Simulation method and system for optimizing heart turn-back induction based on infarction area distance

Abstract

The invention discloses a simulation method and system for optimizing heart turn-back induction based on an infarction region distance, and belongs to the field of heart modeling simulation. A personalized high-fidelity ventricle model is constructed based on a CMRLGE image of an arrhythmia patient, each model is divided into regions, pacing points are randomly selected to induce VT, and induction states such as unstable turn-back, stable turn-back and non-persistent turn-back are obtained. Extracting a left ventricular grid model, calibrating a left ventricular myocardial infarction area and a boundary, and measuring a left ventricular stimulation point-infarction area distance for a pace-making point; and carrying out normality test and significance analysis on the distance data of all patients in the induction state, and if significance is met, optimizing the position of the pace-making point, namely summarizing the distance data in the stable turn-back state, and determining the distance range of easily inducing turn-back. The simulation method disclosed by the invention is easy to induce a turn-back similar to a clinical real situation, and can effectively guide a clinician to accurately find an optimal ablation target spot during a VT ablation operation.

Description

Technical Field

[0001] The present invention relates to the field of cardiac modeling and simulation, and particularly to a simulation method and system for optimizing cardiac reentry induction based on the distance from the infarct region. Background Art

[0002] Ventricular tachycardia (VT) is a potentially life-threatening arrhythmia commonly seen in patients with myocardial infarction (MI) and is one of the main causes of sudden cardiac death (SCD). In recent years, computational modeling has been widely used in non-invasive research and treatment evaluation of fatal arrhythmias, especially in clinically assisting the guidance of VT ablation, showing important value.

[0003] Numerous studies based on clinical observations, ex vivo cardiac experiments, and computational simulations generally propose and support a key hypothesis: the closer the stimulation point is to the infarct region, the easier it is to induce the reentry mechanism. The theoretical basis of this hypothesis is that the infarct border region is often accompanied by fibrosis, gap junction remodeling, and ion channel dysfunction, thereby forming local slow conduction regions and unidirectional conduction block regions, providing an electrophysiological basis for the formation of reentry circuits.

[0004] Although existing studies generally find that the positional relationship between the stimulation point and the scar region significantly affects the induction of reentry, that is, the closer the stimulation point is to the scar region, the easier it is to induce reentry, the specific distance of "closeness" is still in a state of vague definition and lacks a unified and quantitative reference standard. This situation makes it difficult for clinicians to accurately locate the ablation target when formulating ablation treatment plans using cardiac modeling and simulation, and there is a large uncertainty in the treatment effect. At the research level, due to the extremely high requirements for computing resources in cardiac modeling and simulation, there are currently few systematic simulation experiments based on individual patient data in real clinical settings to verify the above hypothesis. This not only leads to the lack of practical support for academic theories but also makes it difficult for clinical practice to obtain effective theoretical guidance. Therefore, it is urgent to construct a method system based on real clinical data and deeply integrated with computational simulation. This method system can provide solid theoretical support and reliable technical basis for individualized VT treatment strategies by systematically evaluating the impact of the distance between the stimulation point and the infarct region on reentry susceptibility. Summary of the Invention

[0005] In view of the above problems, the present invention proposes a simulation method and system for optimizing cardiac reentry induction based on the distance from the infarct region, which optimizes the stimulation point-infarct region distance based on the left ventricular model to support clinical VT treatment.

[0006] The present invention adopts the following technical solutions:

[0007] In the first aspect, the present invention proposes a simulation method for optimizing cardiac reentry induction based on the distance from the infarct region, including the following steps:

[0008] Step 1: Based on the CMR_LGE images of several arrhythmia patients, establish a personalized high-fidelity ventricular model;

[0009] Step 2: For each high-fidelity ventricular model, perform the following operations respectively:

[0010] Divide the ventricular model into regions and randomly select pacing points as stimulation points to induce VT, and obtain the induction status of each stimulation point; the induction status includes unstable reentry, stable reentry, and non-sustained reentry;

[0011] Extract the left ventricular mesh model from the ventricular model, calibrate the left ventricular myocardial infarction region and its boundary, and measure the distance between the left ventricular stimulation point and the infarction region for the pacing points;

[0012] Step 3: Perform a normality test on the distances between the stimulation points and the infarction regions of all induction statuses of several patients, and further perform a significance analysis according to the normality test results. If significance is met, execute Step 4 to optimize the pacing point positions; otherwise, the pacing points in Step 2 are already in the optimal interval and do not need to be optimized;

[0013] Step 4: Summarize the distance data between the stimulation points and the infarction regions in all stable reentry states to obtain the distance range between the stimulation points and the infarction regions that are prone to induce reentry.

