Irreversible electroporation damage evaluation system and method based on myocardial dielectric properties

Abstract

The invention discloses an irreversible electroporation damage assessment system and method based on myocardial dielectric properties, solves the problem of how to assess damage in a pulsed electric field ablation process, and belongs to the technical field of medical equipment. The system comprises an acquisition module and a workstation, wherein the acquisition module acquires an intracardiac ECG signal, a global electric field voltage signal of a target cardiac cavity, and impedance and temperature between ablation electrode pairs in real time; the work station processes the collected signals to obtain electrocardio characteristics, ablation parameters, temperature between the ablation electrode pairs, impedance drop, contact force and contact state between the ablation electrode pairs and myocardial tissue and ablation electrode position stability, the obtained signals are input into the evaluation model, and the evaluation model outputs IRE injury evaluation classification results. The evaluation model is realized based on a neural network. And real-time measurement and evaluation are carried out on PF ablation injury so as to assist a user to complete comprehensive, objective and accurate evaluation of the ablation effect.

Description

Technical Field

[0001] The invention relates to a pulse electric field ablation damage assessment method, belonging to the technical field of medical equipment. Background Art

[0002] Atrial fibrillation (AF) is the most common arrhythmia, significantly increasing the risk of death, stroke, heart failure (HF), and cognitive impairment, and its prevalence increases with age.

[0003] Pulse Field Ablation (PFA) is a major breakthrough in the treatment of atrial fibrillation. Unlike traditional thermal ablation techniques (radiofrequency ablation and cryoablation), PFA uses intermittent biphasic high-intensity electric fields to selectively cause non-thermal IRE damage to myocardial cells to block ectopic electrical signals and achieve pulmonary vein isolation, which can reduce the risk of esophageal, pulmonary vein and phrenic nerve damage.

[0004] During PFA treatment, it is critical to conduct damage assessment on the generated myocardial ablation lesions, which can provide important decision-making information for whether to continue ablation treatment, thereby improving the safety and effectiveness of PFA treatment. At present, doctors can only determine the instantaneous effectiveness of ablation by changes in intracardiac ECG (Electrocardiogram) before and after ablation to achieve acute pulmonary vein isolation, but cannot assess the degree of myocardial damage and cannot increase the probability of sustained pulmonary vein isolation in patients. Summary of the invention

[0005] In order to solve the problem of how to evaluate the damage during pulsed electric field ablation, the present invention provides a system and method for evaluating irreversible electroporation damage based on the dielectric properties of the myocardium.

[0006] The present invention provides an irreversible electroporation injury assessment system based on myocardial dielectric properties, comprising an acquisition module and a workstation;

[0007] An acquisition module, used for real-time acquisition of intracardiac ECG signals, global electric field voltage signals of target cardiac chambers, impedance and temperature between ablation electrode pairs;

[0008] The workstation is used to obtain ablation parameters in real time, and is also used to extract features from intracardiac ECG signals to obtain electrocardiographic features; it is also used to determine the coordinates of the ablation electrode based on the global electric field voltage signal of the target heart chamber, and to determine the position stability of the ablation electrode in combination with the ablation point; it is also used to determine the impedance drop, the contact state between the ablation electrode pair and the myocardial tissue, and the contact force between the ablation electrode pair and the myocardial tissue based on the impedance between the ablation electrode pairs; the electrocardiographic features, ablation parameters, the temperature between the ablation electrode pairs, the impedance drop, the contact force between the ablation electrode pair and the myocardial tissue, the contact state and the position stability of the ablation electrode are input into an evaluation model, and the evaluation model outputs an IRE injury evaluation classification result, and the evaluation model is implemented based on a neural network.

[0009] Preferably, in the acquisition module, the positioning patch is attached around the simulated imaging area, and a global three-dimensional electric field is generated in the simulated imaging area, and the global electric field voltage signal of the target heart chamber during the ablation process is acquired through the ablation electrode or the mapping electrode.

[0010] Preferably, the coordinates of the ablation electrode are determined according to the global electric field voltage signal of the target cardiac chamber, and the position stability of the ablation electrode is determined in combination with the ablation point:

[0011] The collected global electric field voltage signal is preprocessed, and the cardiac cavity image and ablation catheter morphology are reconstructed according to the preprocessed voltage signal, and the fitted ablation catheter morphology is superimposed on the reconstructed cardiac cavity image in the same coordinate system. The coordinates of the ablation electrode are determined according to the relative movement of the ablation catheter in the cardiac cavity, and the position stability of the ablation electrode is determined according to whether the coordinates are within the effective ablation area.

