid aortic valve model is established; According to the bicuspid aortic valve ultrasound image data, the valve offset angle and the valve offset height are determined; According to the valve offset angle and the valve offset height, Boolean operation assembly is performed on the aortic root blood vessel model and the parameterized bicuspid aortic valve model to obtain the aortic root model.

3. The method of claim 1, wherein the method further comprises: The blood flow dynamics parameters include preoperative blood flow dynamics parameters and postoperative blood flow dynamics parameters; Based on the aortic root model, blood flow dynamics calculation is performed by using a machine learning method to obtain blood flow dynamics parameters, specifically comprising: Based on the aortic root model, a virtual operation is performed to obtain an aortic root model after the virtual operation; Based on the aortic root model, blood flow dynamics calculation is performed by using a machine learning method to obtain preoperative blood flow dynamics parameters; Based on the aortic root model after the virtual operation, blood flow dynamics calculation is performed by using a machine learning method to obtain postoperative blood flow dynamics parameters.

4. The method of claim 3, wherein the method further comprises: According to the blood flow dynamics parameters, a regression model and a clustering algorithm are used to generate a diagnosis and treatment suggestion for the target patient, specifically comprising: According to the preoperative blood flow dynamics parameters and the postoperative blood flow dynamics parameters, a regression model is used to determine the influence coefficients of different parts; According to the influence coefficients of different parts and the clinical information of the target patient, a clustering algorithm is used to determine the diagnosis and treatment suggestion for the target patient.

5. The method of claim 1, wherein the method further comprises: The diagnosis and treatment suggestion includes a reference treatment plan and a reference operation strategy.

6. The method of claim 1, wherein the method further comprises: The aortic valve bicuspid malformation diagnosis and treatment suggestion generation method further comprises: Visualize the aortic computed tomography data, the bicuspid aortic valve ultrasound image data, the aortic root model, the blood flow dynamics parameters, and the diagnosis and treatment suggestion.

7. An aortic valve bicuspidization diagnosis and treatment recommendation generation system, applied to the aortic valve bicuspidization diagnosis and treatment recommendation generation method of any one of claims 1-6, characterized in that, The aortic valve bicuspid malformation diagnosis and treatment suggestion generation system comprises: A computer tomography device is configured to acquire computer tomography data of a target patient's aorta; An ultrasound device is configured to acquire ultrasound image data of a target patient's bicuspid aortic valve; A processor is connected to the computer tomography device and the ultrasound device respectively, and is configured to construct an aortic root model according to the computer tomography data and the ultrasound image data; perform hemodynamic calculation based on the aortic root model by using a machine learning method to obtain hemodynamic parameters; and generate a diagnosis and treatment suggestion for the target patient according to the hemodynamic parameters by using a regression model and a clustering algorithm.

8. The aortic valve bicuspidization diagnosis and treatment recommendation generation system of claim 7, wherein, The aortic valve bicuspid malformation diagnosis and treatment suggestion generation system further comprises: A terminal connected to the processor and configured to display the computer tomography data, the ultrasound image data, the aortic root model, the hemodynamic parameters, and the diagnosis and treatment suggestion.

Images