[0014] Furthermore, the process of establishing the high-fidelity ventricular model is as follows:

[0015] Preprocess the CMR_LGE images of arrhythmia patients, including the segmentation of the myocardium and its infarction region, and the interpolation and reconstruction of the ventricle, to obtain a ventricular model;

[0016] Perform finite element mesh division on the ventricular model to obtain a ventricular mesh model;

[0017] Construct the myocardial fiber helix for the ventricular mesh model, and couple the corresponding myocardial cell electrophysiological models for normal tissue, gray matter tissue, and completely infarcted tissue respectively.

[0018] Furthermore, the process of extracting the left ventricular mesh model from the ventricular model includes:

[0019] Project the myocardial region onto the MRI image based on the variational implicit function interpolation method. Starting from the seed points on the selected left and right ventricular endocardiums, gradually expand and label the ventricular cavity through the region growing algorithm. Finally, generate left and right ventricular labels to identify the left and right ventricular regions, and then map the labels to the ventricular model generated by S1 after converting them into a recognizable data format. Finally, generate a mesh data file with only the left ventricular mesh to obtain the extracted left ventricular mesh model.

[0020] Furthermore, the process of calibrating the left ventricular myocardial infarction region and its boundary includes:

[0021] According to the tissue type to which the volume mesh in the left ventricle mesh model belongs, the volume mesh tetrahedron is divided into one or two of normal tissue, gray matter tissue and complete infarction tissue, and a normal myocardial point set, a gray matter point set and a complete infarction point set are obtained respectively;

[0022] Demarcate the infarct area: merge the gray matter point set and the complete infarct point set to obtain the complete infarct area point set on the myocardium;

[0023] Demarcate the boundary of the infarct area: find all subsets that belong to both the normal myocardial point set and the gray matter point set as the transition boundary of the semi-infarct area; find all subsets that belong to both the normal myocardial point set and the complete infarct point set as the transition boundary of the complete infarct area; merge the point set of the transition boundary of the semi-infarct area and the point set of the transition boundary of the complete infarct area to obtain a complete infarct area boundary point set on the myocardium.

[0024] Furthermore, the infarct region point set includes an infarct region boundary point set.

[0025] Further, the process of measuring the distance between the left ventricular stimulation point and the infarct area for the pacemaker point includes:

[0026] For each induced state pacemaker point, first determine whether the point exists in the infarct area point set file. If so, no further calculation is required and the point is directly skipped, and the distance between the stimulation point and the infarct area is recorded as 0; otherwise, enter the iterative search process;

[0027] In each iteration, the tetrahedral mesh where the boundary point set of the current infarct region is located is found, and the points in the boundary point set of the current infarct region are removed from the vertex set of the tetrahedral mesh as the boundary point set of the infarct region after diffusion, until the pacing point appears in the new boundary point set for the first time. The number of iterations at this time is recorded as the corresponding stimulation point-infarct region distance, and the unit of this distance is the number of layers of grid units.

[0028] Furthermore, in step 3, the normality test uses the D'Agostino-Pearson test, the Anderson-Darling test, the Shapiro-Wilk test, and the Kolmogorov-Smirnov test.

[0029] Furthermore, if the p-values ​​of at least two test methods were significantly greater than the significance level of 0.05, one-way analysis of variance was used to evaluate whether there were significant differences in the variable among multiple groups under different induced states; otherwise, the Kruskal-Wallis test was used to evaluate the significant differences.

[0030] Further, in step 4, the distance range between the stimulation point and the infarct region that is prone to induce reentry is obtained through violin plot analysis. The quartile value is taken as the upper limit of the distance range, and the distance point corresponding to the density peak region is taken as the optimized distance between the stimulation point and the infarct region within the range from 0 to the quartile value.

[0031] In a second aspect, the present invention proposes a simulation system for optimizing cardiac reentry induction based on the infarct region distance, which is used to implement the above-mentioned simulation method for optimizing cardiac reentry induction based on the infarct region distance.