[0012] Preferably, the method for preprocessing the collected global electric field voltage signal comprises:

[0013] The collected global electric field voltage signal V is whitened to obtain a voltage signal V′ with zero mean and unit variance;

[0014] Taking the catheter characteristics as a reference, the voltage signal V′ is subjected to a three-dimensional affine transformation to obtain the Euclidean space coordinates of the voltage collection location, i.e., the electrode coordinates R;

[0015] Collect respiratory signals, perform heartbeat and respiratory motion compensation on the electrode coordinates R, and obtain the compensated electrode coordinates R′;

[0016] The cardiac cavity image and ablation catheter morphology fitting are reconstructed according to the compensated electrode coordinates R′.

[0017] Preferably, the breathing signal is extracted by collecting a voltage signal from an electrocardiogram patch attached to the body surface.

[0018] Preferably, the method for determining the contact state between the ablation electrode pair and the myocardial tissue according to the impedance between the ablation electrode pair is as follows:

[0019] The impedance Z between electrode pairs during ablation was collected using ablation electrodes. adj , according to the impedance Z between the electrode pairs adj Calculate the contact state value P:

[0020] P=(Z adj -Z min ) / (Z max -Z min )

[0021] Z max and Z min are the maximum and minimum impedances between electrodes, respectively;

[0022] According to the threshold range of the contact state value P, the sticking state is determined, and then the contact state is determined.

[0023] Preferably, the method for determining the contact force comprises:

[0024] Construct multiple prediction data sets of impedance-contact force between electrode pairs, use the prediction data sets to train the polynomial regression algorithm model, and use the ablation electrode to collect the impedance Z between electrode pairs during the ablation process. adj , the impedance Z between the electrode pairs adj The input is into the trained polynomial regression algorithm model, and the polynomial regression algorithm model outputs the predicted contact force.

[0025] Preferably, the evaluation model includes multiple neural networks, and ECG characteristics, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and resistance, contact state and position stability are simultaneously input into each neural network. The output results of all neural networks are voted on, and the winning output result is the IRE injury assessment classification result. The neural network adopts an extreme learning machine or a BP neural network.

[0026] Preferably, a deep learning network is used to extract features from the ECG signal to obtain ECG features.

[0027] The present application also provides a method for evaluating irreversible electroporation injury based on myocardial dielectric properties, including:

[0028] Establishing a training set, wherein the inputs in the training set include electrocardiographic characteristics, ablation parameters, temperature between ablation electrode pairs, impedance drop, contact force between ablation electrode pairs and myocardial tissue, contact state, and stability of ablation electrode position, and the output is an IRE injury assessment classification result corresponding to the input; using the training set to train an assessment model, wherein the assessment model is implemented based on a neural network;

[0029] During the evaluation, the real-time intracardiac ECG signal, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position are obtained, and the intracardiac ECG signal is feature extracted to obtain ECG features;

[0030] The acquired real-time ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position, and extracted ECG features are input into the trained evaluation model, and the evaluation model outputs the IRE injury assessment classification results.

[0031] The beneficial effects of the present invention are as follows: the present invention comprehensively considers the influence of ECG characteristics, ablation parameters, temperature between ablation electrode pairs, impedance drop, contact force between ablation electrode pairs and myocardial tissue, contact state and stability of ablation electrode position on ablation damage, and provides real-time measurement and evaluation of PF ablation damage to assist users in completing a comprehensive, objective and accurate evaluation of ablation effects. In addition, the present invention has the functions of high-precision imaging of cardiac cavities, electrode positioning navigation and ECG mapping, and can quickly and accurately locate the ablation position. For pulsed field ablation, the intelligent evaluation system for irreversible electroporation damage of myocardial dielectric properties has a specificity and sensitivity greater than 80% for the evaluation of tissue damage after ablation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the irreversible electroporation injury assessment system based on the dielectric properties of the myocardium;

[0033] Figure 2 Attach a schematic diagram for positioning the patch;

[0034] Figure 3 This is a flowchart of imaging navigation and positioning of the present invention;

[0035] Figure 4 Collecting diagrams for dielectric properties between electrodes;