[0032] The beneficial effects of the present invention are as follows:

[0033] (1) Innovative calculation method for the distance between the stimulation point and the infarct region

[0034] The iterative tomographic peeling method (instead of the traditional Euclidean distance) is used to peel off the normal myocardial units adjacent to the scar boundary layer by layer, and the distance is quantified by the number of peeled grids. This method can more accurately reflect the myocardial fiber orientation and anisotropic conduction characteristics, avoiding the defect that the traditional Euclidean distance cannot reflect the myocardial fiber orientation and anisotropic conduction characteristics.

[0035] (2) Scientific and reasonable data analysis method

[0036] The combination of normal distribution test and significance analysis is adopted, and appropriate statistical methods (ANOVA or Kruskal-Wallis test) are selected according to the data distribution to ensure the accuracy and reliability of the analysis results.

[0037] (3) Application value and clinical significance

[0038] It can improve the accuracy of stimulation point selection in cardiac simulation modeling, provide data support for ventricular arrhythmia surgery (such as VT ablation), and improve the success rate of the surgery. Specifically, accurately defining the optimal distance between the stimulation point and the infarct region is prone to induce reentry similar to the clinical real situation, which can effectively guide clinicians to accurately find the best ablation target during VT ablation surgery, improve the success rate of a single surgery, reduce the number of repeated ablations, reduce the risks and pain brought to patients by the surgery, and at the same time can also reduce the economic burden on patients. In addition, this precise treatment strategy can also shorten the hospitalization time of patients, optimize the allocation of medical resources, improve the overall medical efficiency and quality, and has important practical significance and broad application prospects for promoting the development of the field of cardiac disease treatment. Description of the Drawings

[0039] Figure 1 is the overall flowchart of the present invention;

[0040] Figure 2 is the flowchart for left ventricle extraction;

[0041] Figure 3 It is an effect diagram of left ventricular infarction;

[0042] Figure 4 It is a flow chart for quantifying the distance between the stimulation point and the infarction area;

[0043] Figure 5 It is the distance range between the stimulation point and the infarction area that can stably induce reentry. Specific implementation manners

[0044] The present invention will be further described and explained below in conjunction with specific implementation manners. The described embodiments are only examples of the present disclosure content and do not delimit the scope of limitation. Without conflict, the technical features of each implementation manner in the present invention can be combined accordingly.

[0045] The accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0046] The flow chart shown in the accompanying drawings is only an exemplary illustration and does not necessarily include all steps. For example, some steps can be decomposed, while some steps can be combined or partially combined. Therefore, the actual execution order may be changed according to the actual situation.

[0047] A simulation method for optimizing cardiac reentry induction based on the distance of the infarction area provided by the present invention, as Figure 1 shown, the specific steps are as follows:

[0048] S1, construct a ventricular mesh model.

[0049] Obtain the CMR_LGE images and clinical diagnosis and treatment data of several arrhythmia patients, preprocess the image stack, including segmentation of the myocardium and its infarction area, interpolation reconstruction of the ventricle, etc., to obtain a ventricular model; perform finite element mesh division on the ventricular model to obtain a ventricular mesh model; construct the myocardial fiber helix direction for the ventricular mesh model, and couple the corresponding myocardial cell electrophysiological models for normal tissue, gray matter tissue (partial fibrosis area), and completely infarcted tissue respectively.

[0050] First, the steps of constructing the ventricular model are based on the standard preprocessing process of cardiac structure modeling for arrhythmia patients using cardiac magnetic resonance imaging (CMR) technology. The aim is to generate an accurate ventricular model through segmentation and reconstruction, providing a quantitative basis for subsequent analysis. CMR_LGE images, namely Cardiac Magnetic Resonance Late Gadolinium Enhancement (CMR-LGE), are a technique that involves delayed scanning after injecting a gadolinium contrast agent, used to detect fibrotic / infarcted regions in myocardial tissue. Normal myocardium appears hypointense in the delayed phase due to contrast agent clearance, while infarcted regions show hyperintensity due to increased extracellular space leading to contrast agent retention. Image stack preprocessing refers to the standardization of multi-sequence, multi-slice cardiac magnetic resonance images (3D datasets) to improve data quality and adapt it for subsequent analysis. Among them, myocardial and infarct region segmentation separates myocardial tissue from surrounding tissue through image segmentation algorithms and further identifies infarcted regions. It can be achieved using region growing algorithms (such as the Otsu algorithm) or end-to-end segmentation using convolutional neural networks like U-Net. Ventricular interpolation reconstruction is for insufficient image resolution or missing parts of the region, and reconstructs a complete three-dimensional ventricular geometry model through interpolation algorithms.

[0051] The reconstructed ventricular model includes the left ventricle and the right ventricle, represented in the form of a surface mesh or a volume voxel. After completing the finite element mesh generation, a ventricular mesh model is obtained. Further, by combining the helix direction and electrophysiological properties of myocardial fibers, a multi-tissue numerical model is constructed to simulate the dynamic behavior of cardiac electrical activity and reveal the potential mechanisms of arrhythmia. The helix direction of myocardial fibers refers to the arrangement direction of myocardial cells in three-dimensional space, showing a significant anisotropic structure. Subendocardial fibers are circumferentially arranged, middle myocardial fibers are helically twisted, and subepicardial fibers are longitudinally arranged. The coupling of the electrophysiological model of myocardial cells refers to the combination of the mathematical model describing the electrical activity of a single myocardial cell with the macroscopic ventricular model. Here, normal myocardial tissue refers to the healthy myocardial region not affected by fibrosis or infarction; gray matter tissue refers to the region where myocardial fibrosis and normal myocardium are mixed, usually corresponding to the border zone of chronic myocardial infarction, showing partial fibrosis; completely infarcted tissue refers to the completely necrotic and inactive myocardial region.

[0052] This step constructs a high-fidelity ventricular model by integrating the helix direction and electrophysiological model. The above image stack preprocessing method, electrophysiological model of myocardial cells, etc. can be implemented using existing technologies in the field.

[0053] S2. Divide the regions of the ventricular model and randomly select pacing points as stimulation points to induce VT, and obtain the induction status of each stimulation point.

[0054] In this step, according to the induction method similar to that in clinical practice, programmed electrical stimulation is used to pace the ventricular model of each patient from multiple positions. According to the grading standard of the American Heart Association (AHA), 17 pacing points are randomly selected as stimulation points from 17 divided regions of the left ventricle in the ventricular model. The distribution of these pacing points basically covers the potential sites where a larger range of lesions may exist;

[0055] Continuous stimulation is applied to each of these 17 pacing points, and the stimulation scheme for the above 17 pacing points is as follows: First is the S1 stimulation. By selecting a cubic region with a size of 1mm * 1mm * 1mm centered on the pacing site on the endocardium of the personalized ventricular model, continuous stimulation is applied 6 times. Current stimulation is used, with a stimulation duration of 5ms and a stimulation interval of 600ms. Generally, no pacing site will induce VT during the S1 stimulation. If the time node of the first S1 stimulation is at 0ms, then the time node after the S1 stimulation is completed is 3000ms; then the S2 stimulation will be carried out. The initial value of the interval time between the S2 stimulation and the S1 stimulation is set to 250ms. If VT cannot be induced at the 3250ms time node, then it will be decreased in steps of 10ms to induce VT until S2 < 150ms when the stimulation stops. If VT is induced during the S2 stimulation, the start time node of the simulation is t = 600x5 + S2 (ms); if the S2 stimulation cannot induce VT, then the S3 stimulation is carried out. The S3 stimulation method is the same as the S2 stimulation. If VT is induced during the S3 stimulation, the start time node of the simulation is t = 600x5 + S2 + S3 (ms), and so on; if the S3 stimulation cannot induce VT, then the S4 stimulation is carried out. If VT still cannot be induced until the S4 stimulation, it means that this point will not induce VT and there is no need to continue the simulation for this pacing point; the induction state of each pacing point is obtained. The induction state includes unstable reentry, stable reentry, and non-sustained reentry. Here, unstable reentry generally refers to the electroactivity showing dynamic changes of multiple reentry loops, with a cycle length fluctuation > 20ms, being easily terminated or transformed into ventricular fibrillation. Stable reentry generally refers to the continuous existence of a single reentry loop, with a stable cycle length (variation < 5ms), and requires external intervention to terminate. Non-sustained reentry generally refers to the synchronization of electroactivity, no reentry loop is detected, and the excitation time dispersion < 10ms. Combining the AHA standardized zoning and electrophysiological modeling, the reentry risk of each region of the left ventricle is evaluated through systematic stimulation.

[0056] For S3, measure the distance between the left ventricular stimulation point and the infarct region for the pacing point, including three parts: left ventricle extraction, calibration of the left ventricular myocardial infarct region and its boundary, and distance calculation.

[0057] For S31, divide the left ventricular cavity and extract the left ventricular grid model.