[0036] Figure 5 It is an indication diagram of electrode contact force;

[0037] Figure 6 Schematic diagram of the IRE injury assessment model. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0040] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0041] The irreversible electroporation injury assessment system based on the dielectric properties of the myocardium of this embodiment is as follows: Figure 1 As shown, it includes acquisition module, data processing unit, workstation, display, power module and other components;

[0042] The evaluation function of this embodiment will be based on a multimodal fusion model, combining ECG characteristics, ablation parameters, temperature between ablation electrode pairs, impedance drop, contact force between ablation electrode pairs and myocardial tissue, contact status and ablation electrode position stability, to provide users with PF ablation transmural injury evaluation in near real time on the interactive interface.

[0043] With the workstation as the core of the system, the acquisition module collects the required data. The PF ablation device and the workstation are connected to the data processing unit through communication cables to build a data transmission network. The data processing unit pre-processes the input data and transmits it to the workstation. The workstation can configure the parameters of the data processing unit and compatible PF ablation device, and process the reported PF ablation parameters and the data pre-processed by the data processing unit, and then display the imaging, electrode positioning and navigation, mapping, electrode contact pressure, damage assessment and other results on the user interface through the display.

[0044] The acquisition module of this embodiment can use the matching acquisition patch and intracardiac mapping / ablation electrodes to display the 3D electroanatomical map of the human heart and surface / intracardiac ECG activity in near real time, and superimpose the electrode position on the heart chamber image to meet the electroanatomical map, navigation and mapping requirements during cardiac electrophysiology surgery; based on the acquired dielectric properties between the electrodes, before ablation, it provides the contact status, position stability and contact force indications between the electrode and the tissue, and after the ablation is completed, it provides the ablation damage assessment results in near real time.

[0045] The components of the intelligent assessment system for irreversible electroporation injury based on myocardial dielectric properties are described as follows.

[0046] The acquisition module includes acquisition patches, mapping electrodes / ablation electrodes and thermistors, which are connected to the data processing unit through their respective circuits. The acquisition patches include positioning patches and ECG patches.

[0047] Positioning patch: attached around the intended imaging area, such as Figure 2As shown, the data processing unit channel is connected through respective circuits. The data processing unit generates a global three-dimensional low electric field within the imaging range through three pairs of patches in the X\Y\Z directions, and the electric field frequency is lower than 30KHz. The positioning patch RL is the potential reference for the voltage measured by the data processing unit.

[0048] ECG patch: a 12-lead electrode patch, connected to the data processing unit through its own circuits. The RL ECG patch is different from the positioning patch RL and is only used as a reference for the 12-lead ECG. The surface ECG module in the data processing unit will perform high-speed acquisition and data processing of the surface ECG signal, and then regularly package and upload it to the workstation.

[0049] Mapping electrode / ablation electrode: connected to the data processing unit channel through the ablation catheter connection line. The data processing unit collects the intracardiac ECG signal at the electrode and the global electric field voltage signal of the target heart chamber at high speed through the ablation electrode, and can generate a local low electric field between the electrodes with a frequency of 10KHz to 200KHz to obtain the impedance between the ablation electrode pairs. The impedance between the ablation electrode pairs can feedback the contact state between the electrode and the tissue, the contact force, and the damage to the tissue after ablation.

[0050] The mapping electrode is smaller in size than the ablation electrode, has a larger range of motion, can detect more global electric field voltage signals of the target cardiac chamber, and has better positioning and navigation effects, so the mapping electrode can be used additionally for positioning and navigation.

[0051] Thermistor is used to detect temperature signal.

[0052] Data processing unit: includes multiple parallel channels, the number of channels can be expanded, each channel is designed with a transceiver circuit, used in conjunction with the acquisition patch and intracardiac mapping electrode / ablation electrode, the transmitting circuit is a low current source circuit with extremely high output impedance (greater than 100KΩ), which is insensitive to the equivalent load impedance (human body), and the receiving circuit has extremely high input impedance, which can effectively reduce the impact of sampling on the original signal, and obtain the patch voltage signal, surface / intracardiac ECG signal, global electric field voltage signal of the target heart chamber and impedance between ablation electrode pairs in near real time, and after filtering the received signal, it is packaged and uploaded to the workstation at a regular interval.

[0053] Power module: provides power supply for the data processing unit.