[0058] This process is implemented based on the ventricular model obtained in step S1. As Figure 2 shown, the myocardial region is projected onto the MRI image based on the variational implicit function interpolation method. Through the region growing algorithm, starting from the selected seed points (the endocardium of the left and right ventricles), the ventricular cavity is gradually expanded and labeled. Finally, the left and right ventricular labels are generated to identify the left and right ventricular regions. After converting the labels into a data format recognizable by OpenCARP, they are mapped into the ventricular mesh model generated in S1. Finally, a mesh data file with only the left ventricular mesh, that is, the left ventricular mesh model, is generated.

[0059] This process combines the variational implicit function interpolation, region growing algorithm, and format conversion technology to achieve accurate segmentation and reconstruction from MRI images to the left ventricular mesh model. The variational implicit function interpolation method adopted in this embodiment is an implicit surface reconstruction method based on energy minimization, which is used to fit the discrete anatomical structure of the myocardial region into a continuous implicit surface for subsequent projection into the MRI image space. The region growing algorithm is an image segmentation technology based on voxel similarity. Starting from the selected seed points, adjacent voxels are gradually merged through threshold or feature similarity criteria. For example, the left ventricular region is labeled as 1, the right ventricular region is labeled as 2, and the background is 0 to form a binary mask image, and all mesh cells labeled as 1 are extracted.

[0060] S32, calibrate the myocardial infarction region of the left ventricle and its boundary.

[0061] In this step, from the generated left ventricular mesh model (composed of volume meshes), according to the tissue type to which the volume meshes belong, the volume mesh tetrahedrons are divided into one or two of normal tissue, gray matter tissue, and completely infarcted tissue, respectively obtaining three point sets, namely the set of tetrahedron vertices included in normal tissue (abbreviated as the normal myocardial point set), the set of tetrahedron vertices included in gray matter tissue (abbreviated as the gray matter point set), and the set of tetrahedron vertices included in completely infarcted tissue (abbreviated as the completely infarcted point set). A tetrahedron may belong to two tissue types at the same time, and in this case, it is divided into two point sets respectively.

[0062] Calibrate the infarcted region: Combine the gray matter point set and the completely infarcted point set to obtain the complete infarcted region point set on the myocardium. As Figure 3 shown in (a) of

[0063] Calibrate the boundary of the infarcted region: Find all subsets that belong to both the normal myocardial point set and the gray matter point set as the transition boundary of the semi-infarcted region; find all subsets that belong to both the normal myocardial point set and the completely infarcted point set as the transition boundary of the completely infarcted region; further improve the contour of the infarcted region through these two boundaries, as Figure 3As shown in (b) of [reference], a complete set of infarct region boundary points on the myocardium is obtained by combining the set of points on the excessive boundary of the semi-infarct region and the set of points on the excessive boundary of the complete infarct region. It should be noted that the infarct region point set includes the infarct region boundary point set.

[0064] S33. Calculate the left ventricular stimulation point-infarct region distance based on the infarct region and its boundary.

[0065] As Figure 4 shown, for each pacing point in the induced state, first determine whether it is located within the infarct region, that is, check whether the point exists in the infarct region point set file. If the pacing point is already in the infarct region, no further calculation is required, and the point is directly skipped, and the stimulation point-infarct region distance is recorded as 0; otherwise, enter the iterative search process.

[0066] In each iteration process, continuously expand the infarct region boundary, that is, find the tetrahedral mesh where the current infarct region boundary point set is located, and remove the points in the current infarct region boundary point set from the vertex set of the tetrahedral mesh as the expanded infarct region boundary point set until the pacing point first appears in the new boundary point set, and then record the number of iterations at this time as the corresponding stimulation point-infarct region distance. The unit of this distance is the number of layers of the grid cell. In the present invention, the search range is gradually expanded through continuous iteration until the pacing point intersects with the boundary point set, and all pacing points in the induced state are traversed to form an array containing the stimulation point-infarct region distances corresponding to all pacing points.

[0067] The stimulation point-infarct region distances for all induced states of each patient are calculated through the above S1 - S3.

[0068] S4. Perform a normality test on the stimulation point-infarct region distances for all induced states of several patients, and further perform a significance analysis according to the normality test results. If significance is satisfied, execute step S5 to optimize the pacing point position; otherwise, the pacing point in step S2 is already in the optimal interval and no optimization is required.

[0069] In this step, to ensure the applicability of subsequent statistical analysis methods and the reliability of data interpretation, four classic statistical methods, namely the D'Agostino-Pearson test, Anderson-Darling test, Shapiro-Wilk test, and Kolmogorov-Smirnov test, are used to perform a normality test on the stimulation point-infarct region distances for all induced states of several patients to analyze the distribution characteristics of the data set. The input data for each test method is the stimulation point-infarct region distances for different induced states (stable reentry, unstable reentry, non-sustained reentry) (n samples, without distinguishing patients).

[0070] Among them, the D'Agostino-Pearson test determines whether the skewness and kurtosis information of the stimulus point-infarct region distance data are close to the theoretical values of the normal distribution. This method does not rely on sample sorting, is applicable to any sample size, and is sensitive to the joint distribution of skewness and kurtosis.

[0071] The Anderson-Darling test is used to measure the deviation between the sample distribution of the stimulus point-infarct region distance data and the theoretical normal distribution. Higher weights are assigned to the tail regions, making it more sensitive to tail deviations.

[0072] The Shapiro-Wilk test compares the covariance structure between the ordered statistics of the data and the theoretical expectation to determine whether the stimulus point-infarct region distance data comes from a normal distribution. It performs better in small samples, but it relies on sample sorting.

[0073] The Kolmogorov-Smirnov test judges normality by evaluating the maximum difference between the empirical distribution function and the theoretical distribution function of the stimulus point-infarct region distance data. It does not require preset distribution parameters and relies on the parameter estimation of the theoretical distribution.

[0074] After completing the multiple normality tests, further determine whether there are significant differences in the stimulus point-infarct region distance under different induced states (such as stable reentry, unstable reentry, non-sustained reentry). The present invention sets the judgment criteria as follows:

[0075] If among the four normality test methods used, the p-values of at least two are significantly greater than the significance level of 0.05, that is, p > 0.05, then the null hypothesis that the data follows a normal distribution can be accepted. On this basis, for the stimulus point-infarct region distance data with normal distribution characteristics, the one-way analysis of variance (ANOVA) method is used to evaluate whether there are significant differences in this variable among multiple groups under different induced states, so as to judge the statistical relevance between the induced state and the distance from the scar region.

[0076] If the normality test results do not meet the above criteria, that is, the data distribution significantly deviates from the normal distribution hypothesis, then the present invention further selects the Kruskal-Wallis test as a non-parametric method to replace the analysis of variance, so as to compare the distribution differences between multiple groups of stimulus point-infarct region distances without distribution assumptions and judge whether there are significant differences.

[0077] Select the difference analysis method in combination with the normality test results, verify the data distribution characteristics through statistics, and accordingly judge the relevance between different induced states and the distance between the stimulus point and the infarct region. Dynamically select parametric or non-parametric tests according to the data distribution characteristics to ensure the reliability of the conclusion.

[0078] S5. Aggregate the data of the distances between the stimulation points and the infarct regions in all stable reentry states, and analyze through a violin plot to obtain the range of the distances between the stimulation points and the infarct regions that are prone to induce reentry.

[0079] This process focuses on the data of the distances between the stimulation points and the infarct regions in the stable reentry state, visualizes its distribution characteristics through a violin plot, and identifies the high-density interval as the critical distance range prone to induce reentry. The present invention combines visual analysis and statistical inference, and the optimized range of the distances between the stimulation points and the infarct regions provides a quantitative basis for the selection of clinical ablation targets. If there is a large deviation between the simulation effect and clinical reentry, reselect the pacing point within the range of the distances between the stimulation points and the infarct regions that are prone to induce reentry as the new stimulation point, optimize the simulation scheme, obtain the reentry position close to the real situation of clinical patients, and use the key reentry path as the ablation target for the doctor's ablation surgery, thereby improving the treatment effect of patients with myocardial infarction.

[0080] In a specific implementation of the present invention, only the samples with the induction state of "stable reentry" are retained, and the data of unstable reentry and non-sustained reentry are excluded. The construction principle of the violin plot is kernel density estimation. For example, the Gaussian kernel algorithm generates a probability density curve, and the data distribution density is shown through a smooth curve. The horizontal axis is the induction state (i.e., stable reentry), and the vertical axis is the distance value, as Figure 5 shown. A median line, an interquartile range, and a density peak area are superimposed inside the violin. Among them, the median line represents the typical distance value, reflecting the distances between the stimulation points and the infarct regions in most reentry states. The interquartile range represents the core range, and the density peak area represents the data aggregation area. In this embodiment, the interquartile value is taken as the upper limit of the distance range, and the distance point corresponding to the density peak area is taken as the optimized distance between the stimulation point and the infarct region within the range from 0 to the interquartile value. Here, the density peak area refers to the peak point of the ordinate and its nearby area.