[0054] PF ablators and multi-channel recorders: They must be compatible with the system. Through the matching cables, the PF ablators report the ablation data and ablation results to the data processing unit; the surface / intracardiac ECG signals of the data processing unit are forwarded to the multi-channel physiological recorder.

[0055] Workstation: Connect to the data processing unit through a communication cable, and build a communication network with the PF ablation device connected to the data processing unit. Based on this, the workstation can configure the parameters of the data processing unit and the PF ablation device, and perform proprietary algorithm calculations on the ablation parameters reported by the PF ablation device and the data information collected by the data processing module, and display the relevant information in a graphical user interface. The workstation of this embodiment is used to obtain ablation parameters in real time, and is also used to extract features from intracardiac ECG signals to obtain electrocardiographic features; it is also used to determine the coordinates of the ablation electrode according to the global electric field voltage signal of the target cardiac chamber, and determine the stability of the ablation electrode position; it is also used to determine the impedance drop and the contact state of the ablation electrode pair with the myocardial tissue according to the impedance between the ablation electrode pairs, and it is also used to determine the contact force between the ablation electrode pair and the myocardial tissue; the electrocardiographic features, ablation parameters, the temperature between the ablation electrode pairs, the impedance drop, the contact force between the ablation electrode pairs and the myocardial tissue, the contact state and the stability of the ablation electrode position are input into the evaluation model, and the evaluation model outputs the IRE injury evaluation classification results. The evaluation model of this embodiment is implemented based on a neural network.

[0056] In this embodiment, the coordinates of the ablation electrode are determined according to the global electric field voltage signal of the target cardiac chamber, and the method for determining the position stability of the ablation electrode is as follows:

[0057] In order to realize cardiac cavity imaging and electrode positioning navigation functions, the workstation software will process the catheter electrode voltage measured by the data processing unit according to the following process, and finally display it in real time in the graphical user interface.

[0058] Data whitening: The global electric field voltage V at the catheter electrode is whitened to make the data have zero mean and unit variance, and the electrode voltage V′ is obtained to reduce data redundancy. ZCA is the whitening matrix of V.

[0059] V′=VW ZCA (1)

[0060] Voltage positioning algorithm based on three-dimensional affine: With the catheter characteristics as reference, the electrode voltage V′ after whitening is transformed into three-dimensional affine to obtain the Euclidean space coordinates of the voltage collection point, i.e., the electrode coordinates R. The specific catheter characteristics include the physical spacing and electrical spacing of the catheter electrodes. In the following formula, T is the transformation matrix of V′.

[0061] R = f V2R (V′|T) (2)

[0062] Respiratory gating: It is software gating. Respiratory signals are extracted based on the surface electrode voltage values ​​collected by the ECG patch. The electrode coordinates R are compensated for heartbeat and respiratory motion. The compensated electrode coordinates R′ are used for catheter fitting and imaging. In the following formula, R0 is the output of the point cloud filtering (such as radius filtering) of R to remove outliers and reduce the point cloud density; Wg is the heartbeat and breathing compensation coefficient under different phases; g is the stage indication divided according to the breathing phase.

[0063] R′=gating(R0,W g ,g) (3)

[0064] Cardiac cavity reconstruction: As the electrode continues to collect points in the imaging area, the compensated electrode coordinate points increase to form a sparse spatial point cloud. Based on the point cloud 3D reconstruction algorithm, such as the algorithm in the open source library PCL (Point Cloud Library) and Open3D, the imaging model is continuously reconstructed during the imaging process as the point cloud data is updated.

[0065] Catheter fitting: Based on the compensated electrode coordinates R′ and the known physical properties between the electrodes, the ablation catheter morphology is fitted, and the electrode coordinates after fitting are R″.

[0066] Electrode positioning navigation: The fitted ablation catheter shape is superimposed on the reconstructed 3D cardiac cavity image in the same coordinate system to indicate the relative movement of the ablation electrode in the imaging area. The coordinates of the ablation electrode are determined based on the relative movement of the ablation catheter in the cardiac cavity, and the stability of the ablation electrode position is determined based on whether the coordinates are within the effective ablation area.