[0081] In this embodiment, a simulation system for optimizing cardiac reentry induction based on the infarct region distance is also provided, including:

[0082] A ventricular model construction module, which establishes a personalized high-fidelity ventricular model based on the CMR_LGE images of several arrhythmia patients;

[0083] A stimulation point-infarct region distance modeling module, which performs modeling calculations on each high-fidelity ventricular model to obtain the data of the distances between the left ventricular stimulation points and the infarct regions, including:

[0084] Dividing the ventricular model into regions and randomly selecting pacing points as stimulation points to induce VT, and obtaining the induction state of each stimulation point; the induction state includes unstable reentry, stable reentry, and non-sustained reentry;

[0085] Extract the left ventricular mesh model from the ventricular model, calibrate the left ventricular myocardial infarction region and its boundary, and measure the distance between the left ventricular stimulation point and the infarction region at the pacing point;

[0086] A statistical test module, which is used to perform a normality test on the distances between the stimulation points and the infarction regions in all induced states of several patients, and further perform a significance analysis based on the normality test results. If significance is satisfied, the pacing point position needs to be optimized; otherwise, the pacing point is already in the optimal range and no optimization is required;

[0087] A pacing point position optimization module, which obtains the distance range between the stimulation point and the infarction region that is prone to induce reentry by summarizing the distance data between the stimulation point and the infarction region in all stable reentry states, and this range is used as the optimization interval.

[0088] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The implementation methods of the remaining modules are not elaborated here. The system embodiment described above is only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present invention. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0089] The embodiments of the system of the present invention can be applied to any device with data processing capabilities, and the any device with data processing capabilities can be a device or apparatus such as a computer. The system embodiment can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a logically meaningful device, it is formed by the processor of any device with data processing capabilities reading the corresponding computer program instructions in the non-volatile memory into the memory for operation.

[0090] The above embodiments only represent several implementation manners of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. For those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A simulation method for optimizing cardiac reentry induction based on the distance of the infarct region, characterized in that, The following steps are involved: Step 1: Establish a personalized high-fidelity ventricular model based on CMR_LGE images of several patients with arrhythmia; Step 2: For each high-fidelity room model, perform the following operations: The ventricular model is divided into regions and a pacemaker point is randomly selected as a stimulation point to induce VT, and the induced state of each stimulation point is obtained; the induced state includes unstable reentry, stable reentry, and non-sustained reentry; Extract the left ventricular mesh model from the ventricular model, calibrate the left ventricular myocardial infarction area and its boundary, and measure the distance between the left ventricular stimulation point and the infarction area for the pacemaker point; Step 3: Perform a normality test on the distances between the stimulation point and the infarct area in all induced states of several patients, and further perform a significance analysis based on the normality test results. If the significance is met, perform step 4 to optimize the pacemaker position; otherwise, the pacemaker point in step 2 is already in the optimal interval and does not need to be optimized; Step 4, summarizing the stimulation point-infarction area distance data of all stable reentry states to obtain the stimulation point-infarction area distance range that is easy to induce reentry.

2. The simulation method for optimizing cardiac reentry induction based on the distance of the infarct region according to claim 1, wherein The high-fidelity laboratory model establishment process is as follows: Preprocessing of CMR_LGE images of patients with arrhythmia, including segmentation of the myocardium and its infarcted area, interpolation and reconstruction of the ventricle, and obtaining a ventricular model; Dividing the ventricular model into finite element meshes to obtain a ventricular mesh model; The myocardial fiber rotation direction was constructed for the ventricular grid model, and the corresponding myocardial cell electrophysiological models were coupled to normal tissue, gray matter tissue, and complete infarction tissue respectively.

3. The simulation method for optimizing cardiac reentry induction based on the distance of the infarct region according to claim 1, wherein The process of extracting the left ventricle mesh model from the ventricle model includes: The myocardial area is projected onto the MRI image based on the variational implicit function interpolation method. Starting from the selected seed points on the left and right ventricular endocardium, the ventricular cavity is gradually expanded and labeled through the region growing algorithm. Finally, the left and right ventricular labels are generated to identify the left and right ventricular regions. The labels are then converted into a recognizable data format and mapped to the ventricular model generated by S1. Finally, a grid data file with only the left ventricular grid is generated to obtain the extracted left ventricular grid model.