[0067] In this embodiment, the electrode contact state and contact force are realized by a regression algorithm, specifically including:

[0068] Ablation electrode and tissue contact status: A large number of impedances between ablation electrode pairs obtained in the early application process are linearly normalized, and then according to the actual clinical ablation process and result feedback, the scene thresholds (P=0~1) such as catheter electrode not touching, general touching, and good touching are set. When applied, it normalizes the impedances between ablation electrode pairs collected in real time, and indicates the contact status between the electrode and the myocardial tissue on the graphical interface according to the threshold set by the system. The following formula is the normalization of the dielectric properties between electrodes, Z max and Z min is the statistically significant maximum and minimum impedance between electrodes, Z adj The impedance between the ablation electrode pairs currently obtained by the system.

[0069] P=(Z adj -Z min ) / (Z max -Z min ) (4)

[0070] In other embodiments, the abutment indication may also be indirectly characterized based on the contact force between the electrode and the tissue.

[0071] Ablation electrode and tissue contact force: It can be fitted and indicated by the impedance between the ablation electrode pairs. For PFA catheters, it mainly indicates the contact force between the ablation electrode pairs and the tissue. The impedance data between multiple electrode pairs are input into the pre-trained polynomial regression algorithm model, and the contact force between the ablation electrode pairs and the myocardial tissue is output respectively. The electrode pairs with the largest and smallest contact forces and their specific values ​​are marked in the graphical interface, as shown in the attached figure. Figure 5 The contact force between the electrode pair and the tissue is exponentially related to the impedance between the electrodes, and the influence of the blood and tissue near the electrodes needs to be considered comprehensively.

[0072] In other embodiments, the contact force between the ablation catheter and the myocardial tissue may also be measured by a sensor at the catheter tip, and the sensor types include optical fiber type, strain gauge pressure sensor, capacitive pressure sensor, and the like.

[0073] The evaluation model of this embodiment is implemented based on a deep learning algorithm based on fusion features, specifically including:

[0074] Establishing a training set, wherein the inputs in the training set include electrocardiographic characteristics, ablation parameters, temperature between ablation electrode pairs, impedance drop, contact force between ablation electrode pairs and myocardial tissue, contact state, and stability of ablation electrode position, and the output is an IRE injury assessment classification result corresponding to the input; using the training set to train an assessment model, wherein the assessment model is implemented based on a neural network;

[0075] During the evaluation, the real-time intracardiac ECG signal, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position are obtained, and the intracardiac ECG signal is feature extracted to obtain ECG features;

[0076] The acquired real-time ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position, and extracted ECG features are input into the trained evaluation model, and the evaluation model outputs the IRE injury assessment classification results.

[0077] This embodiment integrates known features and features automatically extracted from ECG signals through one-dimensional convolutional neural networks. The known features include real-time acquired ablation parameters, temperature and impedance drop between electrode pairs, contact status, stability and contact force between electrode pairs and tissues, integrates the above features and inputs them into the evaluation model. The evaluation model of this embodiment includes multiple neural networks. The neural network uses an extreme learning machine (ELM) or BP (Backpropagation) neural network to evaluate and classify IRE injury results. In order to make the classification results more stable, prevent overfitting, and improve the generalization ability of the model, IRE injury evaluation will be based on the bootstrap aggregating (Bagging) algorithm. ECG features, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force, contact status and position stability between ablation electrode pairs and resistance values ​​are simultaneously input into each neural network, and the output results of all neural networks are voted. The winning output result is the IRE injury evaluation classification result. Using clinical data sets for training and verification, the specificity and sensitivity of IRE injury evaluation classification are both above 80%. The structure of the IRE injury evaluation algorithm is shown in the attached figure. Figure 6 shown.

[0078] In other embodiments, the fusion features may further include extracting electrocardiographic features, including intracardiac echocardiography features and optical coherence tomography features, by a deep learning algorithm.

[0079] In other embodiments, the ensemble learning algorithm may also use random forest or the like.

[0080] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. An irreversible electroporation injury assessment system based on myocardial dielectric properties, characterized in that: Including acquisition module and workstation; An acquisition module, used for real-time acquisition of intracardiac ECG signals, global electric field voltage signals of target cardiac chambers, impedance and temperature between ablation electrode pairs; The workstation is used to obtain ablation parameters in real time and also to extract features from intracardiac ECG signals to obtain ECG features; It is also used to determine the coordinates of the ablation electrode according to the global electric field voltage signal of the target heart chamber, and to determine the position stability of the ablation electrode in combination with the ablation point; it is also used to determine the impedance drop and the contact state between the ablation electrode pair and the myocardial tissue according to the impedance between the ablation electrode pairs, and is also used to determine the contact force between the ablation electrode pair and the myocardial tissue; the ECG characteristics, ablation parameters, temperature between the ablation electrode pairs, impedance drop, contact force between the ablation electrode pair and the myocardial tissue, contact state and the position stability of the ablation electrode are input into the evaluation model, and the evaluation model outputs the IRE damage assessment classification result, and the evaluation model is implemented based on a neural network.

2. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: In the acquisition module, the positioning patch is attached around the simulated imaging area, and a global three-dimensional electric field is generated in the simulated imaging area. The global electric field voltage signal of the target heart chamber during the ablation process is acquired through the ablation electrode or the mapping electrode.

3. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: The method of determining the ablation electrode coordinates based on the global electric field voltage signal of the target cardiac chamber and determining the ablation electrode position stability in combination with the ablation point: The collected global electric field voltage signal is preprocessed, and the cardiac cavity image and ablation catheter morphology are reconstructed according to the preprocessed voltage signal, and the fitted ablation catheter morphology is superimposed on the reconstructed cardiac cavity image in the same coordinate system. The coordinates of the ablation electrode are determined according to the relative movement of the ablation catheter in the cardiac cavity, and the position stability of the ablation electrode is determined according to whether the coordinates are within the effective ablation area.

4. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 3, characterized in that: The method for preprocessing the collected global electric field voltage signal includes: The collected global electric field voltage signal V is whitened to obtain a voltage signal V with zero mean and unit variance. ′ ; Taking the catheter characteristics as reference, the voltage signal V ′ Perform three-dimensional affine transformation to obtain the Euclidean space coordinates of the voltage collection point, that is, the electrode coordinates R; Collect the respiratory signal, perform heartbeat and respiratory motion compensation on the electrode coordinates R, and obtain the compensated electrode coordinates R ′ ; According to the compensated electrode coordinates R ′ Reconstruct cardiac chamber images and ablation catheter morphology fitting.

5. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 4, characterized in that: The breathing signal is extracted through the voltage signal collected by the ECG patch attached to the body surface.

6. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: Method for determining the contact state between the ablation electrode pair and the myocardial tissue based on the impedance between the ablation electrode pair: The impedance Z between electrode pairs during ablation was collected using ablation electrodes. adj , according to the impedance Z between the electrode pairs adj Calculate the contact state value P: P=(Z adj -WITH min ) / (WITH max -WITH min ) Z max and Z min are the maximum and minimum impedances between electrodes, respectively; According to the threshold range of the contact state value P, the sticking state is determined, and then the contact state is determined.

7. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: Methods for determining contact forces include: Construct multiple prediction data sets of impedance-contact force between electrode pairs, use the prediction data sets to train the polynomial regression algorithm model, and use the ablation electrode to collect the impedance Z between electrode pairs during the ablation process. adj , the impedance Z between the electrode pairs adj The input is into the trained polynomial regression algorithm model, and the polynomial regression algorithm model outputs the predicted contact force.

8. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: The evaluation model includes multiple neural networks, and ECG characteristics, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and resistance, contact state and position stability are simultaneously input into each neural network. The output results of all neural networks are voted, and the winning output result is the IRE injury assessment classification result. The neural network adopts an extreme learning machine or a BP neural network.

9. The irreversible electroporation injury assessment system based on myocardial dielectric properties according to claim 1, characterized in that: A deep learning network is used to extract features from ECG signals to obtain ECG features.

10. A method for evaluating irreversible electroporation damage based on myocardial dielectric properties, characterized in that: include: Establish a training set, the input of which includes ECG characteristics, ablation parameters, temperature between ablation electrode pairs, impedance drop, contact force between ablation electrode pairs and myocardial tissue, contact state and ablation electrode position stability, and the output is the IRE injury assessment classification result corresponding to the input; Using the training set to train an evaluation model, the evaluation model is implemented based on a neural network; During the evaluation, the real-time intracardiac ECG signal, ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position are obtained, and the intracardiac ECG signal is feature extracted to obtain ECG features; The acquired real-time ablation parameters, temperature and impedance drop between ablation electrode pairs, contact force between ablation electrode pairs and myocardial tissue, contact status and stability of ablation electrode position, and extracted ECG features are input into the trained evaluation model, and the evaluation model outputs the IRE injury assessment classification results.

Images