4. The simulation method for optimizing cardiac reentry induction based on the distance of the infarcted area according to claim 1, characterized in that, The process of demarcating the left ventricular myocardial infarction area and its boundaries includes: According to the tissue type to which the volume mesh in the left ventricle mesh model belongs, the volume mesh tetrahedron is divided into one or two of normal tissue, gray matter tissue and complete infarction tissue, and a normal myocardial point set, a gray matter point set and a complete infarction point set are obtained respectively; Demarcate the infarct area: merge the gray matter point set and the complete infarct point set to obtain the complete infarct area point set on the myocardium; Demarcate the boundary of the infarct area: find all subsets that belong to both the normal myocardial point set and the gray matter point set as the transition boundary of the semi-infarct area; find all subsets that belong to both the normal myocardial point set and the complete infarct point set as the transition boundary of the complete infarct area; merge the point set of the transition boundary of the semi-infarct area and the point set of the transition boundary of the complete infarct area to obtain a complete infarct area boundary point set on the myocardium.

5. The simulation method for optimizing cardiac reentry induction based on the distance of the infarct region according to claim 4, characterized in that The infarct area point set includes the infarct area boundary point set.

6. The simulation method for optimizing cardiac reentry induction based on the distance of the infarct region according to claim 4, wherein The process of measuring the distance between the left ventricular stimulation point and the infarct area for the pacemaker includes: For each pacemaker point that induces a state, first determine whether this point exists in the infarct region point set file. If so, there is no need for further calculation, directly skip this point, and record the distance between the stimulation point and the infarct region as 0; otherwise, enter the iterative search process; In each iteration process, find the tetrahedral mesh where the current infarct region boundary point set is located, remove the points in the current infarct region boundary point set from the vertex set of the tetrahedral mesh as the expanded infarct region boundary point set until the pacemaker point first appears in the new boundary point set, then record the number of iterations at this time as the corresponding distance between the stimulation point and the infarct region, and the unit of this distance is the number of layers of the grid cell.

7. The simulation method for optimizing cardiac reentry induction based on the distance of the infarcted area according to claim 1, characterized in that, In step 3, the normality test uses the D'Agostino-Pearson test, Anderson-Darling test, Shapiro-Wilk test, and Kolmogorov-Smirnov test.

8. The simulation method for optimizing cardiac reentry induction based on the distance of the infarcted area according to claim 7, characterized in that If the p-values of at least two test methods are significantly greater than the significance level of 0.05, then use the one-way analysis of variance method to evaluate whether there is a significant difference in this variable among multiple groups under different induced states; otherwise, use the Kruskal-Wallis test method to evaluate the significant difference.

9. The simulation method for optimizing cardiac reentry induction based on the distance of the infarct region according to claim 1, characterized in that, In step 4, obtain the distance range between the stimulation point and the infarct region that is prone to induce reentry through violin plot analysis, take the quartile value as the upper limit of the distance range, and take the distance point corresponding to the density peak region within the range from 0 to the quartile value as the optimized distance between the stimulation point and the infarct region.

10. A simulation system for optimizing cardiac reentry induction based on the distance of the infarcted area, which is used to implement the method described in claim 1, characterized in that, The simulation system includes: A ventricular model construction module, which based on the CMR_LGE images of several arrhythmia patients, establishes a personalized high-fidelity ventricular model; A distance modeling module between the stimulation point and the infarct region, which performs modeling calculations on each high-fidelity ventricular model to obtain the left ventricular stimulation point-infarct region distance data, including: Dividing the region of the ventricular model and randomly selecting a pacemaker point as the stimulation point to induce VT, and obtaining the induced state of each stimulation point; the induced state includes unstable reentry, stable reentry, and non-sustained reentry; Extracting the left ventricular grid model from the ventricular model, calibrating the left ventricular myocardial infarction region and its boundary, and measuring the distance between the left ventricular stimulation point and the infarct region for the pacemaker point; A statistical test module, which is used to perform a normality test on the distances between the stimulation points and the infarct regions of all induced states of several patients, and further perform a significance analysis according to the normality test results. If significance is satisfied, the position of the pacemaker point needs to be optimized; otherwise, the pacemaker point is already in the optimal interval and no optimization is required; A pacemaker point position optimization module, which obtains the distance range between the stimulation point and the infarct region that is prone to induce reentry by summarizing the distance data between the stimulation points and the infarct regions in all stable reentry states, and this range is used as the optimization interval.

Images